Beruflich Dokumente
Kultur Dokumente
version 5
of the Six Sigma®
TM
(DMAIC) process and
User’s Guide
Prepare custom
experimental designs
with Design of
a continuous process.
production through
quality control
User’s Guide
version5
www.jmpdiscovery.com
Worldwide, JMP® is used in the following markets: pharmaceuticals, medical devices, semiconductors, electronics manufacturing, consumer products
manufacturing, chemical manufacturing, agriculture, business and in various areas of industrial and academic research.
The first edition of JMP software was developed in 1989 linking statistics with graphics to explore data, make discoveries, and gain knowledge
for better decision making. JMP provides a complete and flexible design of experiments environment to help engineers, researchers, and
Six Sigma® practitioners create designs with fewer runs, providing quicker results at less cost.
SAS is the market leader in providing a new generation of business intelligence software and services that create true enterprise intelligence. SAS solutions
are used at more than 38,000 sites—including 99 of Fortune 100 businesses—to develop more profitable relationships with their customers and suppliers,
to make better, more accurate and informed decisions, and to drive their organizations forward.
TM
SAS, JMP, and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA
and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Six Sigma
is a registered trademark of Motorola, Inc. 47976US_0402 The Statistical Discovery Software™
Version 5
User’s Guide
JMP, A Business Unit of SAS
SAS Campus Drive
Cary, NC 27513
JMP ® User’s Guide, Version 5
Copyright © 2002 by SAS Institute Inc., Cary, NC, USA.
ISBN 1-59047-070-2
All rights reserved. Printed in the United States of America. No part of this publication may be
reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, or otherwise, without prior written permission of the publisher, SAS
Institute Inc.
Information in this document is subject to change without notice. The software described in this
document is furnished under the license agreement packaged with the software. The software may
be used or copied only in accordance with the terms of the agreement. It is against the law to copy
the software on any medium except as specifically allowed in the license agreement.
First printing, April 2002
JMP®, SAS®, and all other SAS Institute Inc. product or service names are registered trademarks of
SAS Institute Inc. All trademarks above are registered trademarks or trademarks of SAS Institute Inc.
in the USA and other countries. ® indicates USA registration.
Other brand and product names are registered trademarks or trademarks of their respective compa-
nies.
Imageman® is a registered trademark or trademark of Data Techniques, Inc. All rights reserved.
Microsoft Text to Speech Engine® is a registered trademark or trademark Microsoft Corporation.
All rights reserved.
Installer VISE™, Updater VISE®, and MindExpander® are trademarks of MindVision Inc. All
rights reserved worldwide.
Install Shield® is a registered trademark of Install Shield Software Corporation. All rights reserved.
iii
Contents
Contents
JMP User’s guide
Credits and Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
1 Preliminaries 1
What you need to know . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…about your computer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…about statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Learning about JMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…on your own with JMP Help . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…hands-on tutorials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
…reading about JMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
…if you are a previous JMP user . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Conventions and Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Starting JMP 7
Overview of the JMP Starter Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
JMP Starter and the File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
JMP Starter and the Analyze Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
The Basic Stats Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
The Modeling Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
The Multivariate Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
The Survival Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
JMP Starter and the Graph Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
The Graphs Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
The QC Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
JMP Starter and the DOE Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
JMP Starter and the Tables Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
The JMP Starter Index Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4 Characteristics of Data 75
JMP Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Types of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Modeling Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Analysis Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Freq Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Weight Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Numeric Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Date, Time and Date-Time Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Row States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Row State Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Selection Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Data Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Specialized Properties of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Contents
Opening a JMP File from a Web Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Cut and Paste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Drag and Drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Contents
Cursor Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Features and Formatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Formatting Analysis Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Context Menu Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Resizing Plots and Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Selecting Points in Plots and Using Row States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Customizing Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Using the Annotate Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Using the Draw Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Copy, Paste, Drag and Drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
Journal and Layout Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
The JMP Journal Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
The JMP Layout Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Index 267
Credits
Credits and Acknowledgments
Origin
JMP was developed by SAS Institute Inc., Cary, NC. JMP is not a part of the SAS System, though por-
tions of JMP were adapted from routines in the SAS System, particularly for linear algebra and proba-
bility calculations. Version 1 of JMP went into production in October, 1989.
Credits
JMP was conceived and started by John Sall. Design and development was done by John Sall,
Katherine Ng, Michael Hecht, Richard Potter, Brian Corcoran, Annie Dudley Zangi, Bradley Jones,
Charles Soper, Craige Hales, Kevin Hardman, and Chris Gotwalt. In the SAS Institute Technical Sup-
port division, Ryan Gilmore, Wendy Murphrey, Toby Trott, Peter Ruzsa, and Rosemary Lucas provide
technical support and conducted test site administration. Statistical technical support is provided by
Craig DeVault, Duane Hayes, and Kathleen Kiernan. Nicole Jones, Jianfeng Ding, Jim Borek, and
Kyoko Tidball provide ongoing quality assurance. Additional testing and technical support is done by
Noriki Inoue and Kyoko Takenaka from SAS Japan. Bob Hickey is the release engineer.
The JMP manuals were written by Ann Lehman, Lee Creighton, John Sall, Bradley Jones, and Erin
Vang, with contributions from Meredith Blackwelder, Annie Dudley Zangi, and Brian Corcoran. Edit-
ing, creative services, and production was done by SAS Publications. Melanie Drake implemented the
help system.
Thanks also to Georges Guirguis, Warren Sarle, Gordon Johnston, Duane Hayes, Russell Wolfinger,
Randall Tobias, Robert N. Rodriguez, Ying So, Warren Kuhfeld, George MacKensie, Bob Lucas,
Donna Fulenwider, and Mike Stockstill for statistical R&D support.
Acknowledgments
We owe special gratitude to the people that encouraged us to start JMP, to the alpha and beta testers of
JMP, and to the reviewers of the documentation. In particular we thank Michael Benson, Howard Yet-
ter, Andy Mauromoustakos, Al Best, Stan Young, Robert Muenchen, Lenore Herzenberg, Larry Sue,
Ramon Leon, Tom Lange, Homer Hegedus, Skip Weed, Michael Emptage, Pat Spagan, John Frei, Paul
Wenz, Mike Bowen, Lori Gates, Georgia Morgan, David Tanaka, Zoe Jewell, Sky Alibhai, David Cole-
man, Linda Blazek, Michael Friendly, Joe Hockman, Frank Shen, J.H. Goodman, David Ikle, Lou
Valente, Robert Mee, Barry Hembree, Dan Obermiller, Lynn Vanatta, and Kris Ghosh. Also, we thank
Dick DeVeaux, Gray McQuarrie, Robert Stein, George Fraction, Al Fulmer, Cary Tuckfield, Ron This-
ted, Nancy McDermott, Veronica Czitrom, Tom Johnson, and Avigdor Cahaner.
We also thank the following individuals for expert advice in their statistical specialties: R. Hocking and
P. Spector for advice on effective hypotheses; Robert Mee for screening design generators; Jason Hsu for
advice on multiple comparisons methods (not all of which we were able to incorporate in JMP); Ralph
O’Brien for advice on homogeneity of variance tests; Ralph O’Brien and S. Paul Wright for advice on
x
statistical power; Keith Muller for advice in multivariate methods, Harry Martz, Wayne Nelson,
Ramon Leon, Dave Trindade, Paul Tobias for advice on reliability plots; Lijian Yang and J.S. Marron
for bivariate smoothing design; George Milliken and Yurii Bulavski for development of mixed models;
Will Potts and Cathy Maahs-Fladung for data mining; and Clay Thompson for advice on contour plot-
ting algorithms.
For sample data, thanks to Patrice Strahle for Pareto examples, the Texas air control board for the pollu-
tion data, and David Coleman for the pollen (eureka) data.
Translations
For our localized releases, Erin Vang coordinated localization. Noriki Inoue, Kyoko Takenaka, and
Yusuke Ono of SAS Japan were indispensable throughout the project. Thanks to Junji Kishimoto of
SAS Japan, and special thanks to Professor Toshiro Haga (retired, Science University of Tokyo) and Pro-
fessor Hirohiko Asano (Tokyo Metropolitan University). Thanks to Hui Di for Chinese translation
review. Finally, thanks to all the members of our outstanding translation teams.
Past Support
Many people were important in the evolution of JMP. Special thanks Jeffrey Perkinson, David DeLong,
Mary Cole, Kristin Nauta, Aaron Walker, Ike Walker, Eric Gjertsen, Dave Tilley, Curt Yeo, Mike Pez-
zoni, Ruth Lee, Annette Sanders, Tim Christensen, Xan Gregg, Jeff Polzin, Alissa Aungvibool, and Eric
Wasserman. SAS Institute quality assurance by Jeanne Martin, Fouad Younan, and Frank Lassiter.
Additional testing for Versions 3 and 4 was done by Li Yang, Brenda Sun, Katrina Hauser, and Andrea
Ritter.
Also thanks to Jenny Kendall, Elizabeth Shaw, John Hansen, Eddie Routten, David Schlotzhauer, and
James Mulherin. Thanks to Steve Shack, Greg Weier, and Maura Stokes for testing JMP Version 1.
Thanks for support from Charles Shipp, Harold Gugel, Jim Winters, Matthew Lay, Tim Rey, Rubin
Gabriel, Brian Ruff, William Lisowski, David Morganstein, Tom Esposito, Susan West, Chris Fehily,
Dan Chilko, Jim Shook, Ken Bodner, Rick Blahunka, Dana C. Aultman, and William Fehlner.
Technology License Notices
JMP for the Power Macintosh was compiled and built using the CodeWarrior C compiler from Metro-
Works Inc.
SAS INSTITUTE INC.’S LICENSORS MAKE NO WARRANTIES, EXPRESS OR IMPLIED,
INCLUDING WITHOUT LIMITATION THE IMPLIED WARRANTIES OF MERCHANT-
ABILITY AND FITNESS FOR A PARTICULAR PURPOSE, REGARDING THE SOFTWARE.
SAS INSTITUTE INC.’S LICENSORS DO NOT WARRANT, GUARANTEE OR MAKE ANY
REPRESENTATIONS REGARDING THE USE OR THE RESULTS OF THE USE OF THE
SOFTWARE IN TERMS OF ITS CORRECTNESS, ACCURACY, RELIABILITY, CURRENT-
NESS OR OTHERWISE. THE ENTIRE RISK AS TO THE RESULTS AND PERFORMANCE
OF THE SOFTWARE IS ASSUMED BY YOU. THE EXCLUSION OF IMPLIED WARRANTIES
IS NOT PERMITTED BY SOME STATES. THE ABOVE EXCLUSION MAY NOT APPLY TO
YOU.
IN NO EVENT WILL SAS INSTITUTE INC.’S LICENSORS AND THEIR DIRECTORS,
OFFICERS, EMPLOYEES OR AGENTS (COLLECTIVELY SAS INSTITUTE INC.’S LICEN-
SOR) BE LIABLE TO YOU FOR ANY CONSEQUENTIAL, INCIDENTAL OR INDIRECT
xi
Credits
DAMAGES (INCLUDING DAMAGES FOR LOSS OF BUSINESS PROFITS, BUSINESS
INTERRUPTION, LOSS OF BUSINESS INFORMATION, AND THE LIKE) ARISING OUT
OF THE USE OR INABILITY TO USE THE SOFTWARE EVEN IF SAS INSTITUTE INC.’S
LICENSOR’S HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. BECAUSE
SOME STATES DO NOT ALLOW THE EXCLUSION OR LIMITATION OF LIABILITY FOR
CONSEQUENTIAL OR INCIDENTAL DAMAGES, THE ABOVE LIMITATIONS MAY NOT
APPLY TO YOU. SAS INSTITUTE INC.’S LICENSOR’S LIABILITY TO YOU FOR ACTUAL
DAMAGES FOR ANY CAUSE WHATSOEVER, AND REGARDLESS OF THE FORM OF THE
ACTION (WHETHER IN CONTRACT, TORT (INCLUDING NEGLIGENCE), PRODUCT
LIABILITY OR OTHERWISE), WILL BE LIMITED TO $50.
1
1 Preliminaries
Preliminaries
JMP is statistical software that gives you an extraordinary graphical interface to display and analyze
data. JMP is for interactive statistical graphics and includes
• a spreadsheet for viewing, editing, entering, and manipulating data
• a broad range of graphical and statistical methods for data analysis
• an extensive design of experiments module
• options to select and display subsets of the data
• data management tools for sorting and combining tables
• a calculator for each table column to compute values
• a facility for grouping data and computing summary statistics
• special plots, charts, and communication capability for quality improvement techniques
• tools for moving analysis results between applications and for printing
• a scripting language for saving frequently used routines.
JMP is easy to learn. Statistics are organized into logical areas with appropriate graphs and tables, which
help you find patterns in data, identify outlying points, or fit models. Appropriate analyses are defined
and performed for you, based on the types of variables you have and the roles they play.
JMP offers descriptive statistics and simple analyses for beginning statisticians and complex model fit-
ting for advanced researchers. Standard statistical analysis and specialty platforms for design of experi-
ments, statistical quality control, ternary and contour plotting, and survival analysis provide the tools
you need to analyze data and see results quickly.
1
Contents
What you need to know . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…about your computer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…about statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Learning about JMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…on your own with JMP Help . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
…hands-on tutorials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
…reading about JMP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
…if you are a previous JMP user . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Conventions and Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1 Preliminaries—What you need to know 3
1 Preliminaries
What you need to know
…about statistics
Even though JMP has many advanced features, you need only a minimal background of formal statisti-
cal training. All analysis platforms include graphical displays with options that help you review and
interpret the results. Each platform also includes access to help windows that offer general help and
some statistical details.
• Click the Index tab on the JMP Starter window to access the JMP statistical guide, which is a scroll-
ing alphabetical reference that tells you how to generate specific analyses using JMP and lets you
access further help for that topic.
• You can also choose Help from dialogs and from popup menus in JMP report windows.
• After you generate a report, context-sensitive help is available. Select the help tool (?) from the Tools
menu and click the report surface. Context-sensitive help tells about the items in the report window.
…hands-on tutorials
The JMP Introductory Guide is a collection of tutorials designed to help you learn JMP strategies. Each
tutorial uses a file from the Sample Data folder. By following along with these step-by-step examples,
you can quickly become familiar with JMP menus, options, and report windows.
1 Preliminaries
• Special information, warnings, and limitations are noted as such with the word ‘Note’.
• Reference to menu names (File menu) or menu items (Save command) appear in the Helvetica
bold font, similar to the way they appear on your screen.
• The notation to select a command from a menu is sometimes written File > New, meaning “select
the New command from the File menu.”
• Words or phrases that are important or have definitions specific to JMP are in italics the first time
you see them.
The chapters and appendix in this book are organized as follows:
Chapter 2, Starting JMP shows how to start a JMP session and documents the JMP Starter window and
its relationship to the main menus.
Chapter 3, The Menu Bar documents the main menu bar and describes each menu command.
Chapter 4, Characteristics of Data describes types of data, modeling types, data storage, informats and
formats, and other special properties of data.
Chapter 5, JMP Data Tables explains how to create new JMP tables; import and export data; navigate
the table to modify values; add and delete rows and columns; cut, paste, print; and save JMP table
information.
Chapter 6, The Tables Menu explains the Tables menu commands and shows how to summarize data,
and sort, subset, transpose, join, and concatenate JMP tables.
Chapter 7, Using the Formula Editor explains in detail how to use the JMP formula editor, describes the
formula editor components, shows how to create and modify formulas, introduces table templates, and
tells how to create your own table templates.
Chapter 8, Formula Editor Functions lists all the formula editor functions, and includes many examples.
Chapter 9, Report Windows and Surface Features describes the form and features of JMP results in report
windows.
Chapter 10, Personalizing JMP covers preferences and describes how to customize menus and save
them.
Appendix A, What’s New in JMP Version 5 gives an overview of new features in Version 5 of JMP.
2
2 JMP Starter
Starting JMP
The JMP Starter Window
When the JMP application first opens, you see the JMP Starter window. The JMP Starter is a set of tabs
pages, as illustrated below. The JMP Starter provides a guide and gateway to the facilities in JMP. It is a
good way to get started if you haven’t used JMP before. It gives access to most commands found on the
main menu or on toolbars. However, you don’t have to access JMP commands through the JMP
Starter. An experienced JMP user might want to close the JMP Starter window and proceed with other
chapters in this book.
This chapter gives an overview of the JMP Starter window and briefly describes its tab pages and the
items or commands on them. Each tab page description directs you to an appropriate chapter in this
book for details, or to the Statistics and Graphics Guide.
2
Contents
Overview of the JMP Starter Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
JMP Starter and the File Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
JMP Starter and the Analyze Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
The Basic Stats Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
The Modeling Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
The Multivariate Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
The Survival Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
JMP Starter and the Graph Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
The Graphs Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
The QC Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
JMP Starter and the DOE Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
JMP Starter and the Tables Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
The JMP Starter Index Tab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2 Starting JMP—Overview of the JMP Starter Window 9
2 JMP Starter
Overview of the JMP Starter Window
When the JMP application opens you first see the JMP Starter window, which is a set of tab pages as
illustrated in Figure 2.1. It gives access to most commands found on the main menu or on toolbars.
You can close the JMP Starter if you want; it is not required for running JMP. To open the JMP Starter
window, choose the JMP Starter command from the View menu (View > JMP Starter under Win-
dows) or the Macintosh Window menu (Window > JMP Starter) any time during a JMP session. This
window provides easy access to a broad scope of JMP functionality. It gives descriptions and help to
• create and manipulate JMP data tables
• use statistical commands in the Analyze menu
• generate graphs, overlay plots, spinning plots, contour plots, and use statistical quality control tools
from the Graph menu
• design experiments
• manipulate tables using features accessed by commands on the Tables menu.
The JMP Starter also has a tab that gives an extensive scrolling index of features in JMP.
This chapter briefly describes the JMP Starter and shows how to use it. Detailed descriptions and exam-
ples of all commands follow in this book or in the Statistics and Graphics Guide. If you are already famil-
iar with JMP you might want to move ahead and continue with the following chapters.
Figure 2.1 JMP Starter Tab Pages
2 JMP Starter
Figure 2.3 JMP Starter Basic Stats Tab
Distribution
Analyze > Distribution launches the Distribution platform, which describes the distribution of values
in a data table column using histograms and other graphical and text reports. For details about univari-
ate distributions see “The Moments Table for Continuous Variables,” p. 32 in the JMP Statistics and
Graphics Guide.
Fit Y by X
Analyze > Fit Y by X launches a report platform for each pair of columns in the active data table that
are assigned X and Y roles for the analysis. The type of analysis depends on the whether the variables are
continuous or categorical:
Fit model
Analyze > Fit Model fits one or more Y variables to a model of X variables. You can select the kind of
model appropriate to your data from the menu of fitting personalities. The kinds of available personali-
ties depend on the kind of responses you select.
See “Introduction to Model Fitting,” p. 165, “Standard Least Squares: Introduction,” p. 179, “Standard
Least Squares: Perspectives on the Estimates,” p. 211, “Standard Least Squares: Exploring the Prediction
Equation,” p. 233, and “Standard Least Squares: Random Effects,” p. 251 in the JMP Statistics and
Graphics Guide for a detailed discussion of the Fit Model dialog and Fit Model analyses.
Figure 2.4 JMP Starter Modeling Tab
Matched Pairs
Analyze > Matched Pairs looks at the mean difference between two (or more) columns. It plots the
points in a rotated scatterplot that presents them in terms of the differences and means of two variables.
2 Starting JMP—JMP Starter and the Analyze Menu 13
2 JMP Starter
The chapter “Paired Data,” p. 151 in the JMP Statistics and Graphics Guide, discusses matched pairs and
shows a unique graphical representation for that kind of analysis.
Fit Nonlinear
Analyze >Modeling > Nonlinear lets you specify nonlinear models, which are models that are nonlin-
ear in their parameters. Nonlinear launches an interactive fitting facility. You orchestrate the fitting pro-
cess as a coordination of three important parts of JMP: the data table, the Formula Editor, and the
Nonlinear Fit platform. Details on the JMP data table and Formula Editor are in the chapters “Charac-
teristics of Data,” p. 75, “JMP Data Tables,” p. 93, “The Tables Menu,” p. 119, and “Using the Formula
Editor,” p. 143 in this book.
See the chapter “Nonlinear Regression,” p. 335 in the JMP Statistics and Graphics Guide for more about
nonlinear fitting.
Time Series
Analyze >Modeling > Time Series looks at the relationship between a value and its lagged values. The
data must be sorted and have equal intervals between time periods. Time Series plots the time series
and autocorrelations, shows a variogram, and shows the spectral density.
“Time Series Analysis,” p. 499 in the JMP Statistics and Graphics Guide, discusses time series analysis
and gives examples and statistical details
Partition
Analyze > Modeling > Partition recursively partitions rows into groups according to X values that
associate with Y values. This partitioning creates a tree of partitions.
The factor columns (X’s) can be either continuous or categorical (nominal or ordinal). If an X is con-
tinuous, then the splits (partitions) are created by a cutting value, which divides the sample into values
below and values above this cutting value. If the X is categorical, then the sample is divided into two
groups of levels.
The response column (Y) can be either continuous or categorical (nominal or ordinal). If Y is continu-
ous, then the platform fits means, and creates splits which most significantly separate the means by the
sums of squares due to the means differences. If Y is categorical, then the response rates become the fit-
ted value. The most significant split can be determined by the largest likelihood ratio chi-square statis-
tic.
Neural Net
Analyze >Modeling > Neural Net is a particular case of a back propagation feed-forward mulit-
layer-perception neural net. The neural network is a set of nonlinear equations that predict output vari-
ables (Ys) from input variables (Xs) in using layers of linear regressions. JMP fits the neural net using
standard nonlinear least-squares regression methods.
14 2 Starting JMP—JMP Starter and the Analyze Menu
Multivariate
Analyze > Multivariate Methods > Multivariate gives you ways to explore how several variables relate
to each other. The platform begins with a standard correlation matrix. Options give additional options
for correlations and other techniques for looking at several variables.
2 JMP Starter
Discriminant
Analyze > Multivariate Methods > Discriminant is a platform that provides a method of predicting
the level of a one-way classification based on known values of the responses. The technique is based on
how close a set of measurement variables are to the multivariate means of the levels being predicted.
Optionally, you can do stepwise discriminant analysis.
PLS
Analyze >Multivariate Methods > PLS fits models using the partial least squares (PLS) method that
balances the two objectives of explaining response variation and explaining predictor variation. The
PLS techniques work by extracting successive linear combinations of the predictors, called factors (also
called components or latent vectors) that address one or both of these two goals. The PLS platform in
JMP also enables you to choose the number of extracted factors by cross validation, which involves fit-
ting the model to part of the data and minimizing the prediction error for the unfitted part.
See the chapter “Survival and Reliability Analysis I,” p. 433 in the JMP Statistics and Graphics Guide, for
more about the Survival platform.
Chart
Graph > Chart gives a chart for every numeric Y specified where the Ys are statistics to chart. The X val-
ues are always treated as discrete values. Chart automatically updates when values change in the current
data table and you then activate the chart window.
Overlay Plot
Graph > Overlay Plot gives an overlaid plot of a single X column and all numeric Y variables specified in
the Overlay Plot launch dialog. Optionally, the curves can be shown as separate plots for each Y with a
common X axis. Plots can be modified with range and needle options, color, log axes, and grid lines.
Overlay Plots automatically update when you change values in the current data table and then activate
the Overlay Plot report window.
2 Starting JMP—JMP Starter and the Graph Menu 17
2 JMP Starter
Figure 2.7 JMP Starter Graphs Tab
Spinning Plot
Graph > Spinning Plot produces a three-dimensional view of data and, optionally, an approximation
to higher dimensions through principal components. The plot is a spinnable display of the values of
numeric columns in the current data table. The Spinning Plot platform displays three variables at a
time from the columns you select.
To help capture and visualize variation in higher dimensions, the Spinning Plot platform displays a
biplot representation of the points and variables when you request principal components.
Contour Plot
Graph > Contour Plot constructs contours of a response in a rectangular coordinate system. A contour
plot required two variables for the X and Y axes and a third variable (denoted Y in the Contour Plot
Launch dialog) for contours.
Profiler
The output report that results from Analyze > Fit Model has an option that produces a prediction pro-
filer for the effect variables in a model, with one or more Y variables. The prediction profiler displays
prediction traces for each X variable. A prediction trace is the predicted response as one variable is
changed while the others are held constant at the current values. If you save the prediction equations
for a model, you can access the prediction profiler for that data table directly with Graph > Profiler
without rerunning the model.
Contour Profiler
The output report that results from Analyze > Fit Model has an option to do interactive contour pro-
filing. This is useful for optimizing response surfaces graphically. This interactive plot lets you adjust
both multiple responses and multiple factors simultaneously and shows acceptable factor setting regions
18 2 Starting JMP—JMP Starter and the Graph Menu
for response values you specify. If you save the prediction equations for a model, Graph > Contour
Profiler accesses the contour profiler for that data table directly without rerunning the model.
More information is found in the chapters “Bar, Line, and Pie Charts,” p. 519, “Overlay Plots,” p. 533,
“Three-Dimensional Viewing,” p. 543, and “Contour Plot,” p. 559 in the JMP Statistics and Graphics
Guide. The chapter “Standard Least Squares: Exploring the Prediction Equation,” p. 233 in the JMP
Statistics and Graphics Guide, covers the Profiler and the Contour Profiler.
The QC Tab
The QC tab page, shown in Figure 2.8, accesses the commands on the Graph menu that are used in sta-
tistical quality control. Also, the Capability item on this tab page accesses the Distribution command
found on the Analyze menu.
Control Charts
Graph > Control Charts offers a variety of ways to analyze and monitor process data with control
charts. It displays control charts that update dynamically as samples are received. Control charts are a
graphical and analytic tool for deciding whether a process is in a state of statistical quality control.
Shewhart control charts are broadly classified into control charts for variables and control charts for
attributes. Moving average charts and cumulative sum (cusum) charts are special kinds of control charts
for variables.
See the chapters “Statistical Control Charts,” p. 567, and “Shewhart Control Charts,” p. 589 in the JMP
Statistics and Graphics Guide for a discussion of statistical quality control and examples that use differ-
ent kinds of control charts.
Figure 2.8 JMP Starter QC Tab
2 JMP Starter
Variability Chart
Graph > Variability Chart performs a variability analysis, also called a Gage R&R analysis. In a Gage
R&R analysis, a number of supposedly identical parts are taken from a production line. Each one is
measured by several operators a number of times using different measuring instruments. You use vari-
ability charts to examine the magnitudes of the variation due to operators, parts, and instruments.
The chapter “Variability Charts,” p. 615 in the JMP Statistics and Graphics Guide, documents the plots
and reports produced by the Variability platform.
Pareto Plot
Graph > Pareto Plot creates a bar chart (Pareto chart) that displays the severity (frequency) of prob-
lems in a quality-related process or operation. A Pareto chart is a statistical quality improvement tool
that shows frequency, relative frequency, and cumulative frequency of problems. The bars are ordered
by frequency in decreasing order from left to right, which makes a Pareto chart useful for deciding what
problems should be solved first.
The chapter “Pareto Plots,” p. 633 in the JMP Statistics and Graphics Guide, discusses simple
before-and-after Pareto charts and two-way comparative Pareto charts.
Ternary Plot
Graph > Ternary Plot creates ternary plots, which display the distribution and variability of three-part
compositional data. Examples of compositional data are the proportion of sand, silt, and clay in soil, or
the proportion of three chemical agents in a trial drug. You can use data expressed in proportions, or in
absolute measures. The Ternary Plot platform converts absolute values to proportions.
The ternary plot in JMP also has a contour feature, which means you can use a response measure at
each point to form a contour plot within the ternary structure.
See the chapter “Ternary Plots,” p. 647 in the JMP Statistics and Graphics Guide, for details and exam-
ples of ternary plots.
Capability Analysis
The Capability item accesses the Distribution command on the Analyze menu, which has a platform
option for capability analysis. The Capability Analysis option gives a capability analysis for quality
control applications. A capability study measures the deviation of a process from given specification
limits. A dialog prompts you for Lower Spec Limit, Upper Spec Limit, and Target. These are optional
fields, and only those fields you enter are part of the resulting Capability Analysis table. Optionally, you
can enter a known value for sigma, the process standard deviation.
See the chapter “Univariate Analysis,” p. 25 in the JMP Statistics and Graphics Guide, for examples of
capability analysis.
Diagram
The Diagram platform is used to construct Ishikawa charts, also called fishbone charts, or
cause-and-effect diagrams. These charts are useful to organize the sources (causes) of a problem (effect),
20 2 Starting JMP—JMP Starter and the DOE Menu
perhaps for brainstorming, or as a preliminary analysis to identify variables in preparation for further
experimentation.
OC Curves
The OC Curves item accesses a set of JSL scripts that compute and plot operating characteristic curves.
The chapter “Statistical Control Charts,” p. 567 in the JMP Statistics and Graphics Guide, gives details
about how to run the OC curves scripts. The JMP Scripting Guide is the reference for scripting.
Custom Design
DOE > Custom Design generates a D-optimal design for the design you specify and, optionally, an
I-optimal design for repsonse surface designs. Custom designs can have continuous factors and categor-
ical factors with arbitrary numbers of level, interactions, mixture ingredients. A design can also have
covariates, which are factors that already have unchangeable values; a design is built around the covari-
ate. There is also a facility for constructing custom design models, including interaction terms and
polynomial terms for continuous factors. Factors can have inequality constraints. You can also specify
the number of experimental runs, which can be any number greater than or equal to the number of
terms in the model.
2 Starting JMP—JMP Starter and the DOE Menu 21
2 JMP Starter
Screening Design
DOE > Screening Design allows you to specify continuous factors, two- and three-level factors, and
categorical factors. It creates a table of classical screening designs from which you can pick.
Taguchi Arrays
DOE > Taguchi Arrays guides you through the definition of signal and noise factors. The signal factors
form the inner array and the noise factors form the outer array. The inner and outer array designs are
the traditional Taguchi orthogonal arrays such as L4, L8, or L16.
Mixture Design
DOE > Mixture Design lets you define a set of factors that are ingredients in a mixture. JMP displays a
dialog for choosing among several classical mixture design approaches such as simplex, extreme vertices,
and lattice. For the extreme vertices approach, you can supply a set of linear inequality constraints lim-
iting the geometry of the mixture factor space.
Augment Design
DOE > Augment Design can modify existing designs by adding center points, replicating the design,
creating a foldover design, or adding runs to the design using a model with more terms than the origi-
nal design.
2 JMP Starter
The JMP Starter Index Tab
The Index tab page on the JMP Starter window is an alphabetical list of statistical tools in JMP. When
you click on an item, a brief description of that method is displayed with instructions for generating
results. There is also help for these items using the Help main menu.
Figure 2.11 JMP Starter Index Tab
3
3 Menu Bar
The Menu Bar
This chapter is a reference chapter. It contains overview documentation for all the commands in the
JMP main menu bar. It is organized for easy reference. The menus and commands are presented in the
order they appear on the menu bar. Each menu title appears with the page number to help you locate
specific commands quickly.
Chapters later in this book describe the Rows and Cols menus in detail, give more information and
examples of Tables menu commands, explain the Formula Editor with examples of how to use it, and
cover details about report windows, personalizing JMP, and JMP web capabilities.
If you prefer to begin using JMP with a hands-on approach, you can skip this chapter and begin with
“JMP Data Tables,” p. 93, or refer to the Introductory Guide for tutorials.
When the JMP application opens, the first thing you see is the main menu bar and the JMP Starter
window (discussed in the previous chapter, “Starting JMP,” p. 7).
Default tool bars appear beneath the main menu. You can drag the tool bars and dock them on any
edge of the JMP window or tear them off and place them in any convenient location (Windows only).
Under Microsoft Windows, the Rows, and Cols menus appear only when a data table is open.
3 Menu Bar
The File Menu
Macintosh Option-File
The File menu has commands that perform file management or affect the JMP environment. Related
commands are grouped according to their functions. These groups contain commands that
• open and close data files and create new files
• save data tables
• access data from other databases
• set global options
• print data and report windows
• end a JMP session.
You modify a data table by working with its data table view in an active window. In descriptions of
menu commands, data table, table, and spreadsheet are used synonymously unless specifically noted oth-
erwise.
Under Windows, the New command gives a submenu to create a new data table or a new script win-
dow. Use Shift-New on the Macintosh to see the New > Script command.
28 3 The Menu Bar—The File Menu
Data Table
opens an empty untitled data table with one column labeled Column 1. To create more columns,
use the Rows and Columns popup menu commands and type or paste data into the data table.
Use Save As from the File menu to name the table and save it as a JMP file. Use the Open com-
mand to reopen a JMP data table after it has been saved.
Script
opens a new empty script window, which is an editable text window to enter and submit JSL
(JMP Scripting Language) commands.
Note: On the Macintosh the New command creates a new JMP table. Shift-File displays the New
Script command.
Macintosh Option-File
The Open command displays a specialized open file dialog to locate the file you want to open and spec-
ify the file format of the incoming file. The Open dialog displays information appropriate for the
incoming file type. The Open command directly reads JMP data tables, JMP journal files, JMP script
files, SAS data sets, SAS transport files, text files with any column delimiter, Excel® files, and Windows
CMD files. CMD (.JMPCMD) files are created when you modify the JMP main menu and save the
results.
Note: On the Macintosh, JMP opens SAS version 6 data sets Later versions of SAS are not supported
on the Macintosh. However, JMP on the Macintosh can open any SAS data that has been saved in SAS
transport format.
See the chapter “JMP Data Tables,” p. 93, for details and examples of using the Open command.
The Close command closes the active window. When you close a data table window while any of its
analysis windows are open, a dialog asks if you want to hide the data table only or close it and all related
analysis windows. Also, when you close the data table, a dialog asks you whether to save or discard
changes to the table. The Close command is the same as clicking the close box of the active window.
Close does not appear on the File menu unless a window is open.
Note: On the Macintosh, the Close command on the File menu closes the active window. Option-File
displays the Close all Data Windows command, which closes all window of the same type as the active
window.
3 The Menu Bar—The File Menu 29
3 Menu Bar
Macintosh Shift-File
The Import as Data command writes an open script file (JSL file) it to a JMP data table. This is partic-
ularly helpful when extracting tabular information from a web page. To do this, highlight the informa-
tion on a web page and copy it. Then open a new script window in JMP, and paste it to the script
window. Highlight the script text (usually a table of values) and use the Import as Data command to
create a new JMP table that contains that information. The import method conforms to the preference
settings you have for default import setting.
The Save command writes the active data table to a file. If the data table has been saved before, it is
rewritten to the same file name and location, replacing the old information. If the data table is new, the
Save command has the same effect as the Save As command. Saving a data table does not automati-
cally close it.
Note: JMP analysis windows are not saved with the data table. However, you can use JMP tools to copy
reports to other applications or you can use the journaling and layout features to save reports in text or
word processing format.
The Save As command writes the active data table to a file after prompting you for a name and file
type. It can save the data table as a JMP file, convert it to SAS Transport file, save as a Version 7 SAS
data set (Windows only), save it as a text format file, or save in any database format available on your
system. See the chapter “JMP Data Tables,” p. 93, for more information.
The Revert command restores the current data table to its condition when it was last saved. Revert is
dimmed on the File menu unless a data table has been edited.
On the Macintosh, Shift-File displays the Save Selection As command, for saving highlighted graph-
ics from an analysis report window. To highlight output, select the area selection (large plus) tool and
click or drag. When you choose Save Selection As, a dialog lets you choose a Macintosh picture,
PNG, or JPEG graphics format. Under Windows, the Save Selection As command is found in the
Edit menu.
The Database command opens a connection to any database on your system that has an ODBC
driver. File > Database initially displays the Database Open Table dialog, which lets you connect to or
disconnect from a specific database source. When one or more database connections are made, the dia-
log lists the connections and the tables in the selected database file or directory. Optionally, use the
Advanced button on the dialog to write SQL statements to open specific subsets of a table. Use the
30 3 The Menu Bar—The File Menu
dialog shown by the Database > Save Table command to save an open database table. There is a more
detailed example in the chapter “JMP Data Tables,” p. 93.
Under Windows, the Internet Open command lets you open an internet browser within the JMP envi-
ronment. This command displays an open dialog for you to enter a URL, and options to Browse, Edit
HTML, Edit HTML with tags stripped, Open a JMP file, or Extract HTML Table as JMP file.
The Preferences command displays tab pages with the following sets of options:
• General has check boxes for JMP start-up options, analysis destination, information to include
on output, report table style, and others.
• Background Color (Windows only) sets the background color of the report window. You can select
from a standard palette of colors or select a color on a color wheel and add that color to the color
palette.
• Fonts displays a list of fonts to set the style for text, headings, and titles.
• Graphic Formats (Windows only) has check boxes for selecting the graphic format when you copy
or drag graphics from report windows. There are also popup menus to select Metafile, JPEG, or
PNG as the graphic format when saving RTF files and HTML files.
• Communications Settings (Windows only) has settings for real-time data collection.
• File Location lets (Windows only) you enter a default path for standard JMP files such as the sample
data file, Help file, preferences file, and others.
• Platforms lists the analysis and graph platforms, with appropriate options for each.
• Text Import/Export lets you set default options, such as field and line delimiters, for reading and
writing text files.
See “Personalizing JMP,” p. 237 for further discussion of preferences.
Note: On the Macintosh OS 9 and earlier, the Preferences command is located in the Edit menu. On
the Macintosh OS X, Preferences is on the menu for the JMP application.
The Print command prints the active window. It displays the standard dialog for printing. The appear-
ance of the dialog depends on your operating system and printer driver.
3 The Menu Bar—The File Menu 31
3 Menu Bar
The Print Preview command (Windows only) displays the active window in the form it will print.
Print Preview is dimmed on the File menu unless there is an open window.
The Print Setup command (Macintosh only) displays the standard dialog for setting printed page char-
acteristics. The form of the dialog depends on your operating system and current printer driver.
The Send command (Windows only) submits the open window to the mail facility you define.
Under Microsoft Windows, the Recent Files command has a submenu that lists the JMP tables most
recently opened. When you click on a table’s name in this list, JMP opens the table.
The Exit command (Windows) or Quit command (Macintosh) closes all JMP windows (prompting
you to save changes) and quits the JMP application. On the Macintosh OS X, the Quit is the last com-
mand in the JMP application menu.
32 3 The Menu Bar—The Edit Menu
Macintosh Option-Edit
Macintosh Shift-Edit
The Edit menu contains standard commands. These commands operate on rows and columns that are
selected, on selected areas of reports, and on selected formula elements in the Formula Editor.
In the data table, you can select rows, columns, or both rows and columns at the same time, as shown
in Figure 3.1. Edit commands operate on entire rows if no columns are selected. Likewise, they operate
on whole columns if no rows are selected. When both rows and columns, Edit commands affect the
subset of values defined by the intersection of those rows and columns.
Figure 3.1 Example of Selected Rows and Columns
3 Menu Bar
To select a row in a data table, click the space that contains the row number. To select a column, click
above the column name either at the top of the column or in the Columns panel to the left of the data
table. To select multiple rows or columns, drag across them or Shift-click the first and last rows or col-
umns of a range. To make a discontiguous selection, use Control-click (Windows) or Command-click
(Macintosh) to highlight the rows and columns. To select a block of cells formed by the intersection of
rows and columns, drag the cross cursor diagonally across the subset of cells.
Note: You can also select special subsets of rows using the Row Selection command in the Rows
menu, described under “The Rows Menu,” p. 41 later in this chapter.
The Edit menu in conjunction with JMP tools can be used to copy all or part of active analysis report
windows. See “The Tools Menu and Tool Bars,” p. 63, for more information.
The Undo command cancels the effect of the most recent reversible Edit, Rows, or Columns com-
mand. If Undo is available its selection in the Edit menu appears as Undo command where command
is the most recent action. Most destructive data table operations (such as cut, paste, or delete rows) are
reversible. Undo dims when the most recent command is irreversible.
The Redo command reverses the effect of the Undo command.
Macintosh Option-Edit
The Cut command copies selected fields from the active data table to the clipboard and replaces them
with missing values. It is equivalent to Copy, then Clear. You can also use the Cut command to copy all
or part of a report window. However, Cut works like Copy in graphical displays (it does not clear the
copied image).
The Copy command copies the values of selected data cells from the active data table to the clipboard.
Copy copies entire rows when no columns are selected. Likewise, whole columns are copied when no
rows are selected. If you select both rows and columns, Copy copies the subset of cells defined by their
intersection.
The Copy command captures graphical displays or text reports defined by the area selection tool (large
plus) in the Tools menu.
Note: Data you cut or copy to the clipboard can be pasted into JMP tables or into other applications.
Pictures can be pasted into any application that accepts graphics.
If you use the Shift key and select Copy (Option-Copy on the Macintosh), JMP does a Copy with
Labels, which preserves the data table’s column labels in the copied image. Then use Shift-Paste
34 3 The Menu Bar—The Edit Menu
(Option-Paste with Labels on the Macintosh) to paste the information into a new column with the
column name intact.
Macintosh Shift-Edit
The Copy As Text command copies all text from the active report window (no graphical displays) as
unformatted text on the clipboard. On the Macintosh, hold down the Shift key to see Copy as Text on
the Edit menu.
Macintosh Option-Edit
The Paste command pastes information from the clipboard to the selected area in a JMP data table.
Paste can be used with the Copy command to duplicate rows, columns, or any subset of cells defined
by selected rows and columns.
If you use the Shift key and select Paste (Option-Paste on the Macintosh) JMP does a Paste with
Labels, which uses the first line of information on the clipboard as column headers.
The Clear command clears all selected cell values from the active data table and replaces them with
missing values. The values are not copied to the clipboard.
When a data table is the active window, the Select All command selects (highlights) all the rows in the
current data table. When an output, a layout, or a journal is the active window, Select All selects all the
objects in the window.
Under Windows, the Save Selection As command saves highlighted portions of an analysis window
in a graphic format. To highlight output, select the highlighter tool (shown above) and click or drag the
area you want. The Save Selection As dialog lets you choose between PNG (.PNG), JPEG (.JPG), and
WMF (.WMF) graphic formats.
Note: On the Macintosh, the Save Selection As command appears in the File menu when you hold
down the Shift key.
3 The Menu Bar—The Edit Menu 35
3 Menu Bar
Run Script submits a text file containing JMP Scripting Language (JSL) commands. You create a JSL
file two ways:
• Under Windows, select Script from File > New and manually key in JSL commands.
• Complete an analysis and then use Save Script to Script Window from a platform popup menu.
You can save a script that describes the whole analysis window or use a Save Script to Script
Window command for a single part of analysis within the output window.
To submit a portion of a script, first drag to highlight the portion you want and then use Edit > Run
Script. If no script commands are highlighted, the entire script is submitted.
See the JMP Scripting Guide for details on JSL syntax.
The Search command displays the submenu shown above. It gives you the ability to find and replace
text in data tables (including column names) in the usual way found in most word processing and edit-
ing programs. Find and replace actions deal only with character strings; numbers are treated as text and
appear to the Search command as they appear in the data table. All dialog actions also respond to the
keystroke shortcuts shown on the Search popup menu. For multiple search actions, it is sometimes
much more convenient to use the keystroke shortcuts instead of the Search dialog. See the chapter
“JMP Data Tables,” p. 93, for more details about search commands.
Note: the Undo command works only with Replace, and Replace and Find Next.
Under Windows, when a script window is open, the Go to Line command displays a dialog that lets
you enter the line number of the script, then sets the edit focus on that line.
The Journal command copies all information from the active data table or report window to an open
Journal window. The Journal window appears identical to the Report window. Each subsequent
Journal command appends the information from the active JMP window to the current journal file.
You can journal as much information as you need into a single open journal window. The area selection
tool (large plus tool) and the Cut and Paste commands in the Edit menu let you cut and paste sections
of JMP analysis reports into the Journal window.
36 3 The Menu Bar—The Edit Menu
The Layout command creates a new window that contains all information from the report window.
Under Microsoft Windows, when a Layout window is the active (front-most) window, the Tables,
Rows, and Cols menus are replaced with the Layout menu. The Align command in the Layout menu
has standard draw commands to Group and Ungroup selected areas of the Layout window. Ungrouped
pieces of the report can be moved anywhere on the report surface to rearrange them.
Note: You can open or close outline levels in a Layout window only until you have ungrouped layout
elements.
The commands in the Layout menu are discussed further in Chapter 10, “Report Windows and Sur-
face Features” chapter in this book.
Under the Windows operating systems, the Customize command has a submenu to customize Menus
or Toolbars. These commands each display a window that shows a sketch of all main menus or tool-
bars. These facsimiles are completely interactive. You can drag any command or toolbar icon out of its
position and place in any other position. Or, you can drag a command or icon out of its position and
into blank area to eliminate it from showing on the application. Revert to Factory Defaults resets the
menus and toolbars to the arrangement when you first installed JMP.
See “Personalizing JMP,” p. 237 for further discussion and examples of customizing JMP menus.
On the Macintosh, OS 9 and earlier systems, the last command in the Edit menu is the Preferences
command. On the macintosh OS X system, the Preferences command is on the JMP application
menu. Microsoft windows displays Preferences on the File menu. See the Preferences command
described previously in the File menu section for more information.
3 The Menu Bar—The Tables Menu 37
3 Menu Bar
The Tables Menu
The Tables menu commands (or Tables toolbar buttons) modify or create a new JMP table from one
or more existing tables. This section gives you an overview of each Tables menu command. See the
chapter “The Tables Menu,” p. 119, for details and examples of each Tables command.
The Summary command creates a JMP window that contains a summary table. This table summarizes
columns from the active data table, called its source table. The summary table has a single row for each
level of a grouping variable you specify. When there are several grouping variables, the summary table
has a row for each combination of levels of all grouping variables.
The summary table is linked to its source table; when you highlight rows in a summary table, the corre-
sponding source table rows highlight, as shown in Figure 3.2.
Figure 3.2 Summary table with Special Columns Menu
The illustration in Figure 3.2 shows a source table called Companies with a summary table grouped by
type of company. The Type column in the source table has the values “Computer” and “Pharmaceut”.
38 3 The Menu Bar—The Tables Menu
The Companies By (Type) summary table has two rows that show counts of 20 computer companies
and 12 pharmaceutical companies in the N Rows column.
The Subset command produces a new JMP table that is a subset of the active data table. If you click
Selected Rows in the Subset dialog, shown Figure 3.3, the new table has the rows and columns
defined by the highlighted rows and columns from the active data table. Highlight rows by
• selecting them on the data table
• highlighting histogram bars or points on a plot generated by the data table
• using one of the Row Selection commands from the Rows menu
• selecting one or more rows in a summary table produced by the Summary command (described
previously) to define a subset in its source table.
Figure 3.3 The Subset Dialog
If you click the Random Sample radio button, the sample is a random sample of the active data table.
Enter the sample size you want as a number or a proportion in the Sampling Rate text edit box.
Note: When you check the Link to original data table check box on the Subset dialog, the subset table
that results and any plot or graph of that subset table remain linked to the original table. Highlighting
rows in this kind of subset table highlights the corresponding rows in the original table and in all its
plots and graphs. If you hold down the Shift key and select Subset, a linked subset table is automati-
cally generated.
Also, if there are formulas in any columns check boxes appear on the Subset dialog to Copy Formula
from the source table, and Suppress Formula Evaluation of saved formulas when necessary to main-
tain data integrity. If the Copy Formula box is not checked, column formulas are not transferred to the
new data table.
3 The Menu Bar—The Tables Menu 39
3 Menu Bar
The Sort command sorts a JMP data table by one or more columns. The Sort command displays a dia-
log for you to specify columns as sort fields. Sorting a table has these options:
• The order in which you build a list of sort variables establishes the sorting order. Each variable is
sorted within the previous variable in the sort list.
• You can sort any column in the sort list in either ascending or descending order. By default, columns
sort in ascending order.
• If there are formulas in any columns other than the Sort columns of the data table to be sorted, the
Sort dialog shows a Copy Formula and Suppress Formula Evaluation check boxes. If the Copy
Formula box is not checked, column formulas are not transferred to the new data table.These check
boxes don’t appear if there are no formulas in the source table.
• By default, Sort creates a new data table. However, you have the option to sort in place; that is, the
sorted table replaces the original table. This option to replace the original table with the sorted table
is not available if there are any open report windows generated from the original table.
The Stack command creates a new data table from the active table by stacking specified columns into a
single new column. The values in other columns are preserved in the new data table. In addition, Stack
creates an ID column that identifies each row in the new table with values that are the corresponding
column names from the original table.
The table on the left in Figure 3.4, has columns called Column1, Column2, and Column3 that stack to
produce the table on the right. The stacked column, _Stack_, has a single column with values from the
original three columns. The _ID_ variable has the original column names as its values.
Figure 3.4 Example of Stacked Columns
Also, if there are formulas in any columns other than the Stack and ID columns of the data table to be
stacked, the Stack dialog shows Copy Formula and Suppress Formula Evaluation check boxes. If the
Copy Formula box is not checked, column formulas are not transferred to the new data table.
40 3 The Menu Bar—The Tables Menu
The Split command creates a new data table from the active table by splitting one or more columns to
form multiple columns. The new columns correspond to the values (levels) of an ID variable. Split
requires one or more columns whose combined values identify each row in the new table. Optionally,
the values in the other columns can be either dropped or preserved in the new data table.
Stack does the reverse of Split. If you use the Split command on the right-hand table, in Figure 3.4
(nine rows), the result is the original table on the left, with three rows.
Also, if there are formulas in any columns other than the Split and ID columns of the data table to be
split, the Split dialog shows Copy Formula and Suppress Formula Evaluation check boxes.
The Transpose command creates a new JMP table that is a transpose of the active data table. That is,
the columns of the original table are the rows of the new table, and the original table rows are the new
table’s columns. The new table has an additional column called Label, whose values are the column
names of the active table. If there is a label column in the active table, the values of that column are col-
umn names in the new transposed table. If there is no label column, the column names in the trans-
posed table are Row1, Row2, …Rown, where n is the number of rows in the original table.
The Concatenate command appends two or more tables end to end. The Concatenate dialog lists all
open JMP data tables. Select the tables to concatenate and click Concatenate on the Concatenate dia-
log. Concatenate creates a new untitled data table that consists of all rows in the first selected table fol-
lowed by all rows from the second table, and so on.
The arrangement of the new data table depends on the columns in the original tables. Concatenate
creates one column in the new table for each unique column name in all appended tables. Column
names that are the same in multiple tables stack into a single column. See the chapter “The Tables
Menu,” p. 119, for examples of concatenating tables.
Note: The Concatenate dialog also has a Save and Evaluate Formulas check box. If this box is not
checked, column formulas are not transferred to the new data table.
The Join command creates a new data table by joining two tables side by side, or by updating one data
table with values from another table. Tables can be joined
3 The Menu Bar—The Rows Menu 41
3 Menu Bar
• by row number
• by matching the values in one or more columns that exist in both data tables
• in a Cartesian fashion where all values in a column of one data table are merged with all values in a
column of another table.
Optionally, you can include all nonmatches in the new table or eliminate nonmatches that occur in
either of the original tables. Also, you can include all columns from both tables in the joined table or
include only a subset of columns you specify.
See the chapter “The Tables Menu,” p. 119, for details and examples of the Join command.
Also, if there are formulas in any columns of the data table to be joined, the Join dialog shows Copy
Formula and Suppress Formula Evaluation check boxes. If the Copy Formula box is not checked, col-
umn formulas are not transferred to the new data table.
You can access commands that affect highlighted rows from the Rows main menu, the red triangular
popup icon on the Rows panel to the left of the data table, or from the triangular popup menu in the
upper-left corner of the data table above the row numbers.
Rows menu commands can
• exclude highlighted rows from further analysis
• hide highlighted points in current graphical displays
• assign special colors or markers for graphical display of highlighted points
• select all rows or selecting rows assigned specific row state characteristics
42 3 The Menu Bar—The Rows Menu
Exclude/Unexclude is a toggle command used to exclude selected rows from statistical analyses. Data
remain excluded until you choose Exclude/Unexclude again for selected rows. Excluded rows are
marked by the “do not use” symbol (circle with a diagonal bar across it) showing in the row number
area.
Hide/Unhide is a toggle command that suppresses the display of points in all scatterplots. To hide data,
highlight their points in any plot or in the data table and select Hide/Unhide. Data remain hidden
until you choose Hide/Unhide again for selected rows. Hidden rows are marked by the mask symbol
showing in the row selection area, next to the row number.
Warning: Hidden points can be included in statistical analyses although they do not display in plots.
Likewise, points can be excluded from an analysis but not hidden. These conditions could cause mis-
leading plots and analyses.
Label/Unlabel is a toggle command that labels or identifies points on all scatter plots. To label points,
highlight them in the data table or in any plot, and select Label/Unlabel. By default, the row number is
used as the label value on plots. However, if you designate a column in the data table as a Label col-
umn, its values show as labels in plots instead of the row numbers. Data remain labeled until you
choose Label/Unlabel again for selected rows.
Exclude/Unexclude, Hide/Unhide, and Label/Unlabel status are row state characteristics. Active row
states are saved when you save the data table, and can be saved permanently with the data table in a spe-
cial row state column. Figure 3.5 shows examples of how active Hide, Exclude, and Labeled rows
appear on a data table.
3 The Menu Bar—The Rows Menu 43
3 Menu Bar
Figure 3.5 Examples of how Labeled, Hidden, and Excluded Row States in a Data Table
Note: You can have multiple columns as label variables. The value showing on a labeled variable in a
plot is the values of all label variables separated by commas. For more details about row states, see the
chapter “JMP Data Tables,” p. 93.
Colors changes highlighted points in all scatterplots to the colors you choose. To color a group of
points, select the appropriate rows in the data table or select the points in a plot. Then choose a color
from the Colors palette in the Rows menu.
There are 65 colors including shades of gray for distinguishing data points in JMP. The default color is
black.
Markers assigns a plot character to replace the standard points in scatterplots and spinning plots. To
assign one of the markers, select rows in the data table or select the corresponding points in a plot and
choose a marker from the Markers palette. The default marker is a dot.
Note: All row state conditions (including selection status) can be saved permanently with the data table
as active row states or saved in a special row state column in the data table.
The Next Selected command locates the first selected row after the current row and makes it blink
briefly. The current row is the first row when you open a data table. The current row changes to the
44 3 The Menu Bar—The Rows Menu
most recent row you edited or identified by a Next Selected or Previous Selected command. To set
the current row, Alt–click (Option–click on the Mac) anywhere in a row. Each time you choose the
Next Selected command, the next selected row is found and blinks. A beep signals when the last
selected row is located.
The Previous Selected command behaves the same as Next Selected but locates the first selected row
before the current row and makes it blink briefly.
The Row Selection command has a submenu with options for selecting
• a specific row number
• all rows in a data table
• a subset of rows based on the excluded, hidden or labeled row states
• a subset based on criteria you enter
• randomly select rows.
See the section called Selection Status, p. 87 in the “Characteristics of Data” chapter for details about
the Row Selection command.
The Clear Row States command clears all active row states in the data table. All rows become
included, visible, unlabeled, and show in plots as black dots. It does not affect row states saved in row
state columns.
The Color or Marker by Column command displays a dialog with a list of the variables in the current
data table. JMP uses the levels of the variable you select to color or mark points in plots.
The Row Editor command displays the dialog window that is useful for browsing the columns of a
data table one row at a time. When the row editor appears, its shows information for the first selected
row. To navigate the data table, use the buttons on the Row Editor, or the commands in the menu
accessible at the top of the dialog. For more information about using the row editor, see Entering and
Editing Data, p. 106 in the “JMP Data Tables” chapter.
3 The Menu Bar—The Cols Menu 45
3 Menu Bar
The Add Rows command displays a dialog that prompts for the number of rows to add and to specify
their location in the table. Radio buttons let you add rows at the beginning of the table (At start), the
end of the table (At end), or after a row number you specify (After row:). The new rows have missing
values that can be filled by typing or pasting in data.
The Move Rows command moves highlighted rows to the location you specify in the Move Rows dia-
log. Click the appropriate radio button to move highlighted rows to the beginning of the table (At
start), to the end of the table (At end), or after a specific row number (After row:).
The Delete Rows command deletes all selected rows from a JMP data table. Use the Undo command
on the Edit menu to undo an accidental deletion.
You can access commands that affect selected columns from the Cols main menu, from the Columns
panel to the left of the data table, or from the triangular popup menu in the upper-left corner of the
data table.
Commands in the Cols menu act on selected columns in the current data table. To select a column,
click the background area above the column name or click on the column name in the columns panel.
This area is called the column selection area. To highlight multiple columns, drag across their column
selection areas or Shift-click the first and last columns of a desired range. Use Control-click (Com-
mand-click on the Macintosh) to make a discontiguous selection.
Cols menu commands can:
• create or insert new columns and access the Formula Editor to compute column values
46 3 The Menu Bar—The Cols Menu
The New Column command gives a dialog to enter new column information. The dialog prompts for
column name, column characteristics and other column properties.
The Add Multiple Columns command displays a dialog that lets you add more than one column at a
time to a table. You specify the number of columns to add, their location, field width, and type. Specify
whether to add columns at the beginning of the table, after the rightmost column, or inserted between
columns.
By default, the new column are named Column 1, Column 2, and so forth, but the Add Multiple
Columns dialog accepts any prefix to be used in place of Column. Other column characteristics are the
same for all the new columns. Change the column names and characteristics by editing them in the
spreadsheet or in the Column Info dialog.
The Column Info command displays the dialog used by the New Column command, except there is no
Next button for adding new columns. You can use the Column Info command at any time to change
the attributes of one or more columns.
See the chapter “JMP Data Tables,” p. 93, for a complete discussion of column properties.
The Formula command displays the column’s formula editor to create a formula that computes col-
umn values. A plus sign shows next to the column name in the Columns panel when its values are cal-
culated.
The formula can include existing columns, constants, parameters, conditional logic, and a variety of
functions. A column of computed values is locked and cannot be edited. However, you can disassociate
a column from its formula by clicking the Remove Property button on the Column Info dialog for
that column. or clear the Formula Editor contents.
Note: You can also access a column’s Formula Editor with a context-click (right-click under Windows
and Control-click on the Macintosh) in the column heading area or on the column name in the Col-
3 The Menu Bar—The Cols Menu 47
3 Menu Bar
umns panel to access the popup menu shown above. For details about the Formula Editor see the chap-
ters “Using the Formula Editor,” p. 143, and “Formula Editor Functions,” p. 163.
The Preselect Role command assigns a role to the selected column and saves the role with the data
table. Launch dialogs then use the role to automatically fill in the dialog role boxes. Roles have specific
meanings that depend on the analysis being done. Roles are discussed in more detail in the chapter
“Characteristics of Data,” p. 75. Also, each analysis chapter in the JMP Statistics and Graphics Guide
talks about specific analysis roles.
The Validation command lets you enter a list of valid values (List Check) or valid range limit condi-
tions (Range Check) for a column, which then restricts the values that can be entered into the column.
The Label/Unlabel command applied to a column tells JMP to use this column’s values to identify
points in plots. When a column is a label column, the label icon (tag) shows next to the column name
in the Columns panel to the left of the data grid. Multiple columns can be label columns; a labeled
point shows the values of the label variables separated with a comma.
The Scroll Lock/Unlock command locks the selected column into the left-most position in the data
grid. When multiple columns are locked, they show in a block at the left in the data grid; when you
scroll through table columns, locked columns remain visible. A bold vertical line divides the locked and
unlocked columns.The name of a locked column is in italics in the Columns panel to the left of the
data grid.
The Hide/Unhide command hides or shows selected column in the spreadsheet but does not remove
them from the data table. To hide columns, select them and choose the Hide/Unhide command. A hid-
den column appears with the hidden icon (mask) next to the name in the columns panel.
48 3 The Menu Bar—The Cols Menu
When the Exclude/Unexclude command is chosen, that variable does not show in analysis role assign-
ment dialogs. The exclude icon shows next to the excluded column names in the Columns panel to the
left of the data grid.
To unlabel, unlock, or show columns, select them and again select the Label/Unlabel, Scroll Lock/
Unlock, or Hide/Unhide command.
Note: Hidden columns cannot be scroll locked.
The Standardize Attributes command displays a dialog with popup menus to select column attributes
and properties. The selections apply to all the selected columns in the data table when you click Apply.
For example, if you want to change all numeric columns to character, order the data table columns by
data type (or just select all the numeric variables) select all the numeric columns, and use the
Standardize Attributes command to simultaneously change them to character.
Reorder Columns lets you move columns according to the selection you make from its submenu:
Move Selected Columns
moves the selected columns To first (left-most in the data table), To last (right-most in the data
table), or After: (after a column you identify) in the Move Selected Columns dialog.
Original Order
returns the columns to the order they were in at the time the data table was last saved.
Reorder by Name
arranges the columns (except for row state columns) from left to right in alphabetical order by
column name.
Reorder By Data Type
arranges the columns from left to right in alphabetic order by data type (Character, Numeric,
Row State).
3 The Menu Bar—The DOE Menu 49
3 Menu Bar
Reorder By Modeling Type
arranges the columns from left to right in alphabetic order by modeling type (Continuous, Ordi-
nal, Nominal). Row State columns have not modeling type, and are shown last.
Reverse Order
reverses the order of the data table columns.
If you mistakenly move one or more columns, use the Undo command in the Edit menu to restore the
previous order.
The Delete Columns command removes selected columns from the data table. If you accidentally
remove columns, you can use the Undo command in the Edit menu to restore them.
The Design of Experiments (DOE) facility in JMP is an environment for describing the factors,
responses and other specifications needed to create a designed experiment and saving them in a JMP
data table. The specific designs are described briefly here and covered in detail in JMP Design of Experi-
ments, the user’s guide for DOE.
The Custom Design command provides the most flexibility of all design choices. With this option you
can have continuous factors, categorical factors with any number of levels, interactions, mixture ingre-
dients, and covariates (factors that have unchangeable fixed values). There is also a facility for construct-
ing custom-designed models that include interaction terms and polynomial terms for continuous
factors. You can also specify inequality constraints on the factors, and the number of experimental runs,
which can be any number greater than or equal to the number of terms in the model.
50 3 The Menu Bar—The DOE Menu
After specifying all your requirements, the design solution generates an optimal design for those
requirements.
The Screening Design command allows you to specify continuous factors, and 2- and 3-level categor-
ical factors. It creates a table of classical screening designs from which you can pick.
The Response Surface Design command lets you define a number of continuous factors. This envi-
ronment then offers a table of classical response surface designs from which to choose.
For Full Factorial Design command, you specify a set of continuous and categorical factors with arbi-
trary numbers of levels. JMP creates the design containing all possible combinations of those factors.
The Taguchi Arrays command guides you through the definition of signal and noise factors. The sig-
nal factors form the inner array and the noise factors form the outer array. The inner and outer array
designs are the traditional Taguchi orthogonal arrays, such as L4, L8, L16, and so forth.
The Mixture Design command lets you define a set of factors that are ingredients in a mixture. JMP
creates a new dialog for choosing among several classical mixture design approaches, such as simplex,
extreme vertices, and lattice. For the extreme vertices approach you can supply a set of linear inequality
constraints limiting the geometry of the mixture factor space.
The Augment Design command lets you modify existing designs. You can add center points, replicate
the design a specified number of times, create a foldover design, and add runs to the design using a
model with more terms than the original design.
See JMP Design of Experiments for a discussion of each type of design, with details and examples.
3 The Menu Bar—The Analyze Menu 51
3 Menu Bar
The Sample Size and Power command computes power, sample size, or the effect size you want to
detect for a given alpha and error standard deviation. You supply two of these values and the sample size
and power facility computes the third. If you supply only one of these values, the result is a plot of the
other two. This feature is available for the single sample, two sample, and k sample situations.
See JMP Design of Experiments for more discussion of prospective power analysis and examples.
Each Analyze command launches a platform. A platform is an interactive window you use to analyze
data, work with points on plots, and save results. The reports in a JMP analysis are organized hierarchi-
cally. Methods unfold that suit the context of your data. Many results display automatically, and more
are offered through popup menus.
The Analyze commands are briefly discussed in this section. The JMP Statistics and Graphics Guide
describes each command, its options, and gives examples.
Distribution describes a distribution of values with histograms and other graphical and textual reports:
• Continuous columns display a histogram and box plots. You can change the width of the histogram
bars using the hand tool (grabber) from the Tools menu. Optionally, you can test the mean and
standard deviation of the distribution and select from a variety of distribution fits. For continuous
variables, capability analysis is available.
• Nominal or ordinal columns are shown with a histogram of relative frequency for each level of the
ordinal or nominal variable and a mosaic (stacked) bar chart, with options to test probabilities.
52 3 The Menu Bar—The Analyze Menu
Text reports support each of the distribution plots. The reports show selected quantiles and moments of
continuous values. Tables of counts and proportions support nominal and ordinal values. Save com-
mands let you save information such as ranks, level numbers, standardized values, and other statistics as
new columns in the data table.
“Univariate Analysis,” p. 25 in the JMP Statistics and Graphics Guide, covers the Distribution command
in detail and gives examples.
Fit Y by X studies the relationship of two variables. This platform shows plots with accompanying anal-
yses for each pair of X and Y variables. The kind of analysis done depends on the modeling types (con-
tinuous, nominal, or ordinal) of the X and Y columns. Figure 3.6, illustrates each type of plot
produced by the different combinations of X and Y modeling types.
• If both X and Y have continuous modeling types, Fit Y by X displays a scatterplot. Using options,
you can explore various regression fits for the data and choose the most suitable fit for further analy-
sis. Each fit is accompanied by tables with supporting statistical analyses and parameter estimates.
• If X is nominal or ordinal and Y is continuous, Fit Y by X plots the distribution of Y values for each
discrete value of X. You can use options to see means diamonds and a box plot for each X value and
to compare group means with comparison circles. Accompanying text reports show a one-way anal-
ysis of variance table. Optionally, you can request nonparametric analyses, view multiple compari-
sons, and test homogeneity of variance.
• If X has continuous values and Y has nominal or ordinal values, Fit Y by X performs a logistic regres-
sion and displays a family of logistic probability curves. Tables show the log likelihood analysis and
parameter estimates for each curve.
Note: Logistic regression of ordinal columns is parameterized differently from logistic regression of
nominal columns and sometimes produces different results.
• If both X and Y are nominal or ordinal values, Fit Y by X shows a contingency table and a mosaic
bar chart. Accompanying tables show statistical tests, frequency, proportion, and chi-square values
for each cell. Optionally, you can request a correspondence analysis.
Individual chapters in the JMP Statistics and Grpahics Guide describe each type of analysis given by the
Fit Y by X command.
3 The Menu Bar—The Analyze Menu 53
3 Menu Bar
Figure 3.6 Fit Y by X Platforms
The Matched Pairs command handles bivariate data in the special situation where the two responses
form a pair of measurements coming from the same experimental unit or subject. For example, a
matched pair might be a before-and-after blood pressure measurement from the same subject. The
responses are correlated, and the statistical method called the paired t test takes that into account.
The Matched Pairs platform displays the data as a scatterplot of the difference between each pair (Y) by
the mean of each pair (X) and includes the standard paired t test, which is equivalent to testing that the
mean difference between the paired values is zero.
Fit Model lets you tailor an analysis using a model specific for your data. You select columns, assign
roles, and build the model to fit in the Fit Model dialog.
Fit Model performs a fit of one or more Y variables by the X variables selected. You can select the kind
of model appropriate to your data from the popup menu of fitting personalities given in the Fit Model
dialog. The fitting personalities available depend on the kind of responses you select. The following list
briefly describes the different fitting techniques:
54 3 The Menu Bar—The Analyze Menu
The Modeling command has the submenu shown above that launches the platforms:
Nonlinear
Nonlinear fits nonlinear models, which are models that are nonlinear in their parameters. The
Nonlinear command launches an interactive fitting facility. You orchestrate the fitting process as a coor-
3 The Menu Bar—The Analyze Menu 55
3 Menu Bar
dination of three important parts of JMP: the data table, the Formula Editor, and the Nonlinear Fit
platform.
You define the nonlinear prediction formula with the Formula Editor. Then choose Nonlinear with the
response variable as Y and the model column with its fitting formula in the X role. You interact with the
platform through the Nonlinear Fitting Control Panel using:
• buttons to start, stop, and step through the fitting process, and to reset parameter values
• fitting options to specify loss functions and computational methods
• a processing messages area
• a list of current and limit convergence criteria and step counts, current parameter estimates, and
error sum of squares
• options to specify the alpha level for confidence intervals and delta for numerical derivatives.
The Nonlinear platform can show the model and the derivatives of the model with respect to each of its
parameters, and the fitting solution reports. There are features that give confidence intervals on the
parameters and plot the resulting function if it is of a single variable. You can also save the SSE values in
a data table with a grid for plotting them.
The chapter “Nonlinear Regression,” p. 335 in the JMP Statistics and Graphics Guide, describes the
Nonlinear command in detail and gives examples.
Partition
The Partition platform recursively partitions rows into groups according to X values that associate with
Y values. This partitioning creates a tree of partitions.
The factor columns (Xs) can be either continuous or categorical (nominal or ordinal). If an X is contin-
uous, then the splits (partitions) are created by a ‘cutting’ value, which divides the sample into values
below and values above this cutting value. If the X is categorical, then the sample is divided into two
groups of levels.
The response column (Y) can be either continuous or categorical (nominal or ordinal). If Y is continu-
ous, then the platform fits means, and creates splits which most significantly separate the means by the
sums of squares due to the means differences. If Y is categorical, then the response rates becomes the fit-
ted value, the estimated probability for each response level. The most significant split can be deter-
mined by the largest likelihood-ratio chi-square statistic. In either case, the split is chosen to maximize
the difference in the responses between the two.
Neural Net
The Neural Net platform is a standard type of neural network. It is a particular case of a back propaga-
tion feed-forward multilayer-perception neural net. The neural network is a set of nonlinear equations
that predict output variables (Ys) from input variables (Xs) in a flexible way using layers of linear regres-
sions and S-shaped functions. This model is a useful nonlinear regression model, and JMP fits it using
standard nonlinear least-squares regression methods.
56 3 The Menu Bar—The Analyze Menu
Time Series
The Time Series platform lets you explore, analyze, and forecast univariate time series. The launch dia-
log (role assignment dialog) requires that one or more continuous variables be assigned as the time
series. Optionally, you can specify a time ID variable, which is used to label the time axis. If a time ID
variable is specified it must be continuous, sorted ascending, and evenly spaced with no missing values
The analysis begins with a plot of the points in the time series. In addition, the platform displays graphs
of the autocorrelations and partial autocorrelations of the series. These indicate how and to what degree
each point in the series is correlated with earlier values in the series. You can interactively add
• variograms—characterizations of process disturbances
• AR coefficients—autoregressive coefficients
• spectral density plots—period and frequency plots with white noise tests.
These graphs can be used to identify the type of model appropriate for describing and predicting (fore-
casting) the evolution of the time series. The model types include:
• ARIMA—autoregressive integrated moving-average, often called Box-Jenkins models
• Seasonal ARIMA—ARIMA models with a seasonal component
• Smoothing Model—several forms of exponential smoothing and Winters Method.
For more details about the Time Series platform, see the chapter “Time Series Analysis,” p. 499 in the
JMP Statistics and Graphics Guide.
The Multivariate Methods submenu has the commands shown above that launch the following plat-
forms:
Multivariate
Multivariate explores how multiple variables relate to each other and how points fit that relationship.
This platform helps you see correlations between two or more response (Y) variables, look for points
that are outliers, and examine principal components to look for factors.
The Multivariate platform appears showing correlations and a scatterplot matrix. Options give
• inverse and partial, and nonparametric correlations and pairwise correlations with accompanying
bar charts
• a matrix of bivariate scatterplots with a plot for each pair of Y variables
• a Mahalanobis distance outlier plot
• a jackknifed multivariate distance outlier plot where the distance for each point is calculated exclud-
ing the point itself.
3 The Menu Bar—The Analyze Menu 57
3 Menu Bar
There are options with these plots to save the distance scores.
You can also request principal components, standardized principal components, rotation of a specified
number of components, and factor analysis information.
The chapter “Correlations and Multivariate Techniques,” p. 365 in the JMP Statistics and Graphics
Guide, describes the Multivariate command.
Cluster
The Cluster command clusters rows of a JMP table. Cluster can perform a hierarchical or a k-means
clustering method.
The Hierarchical Cluster platform displays results as a tree diagram of the clusters called a dendrogram
followed by a plot of the distances between clusters. The dendrogram has a sliding cluster selector that
lets you identify the rows in any size cluster. There are options to save the cluster number of each row.
Cluster (hierarchical) uses these five clustering methods:
• Average linkage computes the distance between two clusters as the average distance between pairs of
observations, one in each cluster.
• Centroid method computes the distance between two clusters as the squared Euclidean distance
between their means.
• Ward’s minimum variance method (the default) uses the distance between two clusters as the Anova
sum of squares between the two clusters added up over all the variables.
• Single linkage uses the distance between two clusters that is the minimum distance between an
observation in one cluster and an observation in the other cluster.
• Complete linkage uses the distance between two clusters that is the maximum distance between an
observation in one cluster and an observation in the other cluster.
The k-means clustering approach finds disjoint clusters on the basis of Euclidean distances computed
from one or more quantitative variables. Every observation belongs to only one cluster—the clusters do
not form a tree structure as with hierarchical clustering. You specify the number of clusters you want.
The Cluster platform also has options to do normal mixture clustering and SOMs (self-organizing
maps).
The chapter “Clustering,” p. 379 in the JMP Statistics and Graphics Guide, describes the Cluster com-
mand in detail and shows clustering examples.
Discriminant
The Discriminant platform provides a method of predicting the level of a one-way classification based
on known values of the responses. The technique is based on how close a set of measurement variables
are to the multivariate means of the levels being predicted. Optionally, you can do stepwise discrimi-
nate analysis.
sive linear combinations of the predictors, called factors (also called components or latent vectors) that
address one or both of these two goals. The PLS platform in JMP also enables you to choose the num-
ber of extracted factors by cross validation, which involves fitting the model to part of the data and
minimizing the prediction error for the unfitted part.
The Survival and Reliability submenu has a submenu with the following commands:
Survival/Reliability
The Survival platform performs a univariate survival analysis using product-limit life table survival
computations with estimation of Weibull, lognormal, and exponential parameters.
Recurrence Analysis
The Recurrence Analysis platform looks at the age of a system when it requires a repair. A system can
have multiple repairs, each with its associated age, and is followed until it is no longer in service. A typ-
ical system might be some component of an engine or appliance.
Note: You can also use the Nonlinear Fit platform to handle nonlinear models with loss functions for
other parametric survival modeling.
See “Survival and Reliability Analysis I,” p. 433, “Survival and Reliability Analysis II,” p. 453, and
“Recurrence Analysis,” p. 477 in the JMP Statistics and Graphics Guide for a complete discussion of the
Survival submenu commands.
3 The Menu Bar—The Graph Menu 59
3 Menu Bar
The Graph Menu
Graph menu commands produce windows that contain specialized graphs or plots with supporting
tables and statistics. For a detailed discussion of each Graph menu command, see the JMP Statistics and
Graphics Guide.
The Chart command gives a chart for every numeric Y variable specified. The Ys are the variables
whose statistics you want to chart. Initially, a vertical bar chart appears, but there are options to show
horizontal bar charts, line charts, step charts, needle charts, point charts, or pie charts.
You can specify up to two X variables for grouping on the chart itself. The first X is the group variable,
and the second X is the level (subgroup) variable. If there is no X variable, then each row is a bar. The
chapter “Bar, Line, and Pie Charts,” p. 519 in the JMP Statistics and Graphics Guide, describes the Chart
command and shows examples.
The Overlay Plot command overlays a plot of a single numeric or categorical X column and all speci-
fied numeric Y variables. The axis can have either a linear or a log scale. Optionally, the plots for each Y
can be shown separately, with or without a common X axis.
By default, the values of the X variable are in ascending order, and the points are plotted in that order.
You have the option of plotting the X values as they are encountered in the data table.
Note: For scatterplots of two variables with regression fitting options, use the Fit Y by X command
instead of Overlay Plot.
The chapter “Overlay Plots,” p. 533 in the JMP Statistics and Graphics Guide, describes the Overlay Plot
command in detail and shows examples of plotting data.
60 3 The Menu Bar—The Graph Menu
Spinning Plot produces a three-dimensional spinnable display of values from any three numeric col-
umns in the active data table. It also produces an approximation to higher dimensions through princi-
pal components, standardized principal components, rotated components, and biplots. There are
options to save principal component scores, standardized scores, and rotated scores.
The Spinning Plot platform also gives factor-analysis-style rotations of the principal components to
form orthogonal combinations that correspond to directions of variable clusters in the space. The
method used is called a varimax rotation, and is the same method that is traditionally used in factor
analysis.
The chapter “Three-Dimensional Viewing,” p. 543 in the JMP Statistics and Graphics Guide, describes
the Spinning Plot command in detail and shows examples of plotting data and computing principal
components.
The Contour Plot command constructs a contour plot for a response variable, Y, for the values of two
X variables. Contour Plot assumes the X values lie in a rectangular coordinate system, but the observed
points do not have to form a grid. Some Contour Plot options are:
• show or hide data points
• show or hide triangulation and boundary
• specification and labeling of levels
• show a line contour or fill areas
The chapter “Contour Plot,” p. 559 in the JMP Statistics and Graphics Guide, describes the Contour
Plot command in detail and shows examples of plotting data.
The Control Charts command creates dynamic plots of sample subgroups as they are received and
recorded. Control charts are a graphical analytic tool used for statistical quality improvement. Control
charts can be broadly classified according to the type of data analyzed:
• Control charts for variables are used when the quality characteristic to be analyzed is measured on a
continuous scale.
• Control charts for attributes are used when the quality characteristic is measured by counting the
number of nonconformities (defects) in an item or by counting the number of nonconforming
(defective) items in a sample.
The concepts underlying the control chart are that the natural variability in any process can be quanti-
fied with a set of control limits, and that variation exceeding these limits signals a special cause of varia-
tion. In industry, control charts are commonly used for studying the variation in output from a
3 The Menu Bar—The Graph Menu 61
3 Menu Bar
manufacturing process. They are typically used to distinguish variation due to special causes from vari-
ation due to common causes.
The Control Chart platform offers the following kinds of charts:
• mean, range, and standard deviation
• individual measurement and moving range
• p-chart, np-chart, c-chart, and u-chart
• UWMA and EWMA
• CUSUM.
The “Statistical Control Charts,” p. 567 in the JMP Statistics and Graphics Guide, describes the Control
Charts command in detail.
In a Variability analysis, a number of parts assumed to be identical are taken from a production line.
Each one is measured several times by a number of operators using different measuring instruments.
You want to know the magnitudes of the variation due to operators, parts, and instruments. In the
same way that a Shewhart control chart can identify processes which are going out of control over time,
a variability chart can help identify operators, instruments, or part sources that are systematically differ-
ent in mean or variance.
The chapter “Variability Charts,” p. 615 in the JMP Statistics and Graphics Guide, describes the
Variability Chart command in detail.
The Pareto Plot command gives charts that display counts or the relative frequency of problems in a
quality-related process or operation. Pareto plots compare quality-related measures or counts in a pro-
cess or operation. The defining characteristic of Pareto plots is that the bars are in descending order of
values, which visually emphasizes the most important measures or frequencies.
Pareto Plot uses a single Y variable, called a process variable, and gives
• a simple Pareto plot when you do not specify an X (classification) variable
• a one-way comparative Pareto plot when you specify a single X variable.
• a two-way comparative plot when there are two X variables.
The Pareto plot facility does not distinguish between numeric and character variables or between mod-
eling types. All values are treated as discrete, and bars represent either counts or percentages.
The chapter “Pareto Plots,” p. 633 in the JMP Statistics and Graphics Guide, describes the Pareto Plot
command in detail.
62 3 The Menu Bar—The Graph Menu
The Ternary Plot command constructs a plot using triangular coordinates. The Ternary platform uses
the same options as the Contour platform for building and filling contours. In addition it a specialized
crosshair tool that lets you read the triangular axes values.
The chapter “Ternary Plots,” p. 647 in the JMP Statistics and Graphics Guide describes the Ternary Plot
command in detail and shows ternary plot examples.
The Diagram platform is used to construct Ishikawa charts, also called fishbone charts, or
cause-and-effect diagrams. These charts are useful to organize the sources (causes) of a problem (effect),
perhaps for brainstorming, or as a preliminary analysis to identify variables in preparation for further
experimentation. See “Ishikawa Diagrams,” p. 625 in the JMP Statistics and Graphics Guide for exam-
ples of Ishikawa charts.
The Profiler is available for tables with columns whose values are computed from model prediction for-
mulas. Usually, a Profiler plot results when you do a Standard Least Squares analysis and then request it.
However, if you save the prediction equation from the analysis, you can access the Prediction Profile
later from the Graph menu and look at the model using the response column with the saved prediction
formula.
The Prediction Profiler displays prediction traces for each X variable. A prediction trace is the predicted
response as one variable is changed while the others are held constant at the current values. The Predic-
tion Profiler is a way of changing one variable at a time and looking at the effect on the predicted
response. You interact with the Prediction Profiler; as you vary the value of an X variable, the Prediction
Profiler recomputes
• The low and high values show on the X-axis for each factor, showing its current value.
• The current predicted value of each Y variable for the current values of the X variables.
• Lines and markers within the prediction plots show how the predicted value changes when you
change the current value of an individual X variable and include the 95% confidence interval for the
predicted values shown by error bars above and below each marker.
Prediction profiles are especially useful in multiple-response models to help judge which factor values
can optimize a complex set of criteria.
3 The Menu Bar—The Tools Menu and Tool Bars 63
3 Menu Bar
The Contour Profiler command works the same as the Profiler command. It is usually accessed from
the Fit Model platform when a model has multiple response. However, if you then save the prediction
formulas for the responses, you can access the Contour Profiler at a later time from the Graph menu
and specify the columns with the prediction equations as the response columns. See the chapter “Stan-
dard Least Squares: Exploring the Prediction Equation,” p. 233 in the JMP Statistics and Graphics Guide,
for examples of the Profiler and the Contour Profiler.
The Tools menu toolbars are palettes of special tools that determine the effect of mouse actions. There
are general tools, and specific graphics tools that operate on points in plots. The default tool is the
arrow.
• When a data table is the active window, the arrow selects items in the Rows, Columns and Tables
panels. Click the arrow in the upper-left corner of the data grid to deselect rows and columns.
• When a report is the active window, the arrow highlights points in plots and histogram bars.
Other tools are listed next, with a brief description of each. The chapter “Report Windows and Surface
Features,” p. 213, covers the Tools menu in more detail.
The question mark icon accesses the JMP Help system. Select the help tool and then click graphs,
plots, or tables to access specific help in the JMP Help system.
64 3 The Menu Bar—The Tools Menu and Tool Bars
Use the large plus (selection tool) to select rows and columns in the data table or areas of a display win-
dow. When you click the selection tool on a section of a plot or report, that section highlights and can
be copied or dragged. Drag the large plus to extend the selection, or Shift-click on other report ele-
ments. Clicking near the edge of the report window highlights the entire report. To deselect, click a sec-
ond time in a highlighted area. The Copy command in the Edit menu copies highlighted display areas
to the clipboard.
Drag the scroller tool to scroll reports up or down and display only the results you want to see. The
scroller is a more precise way to perform the same function as the vertical scroll bar.
The hand (grabber) tool is for direct manipulation of plots, charts, axes, and formula components.
The brush tool is for highlighting an area of points in plots. When you click, a rectangle appears. Move
the rectangle over points to highlight them. Shift-click to extend the selection. Alt-click (Option-click
on the Macintosh) to change the size of the selection rectangle and also extend the selection.
The lasso tool lets you highlight an irregular area of points in plots. Drag the lasso around any set of
points. When you release the lasso, it automatically closes and highlights the points within the enclosed
area.
The magnifying (zoom) tool automatically zooms in on any area of a plot. When you click the magni-
fier, the point or area where you click becomes the center of a new view of the data. Use Alt-click under
Windows (Option-click on the Macintosh) at any time to restore the original plot. On a ternary plot
you can drag the magnifier tool to zoom the triangular axes.
The crosshair is a movable set of axes used to measure points and distances in graphical displays. The
values where the crosshair intersects the vertical and horizontal axis appear automatically as you drag
the crosshair within a plot. On a ternary plot, this tool displays triangular crosshair lines.
3 The Menu Bar—The View Menu 65
3 Menu Bar
The annotate tool places a text box wherever you click in a report, journal, or layout window. You can
key in notes to use as titles, footnotes, or other report annotation. The annotation box and its text can
be any color or can be transparent. Context-click to see a menu of options to tailor the appearance of
the annotate note.
The line, polygon, and simple shape tools work much like the Annotate tool. They behave in a similar
fashion to those in many draw packages. They can be used on the surface of any report, journal, or lay-
out window. Each has a context (right-click) menu of options to tailor its appearance.
• The line tool draws thin, thick, or dashed lines and can have arrows at either end or both ends.
• The polygon tool draws any shaped polygon and can be spline smoothed.
• The simple shape tool draws either oval shapes or rectangles. Polygons, rectangles, and ovals (simple
shapes) can be filled and raised to give a three-dimensional appearance.
Note: The annotate tool, graphics tools, and the scroller are selected only for a single click and revert to
the standard arrow cursor when you release the mouse button. Use Shift-click to cause the them to per-
sist.
This section covers all the View commands, which differ between Microsoft Window and the Macin-
tosh.
The JMP Starter command opens the JMP Starter window, discussed in the previous chapter, “Start-
ing JMP,” p. 7. When you open the JMP application, the first thing you see is the main menu bar and
the JMP Starter window. If you open a JMP file to start JMP, that file displays as the active window
along with the JMP Starter. You can close the JMP Starter with File > Close when it is the active win-
dow or by clicking its close box. Use the View > JMP Starter to reopen it. Note that on the Macintosh,
the JMP Starter command appears in the Window menu and can be closed by File > Quit when it is
the active window.
66 3 The Menu Bar—The View Menu
(Windows only)
The Log command displays a pane at the bottom of the JMP window that monitors JSL statements
(JSL scripts) as they execute. The log window is editable. For example, can use the log to locate errors in
the script, correct the code, copy it and paste it back into the script window, submit the corrected script
and there is a JSL error.
Note: On the Macintosh, the log window only appears after JSL statements have been run.
(Windows)
There are tool bars for most main menus, and for the colors and markers palettes.The Show Toolbars
command displays a window that lists all available toolbars with a check box to show or hide them. The
Edit > Customize > Toolbars lets you change and rearrange the tool icons that show on each toolbar.
See the chapter “Report Windows and Surface Features,” p. 213, for a complete discussion of toolbars.
(Windows)
The Status Bar command turns the Windows status bar on or off at the bottom window edge.
(Macintosh)
If the toolbar or palette is visible, the Hide Toolbar command hides the toolbar. Likewise, the Show
Toolbar command shows the command icons as a toolbar.
The Unanchor Toolbar command unanchors the visible anchored toolbar so that it is a floating win-
dow with a close box. Likewise, if the toolbar shows as a floating window, the Anchor Toolbar anchors
it to the top of the JMP window. You can also anchor or unanchor the toolbar by simple dragging it to
or from the top of the JMP window.
3 The Menu Bar—The View Menu 67
3 Menu Bar
Initially, the palette contains a default set of command icons, but you can add or delete any command
for which there is an icon by using the Tool Elements Window. The Customize Toolbar window com-
mand displays an alphabetical interactive list of main menu items that you can arrange any way you
want on the tool palette. The example in Figure 3.7 shows the initial tool palette and a modified tool
palette.
Figure 3.7 Macintosh Tool Elements Window and Toolbars
When the Tool Elements window is open, you can delete tool icons from the palette, by simply drag-
ging them off the palette. Add elements from the Tool Elements Window by dragging them from the
window to the tool palette. Rearrange icons on the palette by dragging them to the position you want.
The New Data View command displays a duplicate view of an open data table. The new view is linked
to the original view and all corresponding analysis windows. Changes made to a new view reflect on the
original view when it is made active.
Note: On Windows, the New Data View command is found on the Window menu.
(Macintosh)
The Redraw Window command redraws the active window. It is useful for cleaning up both data table
views and graphical displays that have accumulated stray imperfections resulting from high-speed,
dynamic handling of windows.
Note: The equivalent command on Windows is Redraw, found on the Window menu.
68 3 The Menu Bar—The Window Menu
Macintosh OS 9 Macintosh OS X
The Window menu helps you organize the windows produced during a JMP session. Window menu
items differ as noted below, depending on whether you use a Windows or Macintosh operating system.
(Macintosh)
On the Macintosh, the first two Window menu commands use conventional Mac commands to regu-
late the window size. On Macintosh OS 9 and earlier, Minimize uses the ‘windowshade’ feature to roll
the window into its title bar. On Macintosh OS 10, Minimize shrinks the window into an icon on the
dock. To unroll or expand the window, select Minimize again. The Zoom command enlarges the win-
dow to use the available monitor screen. To reduce the window size, select Zoom again.
The New Data View command displays a duplicate view of an open data table. The new view is linked
to the original view and all corresponding analysis windows. Changes made to a new view reflect on the
original view when it is made active.
Note: On the Macintosh, the New Data View command is found on the View menu.
3 The Menu Bar—The Window Menu 69
3 Menu Bar
(Windows only)
The Close All of Same Type closes all windows that are the same type as the active (front) window. For
example, if the active window is a data table, the Macintosh command appears as Close All Data
Windows, and closes only data table windows. The Windows Close All command closes all open win-
dows.
Note: On the Macintosh, the Close command on the File menu closes the active window. Option-File
displays the Close all Type Window command, which closes all window of the same type as the active
window. Macintosh OS 10 only closes all Data windows (instead of all of any specific type window).
(Windows only)
The Arrange command helps you organize the open windows within JMP:
Cascade
arranges open windows so that the title bar of each window is visible.
Tile
arranges open windows side by side so all of them are visible. Tile Horizontally stacks the win-
dows and Tile Vertically arranges the windows side by side.
Arrange Icons
(Windows) arranges all program-item icons for a selected group into rows. Or, if a group icon is
selected, Arrange Icons arranges all group icons into rows.
The Redraw command redraws the active window. It is useful for cleaning up both data table views
and graphical displays that have accumulated stray imperfections resulting from high-speed, dynamic
handling of windows.
Note: The equivalent command on the Macintosh is Redraw Window, found on the View menu.
(Windows only)
The Move To Back command moves the active window behind all other windows generated by the
current JMP session, leaving the next window in the sequence showing.
70 3 The Menu Bar—The Window Menu
Under Microsoft Windows, the Set Title command lets you change the name of any active JMP win-
dow. This is especially useful if you generate multiple untitled windows and need to distinguish among
them during the JMP session
Note: On the Macintosh, use Control-click on any disclosure icon next to a window title to see report
disclosure commands. These commands include DisplayBox > Set ReportTitle, which performs the
same function as the Set Title command in the Window menu under Microsoft Windows. This Set
Report Title command is available to both the Windows (right-mouse click on a report disclosure icon)
and Macintosh operating systems, but the Windows > Set Title command is Windows only.The report
disclosure commands are discussed in Chapter 9, Report Windows and Surface Features, p. 216.
The Hide command suppresses the display of the active window but does not close it. Under Windows,
to reshow a hidden window, select it from the list displayed by the Unhide command. Choosing All
from the Unhide menu displays all JMP windows. On the Macintosh, select the window from the list
of open windows at the bottom of the Window menu.
Note: Under Microsoft Windows, hidden windows do not appear on the list of open JMP windows
listed at the end of the Window menu.
Windows only
The List All command displays the dialog shown above for you to select the window you want to be the
active window. This command is especially useful if you have a large number of open windows, where
the windows might be data tables, analysis or graphics results, journals and layouts.
3 The Menu Bar—The Help Menu 71
3 Menu Bar
Under Microsoft Windows, the last section of the Window menu lists all open (unhidden) windows.
On the Macintosh, this list includes all windows (both hidden and unhidden). You can bring any win-
dow to the front by clicking on it in this list.
Note that on the Macintosh, the JMP Starter window is accessed from the list of windows at the bot-
tom of the Window menu.
The Windows version of JMP has the standard Windows Help menu. Windows users will find the
menu items similar to those in other applications. For details about the Windows Help menu in gen-
eral, consult your Microsoft Windows User’s Guide. Macintosh users should consult the Macintosh
User’s Guide for the Macintosh OS 8 or later operating system.
The Contents, Search, and Index commands access the JMP Help system. The Help system is con-
structed according to facility provided by your operating system, and also provides navigable online
JMP documentation.
Macintosh OS X
About JMP displays a panel that shows the release, the copyright, the operating system, and the owner
of the copy of JMP that is running. Note that the About JMP and the Preferences commands show
under the application (JMP) menu on Macintosh OS X.
72 3 The Menu Bar—The Layout Menu
On Macintosh, the Help Center command displays a list of Help Books. The JMP Help book accesses
the JMP documentation, with topics and a search facility. Show Balloons shows the standard Macin-
tosh balloon help. The JMP Help command accesses the JMP Help system directly without showing
the list of Help books.
When you choose the Layout command from the Edit menu to create a Layout window, a modified
main menu appears with the Layout menu showing between Edit and Tables. This menu has com-
mands similar to those found in many draw packages. The JMP report in the Layout window is com-
posed of report elements grouped together forming a single object. Each title, plot, chart, axis, and
table item in the analysis report can be ungrouped into separate objects; any subset of report elements
can be grouped into an object. The Layout menu commands operate on selected objects. Click on a
Layout object to select it. Use Shift-click to select multiple objects. You can also drag the cursor to form
a stretch rectangle that selects everything it encompasses.
The Layout window begins with all report elements grouped into a single object. When you click on
the Layout the group highlights and is selected. The Ungroup command ungroups selected objects.
The first time you use Ungroup, the title of the analysis becomes an object, and is the outermost out-
line level of the analysis. The remaining nesting levels become an object. You can continue to select any
portion of Layout window and use the Ungroup command repetitively to ungroup the analysis into its
smallest elements.
3 The Menu Bar—The Layout Menu 73
3 Menu Bar
The Group command combines selected objects into a single object or group; a group can contain
other groups.
The Align command aligns selected objects according to the type of alignment you select from the
Align submenu.You can select as many objects as you want for alignment. However, objects overlap if
they are positioned on the layout window in such a way as to conflict with the alignment request.
Layout objects can be moved to any position in the Layout window, which means objects can overlap.
When objects overlap, the front-most object hides some or all of the objects beneath it. The Clean Up
command moves selected objects such that they are in separate positions in the layout window, and are
all showing.
The Move commands work the same as they do in other draw packages. Each object has a layer relative
to other objects. The Move Forward and Move Backward commands change the layer level of the
selected objects forward or backward one layer at a time. Move to Front moves the selected objects to
lay on top of all other objects. Likewise, Move to Back moves the selected objects to lie beneath all
other objects. Note that when objects are overlaid, the front most objects can hide objects beneath
them.
4
4 JMP Data
Characteristics of Data
The columns of a JMP table can contain different kinds of information. However, all information in a
single column must be of the same type. You create a new column with the New Column command,
and define its characteristics with the New Column dialog.
JMP correctly analyzes data according to column characteristics. This chapter discusses the following
attributes of data, usually defined when the column is created, and by commands in the Rows and
Cols menus:
• data types and modeling types
• analysis roles
• data storage and formats
• a special format for date and time values
• properties assigned to columns that affect the way data behave.
4
Contents
JMP Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Types of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Modeling Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Analysis Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Freq Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Weight Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Numeric Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Date, Time and Date-Time Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Row States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Row State Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Selection Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Data Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Specialized Properties of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4 Characteristics of Data—JMP Data 77
4 Data
JMP Data
The columns of a JMP table can contain different kinds of information. However, all information in a
single column must be of the same type. You create a new column with the New Column command,
and define its characteristics with the New Column dialog.
JMP analyzes data and presents results according to column characteristics. This chapter discusses the
following attributes of data:
• data types and modeling types
• the roles data play in an analysis
• data storage and formats
• a special format for date and time values
• computed values
• row states
• specialized properties assigned to columns that affect the way data behave.
Types of Data
The New Column dialog lets you give JMP a description of a column’s data characteristics by assigning
the column one of three data types and one of three modeling types. Figure 4.1 shows a New Column
dialog, with menus for selecting a data type and a modeling type. These characteristics are described in
detail in the following sections.
Figure 4.1 New Column Dialog
Data Types
The first characteristic assigned to a column is its data type. The data type of a column determines how
its values are formatted in the data grid, how they are stored internally, and whether they can be used in
calculations. The three data types are numeric, character, and row state, as shown in Figure 4.1.
You designate a data type for each column with the Data Type popup menu in the New Column dialog
or change the data type using the Column Info dialog.
78 4 Characteristics of Data—Types of Data
You can change a column’s data type as long as the column values are as follows:
• Numeric columns must only contain numbers, with or without a decimal point. A numeric column
can have any modeling type. Numeric column values display right-aligned and can be formatted
with the Format popup menu. You can change a numeric column to character or row state with the
Data Type popup menu.
• Character columns can contain any characters including numbers. In character columns, numbers
are seen as characters only and are treated as discrete values instead of continuous values. You can
change a character column to numeric. Any values that are numbers convert correctly. Character
values become missing.
• Row State columns contain special information that can affect the appearance of graphical displays.
A row state column does not have a modeling type because its values are not used in analyses. See
Row States, p. 84 later in this chapter, for details about using row state information.
Note: You cannot undo the change of a modeling type.
Modeling Types
Numeric and character columns must be assigned one of the three modeling types listed by the
Modeling Type popup menu on the New Column or on the Column Info dialog. The Modeling Type
assigned to a column tells JMP how to treat its values in analyses. The icon to the left of the column
name in the Columns panel also accesses the modeling type menu. Changing the modeling type lets
you look at a variable in different ways in an analysis.
You can change a column’s modeling type as long as the assigned modeling type corresponds to the data
types as follows:
• Continuous columns must contain numeric values and be assigned the Numeric data type. Contin-
uous values are treated as continuous measurement values. JMP uses the numeric values directly in
computations.
• Ordinal columns can have either Numeric or Character data types. JMP analyses treat ordinal val-
ues as discrete categorical values that have an order. If the values are numbers, the order is the
numeric magnitude. If the values are character, the order is the sorting sequence.
• Nominal columns can have either Numeric or Character data types. All values are treated in JMP
analyses as though they are discrete values with no implicit order.
The table to the right in Figure 4.2 illustrates possible combinations of data types and modeling types.
4 Characteristics of Data—Analysis Roles 79
4 Data
Figure 4.2 Modeling Types
Analysis Roles
To analyze data, you must cast columns into roles appropriate for the analysis you want to perform.
The following roles are the most typical.
• X identifies a column as an independent, classification, or explanatory variable whose values divide
the rows into sample groups.
• Y identifies a column as a response or dependent variable whose distribution is to be studied.
• Weight identifies a numeric column whose values supply weights for each response.
• Freq identifies a numeric column whose values assign a frequency to each row for the analysis.
Assigning roles is usually done in the launch dialogs presented by each Analysis or Graph menu com-
mand. The launch dialog uses the assigned roles only for a single analysis. However, there is a Recall
button on most launch dialogs that recasts the variables as they were in the previous analysis that used
that dialog.
The names of the role buttons on the different platform launch dialogs vary slightly to convey informa-
tion specific to the analysis. Figure 4.3 shows the launch dialogs for a Distribution analysis (top dialog)
and a Recurrence Survival analysis (second dialog). Both analyses require a response (Y) column, but
the recurrence response is denoted as Y, Time to Event to indicate that the response for this kind of
analysis is always a time measurement.
All platforms also support By variables entered on the Launch dialog. When there is a By variable, the
analysis is performed for each level of the By variable, or combination of levels if there is more than one
By variable.
80 4 Characteristics of Data—Analysis Roles
You can assign the most common analysis roles with the Preselect Roles command on the Cols
menu. Multiple columns can have preselected roles. Preselected roles remain in effect for all analyses
until you change them.
When you select an Analyze or Graph command and there are sufficient columns with preselected
roles, the analysis completes without the intermediate step of showing the launch dialog. If an analysis
requires more columns than have been assigned roles, the launch dialog shows the preselected roles and
prompts you for additional role information.
Often, data are collected and arranged in a data table with columns that always serve as Weight or Freq
variables. Preselected roles are especially useful for these kinds of columns because their values are an
integral component of the data and are needed for most analyses.
Freq Variables
The Freq role identifies a variable whose values are frequencies for each row. Specifically, if f is the value
of the Freq variable for a given row, then that row is used in computations f times. When there is no
Freq column, each row contributes the implied value of one (1) to the frequency count. The total
number of rows with nonmissing values is the denominator of the mean computation, and is used in
determining degrees of freedom for tests of hypotheses.
An analysis that uses a frequency variable reflects an expanded number of rows; means and degrees of
freedom reflect this expanded table. You could produce the same analysis by first creating a new data
table that contains the number of observations expanded by the value of the Freq variable for each row.
4 Characteristics of Data—Numeric Formats 81
4 Data
For example, if the value of Freq variable is 5 for the first row, then the first 5 rows in the new expanded
table would be copies of the first row of the original table.
You usually think of a frequency value as being an integer greater than 1. However, there are situations
where its value has a fractional part, or may be between zero and one. If the value is not an integer, only
the floor of the value (the largest integer smaller than the value) is used. If the frequency value is missing
or is less than one, the floor of its value is zero and that observation is not used in the analyses.
Note: A negative frequency value causes an error message. If all frequencies are less than one, the fre-
quencies for computations are all zero and generate an error condition.
Weight Variables
You assign a column the Weight role when you want the analysis to use relative weights for each row.
The response in each row is multiplied by its Weight variable for all analyses. The sum of the weights is
used in statistical computations; however, the weight variable does not alter the degrees of freedom.
One common use for a Weight variable is to compute weighted product-moment (Pearson) correlation
coefficients. A weight variable with values proportional to the reciprocals of the variances is sometimes
used when the variance associated with each row is different.
Unlike the Freq variable, the values of a Weight variable can be either integer or noninteger, positive or
negative.
Numeric Formats
A numeric variable has a format associated with it. Formats for numeric column values are listed in the
Format popup menu found in the New Column and Column Info dialogs, as shown in Figure 4.4.
Note that there is an editable Width associated with numeric formats. Formats appear with a default
width that you might need to increase to display large numbers.
Figure 4.4 Column Info Dialog and Date-Time Formats
The following numeric formats tell JMP how to display numbers in the data grid:
• The Best format means JMP considers the precision of each cell value and chooses the best way to
show it. By default, the physical width of the column is 10 characters. You can increase the column
width in the Column Info dialog if you need to show large values with many decimal places. If a col-
82 4 Characteristics of Data—Date, Time and Date-Time Formats
umn has computed values, the Best format often leads to messy-looking numbers. To show all val-
ues with the same number of decimal points, select Fixed Decimal as the numeric format.
• Fixed Dec displays all column values rounded to the number of decimal places you specify. You can
set the number of decimal places to zero to see only whole numbers.
• PValue is a specialized format that best displays probability values. It has a default width of 6 and
shows 4 decimal places. If a number is less than 0.0001, it displays with a less than sign, <.0001. The
format is mostly used in JSL programming scripts and rarely needed for a data table column.
• Scientific displays a number in standard scientific notation. If you enter the number 123456, it dis-
plays as 1.23456e+5
• Currency formats values with two decimal positions, commas as needed, and the dollar sign ($) pre-
fix. If you enter the number 123456, it shows in the data table as $123,456.00. Entering the num-
ber 1234.56 formats as $1,234.56. The default width of the Currency format is 10. If you have a
number that requires a wider field width, the format defaults to the Best format. Use the Column
Info dialog to increase the column width sufficiently to display your data.
• Date and Time formats are a special numeric format, and are described in the next section.
4 Data
Figure 4.5 Date, Time, and Datetime Formats
The following examples show formats for reading or displaying the date, December 31, 2002. Its
unformatted (numeric) value is 3,124,137,600.
• m/d/y has the form mm/dd/yyyy, giving 12/31/2002.
• mmddyyyy has the form mmddyyyy, giving 12312002.
• m/y has the form mm/yyyy, giving 12/2002.
• d/m/y has the form dd/mm/yyyy, giving 31/12/2002.
• ddmmyyyy has the form ddmmyyyy, giving 31122002.
• ddMonyyyy has the form ddMonyyyy, giving 31Dec2002.
• Monddyyyy has the form ddMonyyyy, giving Dec312002.
• y/m/d has the form yyyy/mm/dd, giving 2002/12/31.
• yyyymmdd has the form yyyymmdd, giving 20021231.
• Date Long and Date Abbrev are for display only.
Date Long shows a date value as weekday, month day, year, which gives Tuesday, December 31,
2002.
Date Abbrev is the same as the Long format except that weekday and month have three-character
abbreviations, giving Tue, Dec 31, 2002.
You might need to increase the column width to see the entire formatted value for Date Long and
Date Abbrev date formats, and for date-time formats. Edit the Width text box on the New Col-
umn dialog to specify the field width needed to accommodate the largest number of digits or char-
acters you think the new column could have. The maximum field width is 40 for numeric values
and 255 for character values.
• m/d/y h:m, m/d/y h:m:s, d/m/y h:m, d/m/y h:m:s, y/m/d h:m, y/m/d h:m:s, ddMonyyyy h:m,
ddMonyyyyh:m:s, Monddyyyy h:m, and Monddyyyy h:m:s, show the number of hours, minutes,
and seconds after midnight of the prepended date. The date displays in the formats described above.
84 4 Characteristics of Data—Row States
The hours, minutes, and seconds are separated by a colon. This example (12/31/2002) has no hours
or minutes. Examples of its formatted values with zero hours minutes and seconds are 12/31/2002
12:00 AM and 12/31/2002 12:00:00 AM.
• :days:hrs:mins and :days:hrs:mins:secs show the number of days, hours, minutes, and seconds
since January 1, 1904. The results for December 31, 2002 are :36159:00:00 and :36159:00:00:00.
• h:m:s gives hours, minutes, and seconds portion of the date in the date field. The example used here
(December 31, 2002) has no time entered, so the result is 00:00:00.
Selecting a date or time format for a column also lets you enter values using any representation recog-
nized by your machine. When you press Return, JMP stores the date entry internally as the numeric
number of seconds between January 1, 1904 and the date you entered. The date displays according to
the format you assigned to the column.
The JMP Formula Editor offers full support for dates with functions that accept date columns as argu-
ments and return date-related elements such as day of the week, day of the year, week of the year, or
month. See the chapter Formula Editor Functions, p. 163, for a description of date functions.
If you are importing variables from a SAS file, JMP looks for a SAS date format and translates it to a
JMP date column. When you are exporting data to a SAS file, JMP date columns become SAS date val-
ues with the appropriate SAS format.
Row States
Row states are characteristics associated with a row. They can distinguish subsets of your data, exclude
data from analyses, and customize the appearance of graphical displays. Row state commands are in the
Rows menu and affect only highlighted rows. Row state assignments appear in the row number area at
the left of the data grid (see Figure 4.6).
4 Characteristics of Data—Row States 85
4 Data
Row State Conditions
JMP uses the following row states:
Exclude/Unexclude
is a toggle that excludes selected rows from statistical analyses. To do this, select the rows that con-
tain unwanted values and choose Exclude/Unexclude from the Rows menu. Data remain
excluded until you choose Exclude/Unexclude again for those selected rows.
Warning: Excluded data are not automatically hidden in plots even though they are excluded from
calculations in text reports and graphical displays.
Hide/Unhide
is a toggle that suppresses the display of points in all scatterplots. To hide data, select rows in the
data grid and choose Hide/Unhide from the Rows menu. For example, you can exclude points
from analysis and then hide those same points in scatterplots. The data remain hidden until you
choose Hide/Unhide again for selected hidden rows.
Warning: Hidden points are not automatically excluded from statistical computations that affect
text reports and graphical displays even though they are not displayed in the plots. To exclude hid-
den observations from analyses, you must also use assign them the Exclude/Unexclude row state.
Label/Unlabel
a toggle that labels points on all scatterplots. To label points, select the rows containing the points
and use the Label/Unlabel command. This row state remains in effect until you choose Label/
Unlabel again for those selected rows.
Note: There is also a Label/Unlabel command in the Cols menu. JMP uses the row number as
the label value on scatterplots if you don’t assign a label column. However, if you assign a label col-
umn, its values appear on plots for labeled rows. If there are multiple label columns, their values
appear on plots separated by a comma.
Colors
lets you assign any colors to highlighted rows. The points in scatterplots and spinning plots display
using the color you select from the Colors palette. The active color assigned to a row displays next
to the row number in the data grid.
Markers
assigns a character from the JMP markers palette to replace the default dot in scatterplots and
spinning plots. Each active marker displays next to its row number in the data grid, as shown in
Figure 4.6.
or activate a row state condition saved in a column, use the Copy to Row States or Add to Row
States command (Figure 4.6).
Copy to Row State
copies the row state characteristics from the row state column to the row number area, which
makes them active. Copy to Row States replaces existing active row states.
Add to Row States
adds the row state characteristics from the row state column to the row number area, which makes
them. Add to Row States preserves existing active row states.
Copy from Row States
copies the active row state characteristics from the row number area to the row state column.
Copy from Row States replaces existing row state values in the column.
Add from Row States
adds the active row state characteristics from the row number area to the row state column. Add
from Row States preserves existing row state values in the column.
Figure 4.6 Row States Columns and Row State Conditions
Note: To add a row state to a row state column for every row, context-click (Control-click on the
Macintosh) to see Rows menu commands, and select from the list of row states.
4 Characteristics of Data—Row States 87
4 Data
Selection Status
Whether or not a row is selected is also a row state condition. Like other row state characteristics, selec-
tion status can be saved in a row state column. In the data grid, the row number area of a selected row
is highlighted. If you use Copy from Row States, the cell of the row state column also appears high-
lighted.
The Row Selection command in the Rows menu is useful for selecting all rows or a subset of rows
with specific row state characteristics.
Note that the Invert Row Selection command deselects all selected rows, and selects all previously
deselected rows. The Select All Rows command selects all the rows in a JMP data table.
You can select rows with the excluded, hidden, or labeled row state characteristic using the following
submenu items:
Select Excluded
selects all excluded rows regardless of their current selection status and deselects any other previ-
ously selected rows.
Select Hidden
selects all hidden rows regardless of their current selection status and deselects any other previously
selected rows.
Select Labeled
selects all labeled rows regardless of their current selection status and deselects any other previously
selected rows.
Select Where
displays the dialog on the left in Figure 4.7. To use the Select Where command, highlight a vari-
able in the variable list on the dialog, choose a condition from the list of conditions (equals, does
not equal, and so forth), enter a comparison value in the text box, and click OK.
88 4 Characteristics of Data—Data Validation
Select Randomly
displays the dialog on the right in Figure 4.7 to enter the absolute sample size or proportional sam-
ple size you want in the text edit box.
Figure 4.7 Select Where and Select Randomly Dialogs
See Selecting and Deselecting Rows and Columns, p. 101 in the “JMP Data Tables” chapter for
more details about selecting rows.
Data Validation
The List Check and Range Check Validation column properties the New Column dialog display the
dialogs shown in Figure 4.8. You complete a dialog by entering acceptable values or an acceptable range
of values for a column. For a single-sided range check, omit one of the bounds (upper or lower).
(list check)
The cursor changes to the list check form when positioned over any editable area (cell or column name)
in a column with list check data validation checking in effect. When you click a cell, a blinking vertical
bar appears for entering or editing text as usual. However, you can only enter values included on the
validation list. A beep and a message warn if you try to enter invalid text. If you right-click a cell (Con-
trol-click on the Macintosh), a popup menu of acceptable values shows. This lets you select the cell
value from this menu instead of typing it into the cell.
(range check)
The cursor changes to a large I-beam on numeric column values with a specified range and range con-
dition. If you try to enter a value outside the validation range, a dialog asks if you want to leave the
value as is or change it to a missing value.
4 Characteristics of Data—Specialized Properties of Data 89
4 Data
Figure 4.8 List Check and Range Check Dialogs
When you select a property from the New Property list, a panel specific to that property appears, and
the property is then listed in the Current Properties list. You can select any property to see its dialog and
change information about it. To delete a property, select in the list of properties and click Remove but-
ton. Column properties can be added or deleted at any time.
90 4 Characteristics of Data—Specialized Properties of Data
Formula
gives the column access to the JMP Formula Editor. To display the Formula Editor, click the Edit
Formula button that now appears on the Formula panel in the New Column dialog. To modify a
column’s formula after the column has been created, highlight the column in the data grid and
select Formula from the Cols menu.
The Formula column property locks the column so that values cannot be edited in the data table.
Notes
gives a text edit box for you to enter documentation for the column, such as the source of the data
or any other helpful notes.
Range Check and List Check
display the dialogs shown in the previous section, Data Validation, p. 88, for constraining the val-
ues that can be entered into a column.
Axis
displays the same panel as you see when you context-click on the axis of a graph. It lets you specify
minimum, maximum, increment, number of tick marks, and so on for these axes. These values
become the default whenever the column is used in an analysis.
Coding
has edit boxes to enter high and low values. When you fit the column in a model, the low and high
data values are transformed into –1 and 1, which makes tests and parameter estimates more mean-
ingful. Coding can be used for any continuous variable, and is the default for continuous factors
generated by the DOE facility in JMP.
Mixture
is a property assigned to a column if it is one of several factors that form 100% of a mixture. When
you assign the Mixture property to a set of variables, a no-intercept model is automatically gener-
ated by the Fit Model facility when you analyze the data.
Row order levels
has a check box to indicate that rows be ordered by the order they are found in the data table
instead of ordered by value.
Spec Limits
creates a tab with edit fields to enter Lower Spec Limit, Upper Spec Limit, and Target, as you do
when you request a capability analysis in the Distribution platform. These limits are stored with
the column and automatically used in analyses.
Control Limits
lets you enter values for average (Avg), lower confidence level (LCL), and upper confidence level
(UCL) to be saved with the column and used automatically by the Control Charts platform. Also,
a popup menu on this property panel lets you select the type of control chart to which the values
apply.
4 Characteristics of Data—Specialized Properties of Data 91
4 Data
Response Limits
gives edit fields to enter Lower, Middle, and Upper limits, and Desirability values for a response
column in an experimental design. There is also a popup menu listing the selections Maximize,
Match Target, Minimize, and None, which are possible goals for a DOE response variable.
Design Role
provides a popup menu with selections that tell how a factor column is to be used in a model to
design an experiment. Factor roles are Continuous, Categorical, Blocking, Covariate, Mixture,
Constant, Signal, and Noise.
Sigma
lets you enter a known sigma to be saved with the column. It is used by applications such as con-
trol charts that require a sigma value to complete computations. If no value is supplied, Sigma is
calculated from the sample.
Units
gives a text edit box for entering measurement units. For example, you might want to indicate that
age values are measured in months or that a monetary value is in thousands of dollars.
Other
is for user-defined properties. You assign any name you want to the property, and enter the value
for the property. This property is then available for JSL programming.
5
5 Data Tables
JMP Data Tables
Create and navigate
JMP data are organized in memory as rows and columns of a table referred to as the data table. The col-
umns have names and the rows are numbered. An open data table is kept in memory, and you commu-
nicate with it through an active window. You can open as many data tables in a JMP session as memory
allows.
Commands in the File, Edit, Tables, Rows, and Cols menus give you a broad range of data handling
operations and file management tasks such as flexible data entry, data validation, text editing, and
extensive table manipulation. This chapter describes the elements of a JMP data table, tells how to cre-
ate a new data table, and shows how to navigate about on the data table window. The chapter “The
Tables Menu,” p. 119, covers each Tables menu command in detail.
5
Contents
Elements of a JMP Data Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Table Variables and Table Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
New Table Variable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
New Property/Script . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
The Columns Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
The Rows Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
The Data Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Cursor Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Selecting and Deselecting Rows and Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Creating a New Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Adding and Deleting Rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Adding and Deleting Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Filling a Data Table with Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Entering and Editing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Finding and Replacing Cell Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
Importing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Exporting Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Accessing a Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Opening a JMP File from a Web Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Cut and Paste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Drag and Drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5 JMP Data Tables—Elements of a JMP Data Table 95
5 Data Tables
Elements of a JMP Data Table
JMP data are organized in memory as rows and columns of a table referred to as the data table. The data
table opens as a standard window. The following is a list of some important data table features:
• Data can be manually keyed in, calculated, imported from a text file, pasted into a JMP table, or
read directly from SAS data sets, Excel files, and other file formats. Under Windows, data can be
received as real-time external measurements.
• Column names use any keyboard character including spaces. The size and font for names and values
is a preference setting that you control. Column names automatically wrap in the column name area
to accommodate the column width you specify.
• Character fields can be up to 255 characters long. You can move column boundaries and enlarge the
column to view long values.
• Commands from the Rows and Cols menus add, move, hide, and delete rows and columns.
• There is no limit to the number of rows or columns in a data table. However, the table must fit in
memory.
You can start JMP by opening an existing JMP data table, which displays the data in the form shown in
Figure 5.1. The table has two parts: the data table panels and the spreadsheet view of the data, called
the data grid.
Figure 5.1 Active Areas of JMP Data Table Panels and Data Grid
Data Table Panels show or hide Cols menu click to
data grid Data Grid select or ws
Tables menus and or columns
table pro perties Table Panel Rows
Menu
Cols menu
Columns Panel
movable
column
Rows menu boundaries
rows state Rows Panel
information
There are three data table panels—the Table panel, the Columns panel, and the Rows panel. The data
table panels are arranged to the left of the data grid as shown in Figure 5.1. These panels contain all the
pertinent information about the table and its contents. Each has a popup menu with commands that
perform most table management operations, column actions, and row actions.
The panels are divided by horizontal splitters, which you drag up or down to lengthen or shorten the
panels. Likewise, a movable vertical splitter divides the table panels from the data table itself. You can
click the diamond in the upper left of a panel to show or hide the data table panels.
96 5 JMP Data Tables—Table Variables and Table Properties
New Property/Script
The New Property/Script command lets you save a JSL script with a data table. This command dis-
plays a dialog to name the script and space to type in JSL commands.
5 JMP Data Tables—The Columns Panel 97
5 Data Tables
As a simple example, suppose you create a new table property in a table that has height and weight as
columns (the Big Class sample data table). Figure 5.3 shows a Property/Script namedBivariate that
runs the Fit Y by X command with weight as Y and height as X, and then fits a line to the result. The
Table Panel shown in Figure 5.3 displays the new Table Property, which runs the bivariate script when
you click its icon and select the Run Script command. Notice in the table panel that the data table also
has a Table Variable called Notes.
Table Properties are mostly used to create permanent JSL scripts for a table, but they are available to the
Formula Editor and could be used in formulas.
Double-click on a Table Property to edit its name or content. Click on the triangle next to the Table
Property name to see a menu with commands to Run Script, Edit or Delete the Table Property.
Figure 5.3 Create a New Table Property to Run a JSL Script
Modeling types are described in the chapter “Characteristics of Data,” p. 75. The Cols menu com-
mands are covered in the chapter “The Menu Bar,” p. 25.
5 Data Tables
Figure 5.6 Rows Panel Information
Cursor Forms
To navigate the data grid, you need to understand how the cursor works. The cursor has different
forms, and the actions it performs depend on its location in the data grid:
Arrow Cursor
The cursor is the standard arrow when it is in the panels area to the left of the data table, in the triangu-
lar rows and columns area in the upper-left corner of the data grid, or on the title bar of the Tables
panel.
• To select a column, click once in the area above the column or on its name in the Columns panel.
• Click the table name on the title bar of the Tables panel to edit it.
• Click in the column name area to edit a name, or double-click in the column name area to see the
Column Info dialog.
• Click in the triangular areas in the upper-left corner of the data grid to deselect rows and columns.
100 5 JMP Data Tables—The Data Grid
When the cursor is within a column heading or a row number area it becomes a large plus, indicating it
is available to select rows or columns. When you click the large plus cursor, that row or column high-
lights. Drag to highlight multiple rows or columns, and Control-click (Command-click on the Macin-
tosh) to select discontiguous rows or columns.
• Double-click in a column heading area to see the Column Info dialog for that column.
• Click in a column name area at the top of a column to edit the name. The cursor becomes an
I-beam.
• Double-click in the row number area to see the Row Editor for editing values.
The cursor then becomes an I-beam and you can edit the text if it is not in a locked column or locked
data table.
You can also edit selected column names in the Columns panel to the left of the data grid.
I-beam Cursor
When you select (click) editable text, the cursor appears as the I-Beam. To edit text, position the
I-beam within highlighted text. Click the I-beam to mark an insertion point, or drag it to select text for
replacement. The I-beam deposits a blinking vertical bar to indicate a text insertion point or a high-
lighted area of text to be replaced. The default selection is the entire cell. Use the keyboard to make
changes.
The cursor changes to a double arrow when it is on a column boundary. Drag this cursor left or right to
change the width of a column. However, this physical width does not affect the column field width
specified in the Column Info dialog.
Note: You can adjust all widths of selected column at once with Control key (Option key on the
Macintosh) as you drag the double arrow cursor on any of the selected column boundaries.
The cursor changes form when you move the mouse over values in columns that have data validation in
effect. It becomes a small, downward-pointing arrow on a column with list checking, and a large
I-beam on a column with range checking. When you click, the value highlights and the cursor becomes
the standard I-beam; you enter or edit data as usual with any values defined as valid text or valid num-
bers. List Check and Range Check values are defined in the Column Info dialog of a column.
Popup pointer cursor
5 JMP Data Tables—The Data Grid 101
5 Data Tables
The cursor changes to a pointer over any popup menu icon or diamond-shaped disclosure icon. Click
and drag to select a menu icon and to open or close a window panel.
selected column
selected rows
To extend a selection of rows or columns, drag the cursor across the array or Shift-click the first and last
rows or columns of the desired range. To make a discontiguous selection, use Control-click (Com-
mand-click on the Macintosh) on the desired selections. These commands are also toggles that deselect
already selected rows or columns.
Note: To select all rows or all columns at once, Shift-click in the triangular areas in the upper-left corner
of the data grid. Click in the upper-left corner to deselect all rows or columns.
There are other ways to highlight specific rows:
Summary Tables
You can use the Summary command in the Tables menu to select subsets of rows. Summary creates a
JMP window that contains a summary table. This table summarizes columns from the active data table,
called its source table. It has a single row for each level of a grouping variable you specify. Each row in
the summary table identifies the corresponding subset of rows in the source table. When you highlight
a summary table row, all corresponding rows highlight in its source table. See the chapter “The Tables
Menu,” p. 119, for more details about the Summary command.
5 Data Tables
comparison operation from the Select Where dialog popup menu, and a selection criterion value.
When you click OK, Select Where highlights all rows that meet the search criterion.
The Select Where command looks only at text strings. Before it completes the comparison,
Where converts a numeric value to the character string as it appears in the table cell. The Select
Where popup menu lists the comparison operators.
Figure 5.9 Select Rows Using the Select Where Dialog
Select Randomly
randomly selects the number of rows you enter into a dialog. If the number entered is greater than
1, that many rows are selected. If the number entered is a fraction, JMP selects that fraction of the
total rows. For example, entering 10 randomly selects 10 rows. Entering 0.1 randomly selects 10%
of the rows.
5 Data Tables
By default, the new column names are Column 1, Column 2, and so forth. However, you can specify
other text that becomes the prefix of the new column names.
Using the Add Multiple Columns command to define multiple columns is different than using the
New Column command because all columns you request with the Add Multiple Columns dialog have
the same data characteristics. The Add Multiple Columns dialog creates the number of new columns
you request with the characteristics you enter.
See the chapter “Characteristics of Data,” p. 75, for a description of each Column Property.
Figure 5.10 The New Column Dialog (left) and Add Multiple Columns Dialog (right)
5 Data Tables
Figure 5.12 The Search Dialog
replaces the currently highlighted cell value with the contents of the Replace with box (the replace
value). If the replace value is missing and you select Replace or use Control-H (Command-= on
the Macintosh), the currently highlighted cell content becomes a missing value.
Replace & Find Next
functions the same as Replace, but continues to search for the find value.
Replace All
replaces all occurrences of the find value with the replace value except in locked columns. You can-
not undo Replace All.
Check boxes on the Search dialog further refine a search:
Match Case
gives case sensitive search, useful for locating proper nouns or other capitalized words.
108 5 JMP Data Tables—Filling a Data Table with Values
Importing Data
The Open command displays a specialized open file dialog that lets you locate the file you want to open
and tell JMP the file format of the incoming file. The Open command then reads the file into a JMP
data table.
JMP directly reads JMP data tables, JMP journal files, JMP Script files, SAS data sets, SAS transport
files, text files with any column delimiter, and Excel files. Also, under Windows, the Database com-
mand (described next) can access any database on your system that has an ODBC driver. This feature is
not currently available for the Macintosh.
Windows Import
If you indicate what kind of file to expect with a Files of type selection appropriate for the file type to
be read, JMP gives helpful information when possible. The example on the right in Figure 5.13, shows
an Open Data File dialog for reading a JMP data table. The table notes (if there are any) also show.
If the incoming file is not a JMP data table and you choose All(*.*) from the Files of type drop-down
list, JMP looks at the type of file given by the 3-character extension appended to its file name and opens
it accordingly. This works as long as the file has the structure indicated by its name.
5 JMP Data Tables—Filling a Data Table with Values 109
5 Data Tables
Figure 5.13 The Open Data File Dialog for Macintosh (left) and Windows (right)
The Text Import selection from the Files of Type drop-down list attempts to discern the arrangement of
text data. This is adequate for a rectangular text file with no missing fields, a consistent field delimiter
and end-of-line delimiter.
Note: If double-quotes are encountered when importing text data, they change the delimiter rules to
look for an end double-quote. Other text delimiters including spaces embedded within the quotes, are
ignored and treated as part of the text string.
To see a preview of an incoming text file, choose Text Import Preview from the Files of Type
drop-down list.
Then, when you click Open, a panel asks whether the data have Delimited fields or have Fixed Width
fields.
If you click Delimited, the data arrangement is described as shown in Figure 5.14. This dialog begins
with settings from your Preferences file. It also shows the columns’ names, data types and the first two
rows of data. In this example, preferences are set that indicate the incoming table contains column
headers to be used as the column names. The column names are species, subject, miles, and season.
If there are no column names the Name fields are called Column 1, Column 2, and so on.
You can also identify one or more end-of-field delimiters, end-of-line delimiters, choose to strip enclos-
ing quotes, and see how many rows and columns will be read.
If you click Fixed Width, you can enter the width of each field and specify column names.
110 5 JMP Data Tables—Filling a Data Table with Values
5 Data Tables
To limit the number of rows to be read, enter the number in the Number of Lines edit box above the
Specify Columns button (see Figure 5.14).
You can also select the columns to read from an existing JMP data table and SAS Transport files. The
Select Columns checkbox shown in Figure 5.15 appears in the File Open dialog for the those file
types.
Figure 5.15 Selecting Rows and Columns for Importing
Import as Data
The Import as Data command on the File menu reads scripts and text windows. Selecting this com-
mand instantly imports the text in the active text window as a JMP data table using the settings from
the Text Import preference. To change the Text Import preferences, select File > Preferences and select
the Text Import/Export tab on the Preferences panel.
Using the Import as Data command is useful for situations where text is pasted into a text window,
then subsequently formatted and adjusted in preparation for import into JMP.
Imported File Sizes
The limitations on the size of imported text files are:
• a limit of 10,000 columns that can be imported per table.
• a limit of 128K per line (row) of data that can be imported.
The size limit of a file to be imported is restricted by the amount of free RAM/Virtual Memory that the
system possesses. For example, JMP doesn’t import 1 Gigabyte file into a system that has a combined
500Meg of RAM/Virtual Memory. For realistic performance, the size of the file to import should be
less than the physical RAM present in the machine.
112 5 JMP Data Tables—Filling a Data Table with Values
Exporting Data
The Save As command writes the active data table to a file after prompting you for a name and file
type. It can save the data table as a JMP file, convert it to a SAS Transport file, save it as a text format
file, or save it in any database format available on your system.
The Save As command has the same options as the Open command. The .TXT file extension and Text
Export from the Save As Type drop-down list displays the Text Export Options dialog in Figure 5.16:
JMP Data Tables
save the table in JMP format. This is the default Save As option.
SAS Transport Files
converts a JMP data table to SAS transport file format and saves it in a SAS transport library.
Append To
appends the data table to an existing SAS transport library. If you don’t use the Append To option,
a new SAS transport library is created using the name and location you give. If you do not specify
a new file name, the SAS transport library replaces the existing JMP data table.
Text Format Files
converts data from a JMP file to standard text format, with the rows and columns. The options
button on the Save As dialog displays choices to describe specific text arrangements:
Export Column Names to Text File has Yes and No radio buttons to request that JMP column
names be written as the first record of the text file, or that no labels or header information be saved
with the data.
End of Field and End of Line designate the characters to identify the end of each field and end of
line in the saved text file. These options are described previously in the Importing Data section.
There is an option to have no delimiter in exported text files, which exports the variables with no
quotation marks or spaces between them.
When saving a SAS V7 file, the following rules apply. Suppose you are saving a JMP data table called
baseball.jmp:
• If there are existing.sd2 or.sd7 files in the directory, JMP saves the file as Baseball.sd7, if V7 format
is specified, and Baseball.sd2 if V6 format is specified.
• If there exists files in the directory with the longer extension (.sas7bdat), then JMP saves the file as
Baseball.sas7bdat if V7 format is specified.
• If there are no SAS data sets in the directory, JMP saves the file as Baseball.sas7bdat if V7 format is
specified.
5 JMP Data Tables—Filling a Data Table with Values 113
5 Data Tables
Figure 5.16 Save As Dialog and Text Formatting Options
Accessing a Database
Under Windows, the Database command opens a connection to any database on your system that has
an ODBC driver. Figure 5.17 shows the dialog initially displayed when you select
File > Database Open Table. This Database Open Table dialog lets you connect to or disconnect from
a specific database source.
114 5 JMP Data Tables—Filling a Data Table with Values
When one or more database connections are made, the list of connections shows in the Connections
list, and a list of tables in the selected database file or directory shows in the Tables list. Optionally, you
can click the Advanced button and write SQL statements to open specific subsets of a table.
To use the database importing facility in JMP, click Connect and follow these steps:
1. When the Select Data Source dialog (see Figure 5.18) appears, click the Machine Data Source tab
to see a list of database types on your system.
2. Select the data source you want from the list of names and click OK. Then identify the file or the
directory of the file you want to import into JMP.
Figure 5.18 Select a Database Source, and Navigate to a File or Directory
3. After you identify a directory or file location and click OK, the Database Open Table dialog shows
the database connection source and available files (see Figure 5.18).
4. Select the database name and the file you want to open and click OK to import the entire table.
If you want to import only selected portions of a table, click the Advanced button. An enhanced Data-
base Open Table dialog appears with a default SQL Select statement appropriate for the selected file.
The example shown in Figure 5.19, reads five variables (name, age, sex, height, and weight) from the
Big Class table stored in an Excel directory.
5 JMP Data Tables—Filling a Data Table with Values 115
5 Data Tables
You could type in the SQL statement yourself, or modify the generated statement with additional SQL
clauses. For example, the first SQL statement below could be modified as shown in the second state-
ment:
SELECT NAME, AGE, SEX, HEIGHT, WEIGHT FROM Big Class
SELECT NAME, AGE FROM Big Class where AGE>14 and SEX=‘F’
Then, when you click Execute SQL, a JMP data table appears with columns Name and Age. The SQL
statement becomes the JMP data table notes.
Figure 5.19 Open a Table, or SQL Selection
5 Data Tables
Paste
pastes information from the clipboard to the selected area in a JMP data table. Paste can be used
with the Copy command to duplicate rows, columns, or any subset of cells defined by selected
rows and columns.
To duplicate an entire row or column,
• select and Copy the row or column to be duplicated
• select an existing row or column to receive the values
• use the Paste command.
Note: You can also copy or move rows by dragging them to a new position in the data table.
To duplicate a subset of values defined by selecting both rows and columns, follow the steps above,
but select the same arrangement of rows and columns to receive the copied values as originally
contained them. If you paste data with fewer rows into a destination with more rows, the source
values recycle until all receiving rows are filled. If you paste more information than space available,
the excess values are lost.
Also, the Paste command extends a JMP data table by adding rows and columns to a data table as
needed to accept values from the clipboard.
If you select rows before choosing this command, the effect is similar to joining data tables by row
number; Paste adds cells to the bottom of the data table and fills them with values from the clip-
board.
To transfer data from another application into a JMP data table, first copy the data to the clip-
board from within the application. Then use the Paste command to copy the values to a JMP
data table. Pasting automatically creates rows and columns as needed.
If you use the Shift key and select Paste (Option-Paste on the Macintosh) JMP does a Paste
with Labels, which uses the first line of information on the clipboard as column headers.
See the chapter “JMP Data Tables,” p. 93, for more information about the Paste command.
You can drag multiple columns (contiguous or discontiguous) at the same time. Click to highlight the
set of columns you want to move or duplicate. Click a second time and drag the column set to the new
position. If you do this with discontinuous columns, they become contiguous when dropped.
To duplicate a column use Control-drag (Option-drag on the Macintosh). The new column has the
original column’s display format, but not necessarily its modeling type.
You can also move or duplicate only highlighted values. Cell values appear highlighted when you select
both rows and columns. Use the grabber (hand) tool, then click and drag from within a highlighted
cell. All selected cells are moved or duplicated.
Note: If you drag a set of cells to an empty area of the table, new columns are automatically created. If
you drag selected cells to a destination within the data grid, they are moved, leaving missing values in
the original cells, and replacing the values at the destination if the data types are the same. Using Con-
trol-drag (Option-drag on the Macintosh) copies the values to the destination leaving the original val-
ues intact.
6
6 Tables Menu
The Tables Menu
Manipulating tables
The Tables menu has commands that perform a wide variety of data management tasks on JMP data
tables. These commands let you
• sort by any number of columns
• subset by selected rows, or randomly
• stack multiple columns into a single column
• split a column into two columns
• concatenate multiple tables end to end
• join two tables side by side
• update columns in a table with values from another table
• transpose tables.
6
Contents
Overview of Tables Menu Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
The Summary Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
The Summary Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Summary Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Marginal Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
The Subset Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Selecting Rows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
The Sort Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
The Stack Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
The Split Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
The Transpose Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
A Simple Transpose Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Transpose with a Label . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
The Concatenate Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Concatenate Tables with the Same Column Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Concatenate Tables with Different Column Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
The Join Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Join by Row Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Keep a Subset of Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Join by Matching Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Cartesian Join . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Update a Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6 The Tables Menu—Overview of Tables Menu Commands 121
6 Tables Menu
Overview of Tables Menu Commands
The Tables menu commands are accessible from the main menu and as a submenu in the Tables panel
to the left of the data grid (Figure 6.1). The commands in the Tables menu modify JMP data tables or
create a new table by combining one or more tables.
Figure 6.1 The Tables Panel
The following list is a brief description of each command. Details and examples are covered later in this
chapter.
Summary
creates a JMP window that contains a summary table. This table summarizes columns from the
active data table, called its source table. The summary table has a single row for each level of a
grouping variable you specify. Optionally, you can add columns of summary statistics to this table.
Subset
creates a subset data table defined by the selected rows and columns of the active data table win-
dow.
Sort
sorts a JMP data table by one or more columns in either ascending or descending order.
Stack
creates a new data table by combining specified columns from the active data table into a single
new column.
Split
creates a new data table by dividing a specified column into several new columns according to the
values of an ID variable.
Transpose
creates a new data table that is the transposed version of the active data table. The columns of the
active table are the rows of the new table, and its rows are the columns of the new table.
122 6 The Tables Menu—The Summary Command
Concatenate
creates a new data table from two or more data tables by combining them end to end.
Join
creates a new data table by joining (merging) two tables side by side.
6 Tables Menu
Figure 6.2 JMP Table for Grouping Examples
The Summary command displays the dialog in Figure 6.3. To create the summary table shown to the
right of the dialog, first select the variable Type in the columns list of the dialog. Click Group to see
Type in the grouping variables list. You can select as many grouping variables as you want. Click OK to
see the summary table.
The new summary table appears in an active window. This table is linked to its source table and is not
saved when you close it unless you use the Save As command to give it a name and location.
Figure 6.3 The Summary Dialog
A summary table has one column for each grouping variable plus a column named N Rows with counts
for each grouping level. This example shows 20 computer companies and 12 pharmaceutical compa-
nies. The values of the grouping variables are listed in either ascending or descending sort order. To
specify the order of grouping levels, select the variable in the grouping variable list and click the a…Z/
Z…a toggle. By default, the grouping levels appear in ascending order.
The example in Figure 6.4 shows the Companies.jmp data table summarized by both type of company
and size of company, with the values of size listed in descending order within company type.
The summary table behaves in the following ways:
124 6 The Tables Menu—The Summary Command
• When you select rows in the summary table, the corresponding rows are highlighted in its source
table.
• When a summary table is active, there is an Add Statistics Column command in the Column
popup menu at the upper-left corner of the data table, or in the Columns panel to the left of the
data grid. This command accesses the Summary dialog so that you can add statistical summary col-
umns to the table at any time.
• The summary table, like a graphical display, is not saved when you close it. However, you can use
the Save As command to specify a name and disk location for the table.
Figure 6.4 Second Grouping Variable in Descending Sequence
Summary Statistics
Initially, a summary table displays frequency counts (N Rows) for each level of the grouping variables.
However, you can add columns of descriptive statistics to the table.
Click the popup menu icon in the upper-left corner of the data table and select the Add Statistics
Column command. This displays the Summary dialog again. The Statistics popup menu in the Sum-
mary dialog lists standard univariate descriptive statistics. Use the Statistics menu to request any of
these statistics for any numeric source table column. The example in Figure 6.5 shows how to generate
the Mean of Profits/Emp for each level of the grouping variable.
6 The Tables Menu—The Summary Command 125
6 Tables Menu
Figure 6.5 Add Summary Statistics
The Statistics popup menu gives these summary statistics for numeric columns:
N
is the number of nonmissing values and is used to compute statistics when there is no column
assigned the weight role.
% of Total
expresses each value of the selected variable as a percent.
N Missing
is the number of missing values.
Min
is the least value in a column, excluding missing values.
Max
is the greatest value in a column.
Sum Wgt
is the sum of all values in a column assigned the Weight role and is used instead of N to compute
other statistics.
Sum
is the sum of all values in a column.
Mean
is the arithmetic average of a column’s values. It is the sum of nonmissing values divided by the
number of nonmissing values (N–N Missing).
Variance
is the sample variance, computed for the nonmissing values. It is the sum of squared deviations
from the mean, divided by (N–N Missing–1).
126 6 The Tables Menu—The Summary Command
Std Dev
is the sample standard deviation, computed for the nonmissing values. It is the square root of the
sample variance.
Std Err
is the standard error of the mean. It is the standard deviation divided by the square root of N. If a
column is assigned the role of weight, then the denominator is the square root of the sum of the
weights.
Range
is the difference between Max and Min.
Median
is the 50th percentile, which is the value where half the data are below and half are above or equal
to the 50th quantile (median).
Quantiles
gives the value at which the specific percentage of the argument is less than or equal to. The sum-
mary dialog has an edit box for entering the quantile percentage you want.
Figure 6.6 shows the result of the example in Figure 6.5. By default, the name of the statistics columns
is of the form statistic(source column), where statistic is the selection made on the summary dialog, and
source column is the column from the source table selected in the summary dialog. The form of the col-
umn names can be changed to one of the selections for the drop-down menu found on the lower-left of
the Summary dialog (see Figure 6.3)
Figure 6.6 Example of Summary Tables with Summary Statistics
See “The Moments Table for Continuous Variables,” p. 32 in the JMP Statistics and Graphics Guide for
more information about summary statistics.
Another way to add summary statistics to a summary table is with the Subgroup button on the Sum-
mary dialog. This method creates a new column in the summary table for each level of the variable you
specify with Subgroup. The subgroup variable is usually nested within all the grouping variables. The
summary table now becomes a two-way table of summary statistics.
The table that groups the Companies.jmp data table by Type and Size Co can use a subgroup variable
instead of two grouping variables, as illustrated in the next example. Figure 6.7 begins with the
Companies.jmp data table grouped by Size Co. The Statistics popup menu requests the mean of
profit/emp with Type as the subgroup variable. This creates a column for each statistic for each level of
6 The Tables Menu—The Summary Command 127
6 Tables Menu
subgroup variable. The result is the summary table shown in Figure 6.7. There is a row for each size
company and a column for the mean of each type of company. The cells hold the mean for the sub-
group defined by the intersection of the row and column.
Figure 6.7 Summary Statistics for Subgroup
Marginal Statistics
Note that there is a check box on the Summary dialog to Include marginal statistics (Figure 6.7).
When marginal statistics are requested, the resulting summary table has an extra row for each level of
the first variable in the Group list that gives the requested summary statistics for that level. There is also
an additional row at the end of the table that shows the summary statistics for the whole table.
Figure 6.8 shows the Summary Table for group variables Type and Size Co with marginal statistics.
Figure 6.8 Summary Table with Marginal Statistics
128 6 The Tables Menu—The Subset Command
Selecting Rows
To select rows, click row number areas in the data grid or define subsets with a summary table as
described previously, or highlight points in plots. If no rows or columns are selected, then Subset
reproduces the entire data table.
6 Tables Menu
Figure 6.10 Subsetting Highlighted Columns with a Histogram
The columns you add to the By list establish the order of precedence for sorting. The first column is the
major sort field. Each successive column in the list sorts after the preceding column. The icon to the left
of the sort column indicates whether the values sort in ascending or descending order. By default, val-
ues sort in ascending order. You can toggle the sort between ascending and descending for any selected
column in the By list with the a…Z/Z…a button. Click Sort when the dialog is complete.
Figure 6.11 shows the JMP sample data table Popcorn.jmp in its original form (top table) and sorted
by popcorn and yield (within popcorn) (bottom table). In this example, Sort created a new data table.
However, if you check the Replace Table option, the sorted table overwrites the original data table.
Figure 6.11 Sort in Ascending or Descending Order
Note: If there are formulas in any columns other than the Sort columns of the data table to be sorted,
the Sort dialog shows a Copy Formula check box option. If you don’t check this option, no formulas
are included in the sorted table. There is also a Suppress Formula Evaluation check box that suppresses
formula evaluation when the sorted table is initially created.
6 Tables Menu
The Stack command creates a special Source Lable Column (ID column) to identify each row in the
new table. Its values are the column names in the original table from which the stacked values origi-
nated. In the example below, the name of the ID column is the default name (_ID_), and it has the lit-
eral values ‘yield1’ or ‘yield2.’
Note: Be aware that if you delete the default column name, _Stack_ and _ID_, and do not specify new
names, these new columns are not written to the data table.
You can enter a name for the data table in the Output table name box. This example uses the default
table name, Untitled.
Click Stack when the dialog is complete. In this example, the stack operation creates the new untitled
data table shown at the bottom in Figure 6.12. This new table has twice as many rows as the original
table because two columns are stacked. The values of _Stack_ are the experimental results, and values
in the _ID_ column tell which trial each row represents.
Figure 6.12 Stack Command Dialog and Stacked Tables Example
Check box options on the Stack dialog affect the rows and columns in the new stacked table,
• If no options are used, the rows in the first column in the Stack list are first in the new stacked
table, followed by all the rows in the second column.
• If Stack by Row is checked, as in Figure 6.12, the stacked results are interleaved.
• If all values to be stacked are missing for a row, the Eliminate Missing Rows option drops that row
from the new table.
• When Drop All Other Columns is checked, only the stacked column and the ID column appear in
the new table.
Note: If there are formulas in any columns other than the Stack and ID columns of the data table to be
stacked, the Stack dialog shows a Copy Formula check box option to save formulas from the original
tables. If you don’t check this option, no formulas are included in the new table. There is also a
Suppress Formula Evaluation check box that suppresses formula evaluation when the new table is ini-
tially created.
132 6 The Tables Menu—The Split Command
Note: If there are formulas in any columns other than the Split and ID columns of the data table to be
split, the Split dialog shows a Copy Formula check box option to save formulas from the original table.
If you don’t check this option, no formulas are included in the new table. There is also a Suppress
Formula Evaluation check box that suppresses formula evaluation when the new table is initially cre-
ated.
6 The Tables Menu—The Transpose Command 133
6 Tables Menu
The Transpose Command
The Transpose command creates a new JMP table that is a transposed version of the active data table.
The columns of the active table are the rows of the new table, and its rows are the new table’s columns.
The new table has an additional column called Labels whose values are the column names of the origi-
nal table. If there is no label column, the column names in the transposed table are Row 1, Row 2, …,
Row n where n is the number of rows in the original table. If there is a variable assigned the label role,
the values of that column are column names in the new transposed table.
The Transpose command has the following characteristics:
• The columns of the original table must be either all character or all numeric, except a column used
as an ID variable or columns used for grouping.
• Transpose can transpose any selected subset of rows.
• Transpose can transpose groups of rows. Subsets defined by a summary table created with the
Summary command (described previously in this section) transpose independently and stack to
form the new transposed table.
• Transposing a table with columns but no rows gives a new table with one column that lists those
column names. Likewise, if you create a table with one column and assign it the Label role, its val-
ues become the column names in the transposed table.
Figure 6.14 Simple Transpose (left) and Transpose with a Label (right)
6 Tables Menu
Figure 6.15 Concatenate Tables with the Same Column Names
You can concatenate as many tables as you want. The number of rows in the new table is the sum of the
number of rows in all the tables.
136 6 The Tables Menu—The Join Command
Note: The Concatenate dialog (Figure 6.15) has a check box option to Save and Evaluate Formulas
from the original tables. If you don’t check this option, no formulas are included in the new table.
When that option is checked, all formulas found in the first table in the concatenate list on the dialog
are included in the new table.
When you select Join, the dialog in Figure 6.18 displays the name of the active table next to the word
Join. You select an open data table to join With from the table selector list.
After you select a With table, choose a matching specification from the Join dialog. The default match-
ing option is By Row Number. Click Join to see the new data table shown in Figure 6.18. The two
original tables are joined side by side, and the new table has all columns from both tables.
If a column name is the same in the original tables, the names of these columns in the new table are of
the form variable name of table name.
If a name (such as oil amt) occurs in only one table, it is written directly to the new table.
6 The Tables Menu—The Join Command 137
6 Tables Menu
Figure 6.18 Join Dialog and Joined Table
Note: The Join dialog (Figure 6.18) has the Copy Formula and Suppress Formula Evaluation
check box options to save and evaluate formulas from the original tables. If you don’t check this
option, no formulas are included in the new table. When the Copy Formula option is checked, for-
mulas in columns other than matching columns are included in the new table. The Suppress
Formula Evaluation suppresses formula evaluation when the new table is initially created.
To join these tables correctly, the values for popcorn and batch must match in both tables, and oil amt
in the Trial1.jmp table must match with oil in the Little.jmp table. The By Matching Cols option on the
Join dialog lets you specify columns whose values must match to complete a join.
By Matching Cols displays the Match Columns dialog, which prompts you to select columns whose
values must match in both tables for rows to be joined. After selecting a column from each upper list,
click Match to see the selected pair of columns displayed in the lower box. Select additional pairs of
6 The Tables Menu—The Join Command 139
6 Tables Menu
columns as needed. If you want to remove matched columns, select them in the lower lists and click
Remove.
Matching columns do not have to have the same names and do not have to be in the same relative col-
umn position in both tables. When you click Done, the first column in the left-hand list pairs with the
first column in the right-hand list, the second columns are paired, and so on. Rows join only if values
match for all the column pairs.
Figure 6.21 Join by Matching Columns Dialog
After you choose columns whose values must match, additional options show beneath each table’s col-
umn list to further qualify the join operation. For example, some rows in the Trial1.jmp table (see
Figure 6.17) do not have a match in the Little.jmp table.
Click Include Non matches box for the Trial1.jmp table so that the new table has all rows in Trial1.jmp
even if there is no match in the Little.jmp table.
After the Match Columns dialog is complete, click Done to return to the Join dialog. When you click
Join, JMP creates the table shown in Figure 6.22. The yield column from the Little.jmp data table
(Yield of Little) has missing values whenever there were no matching values in the Trial1 table.
Notice that the new table is now sorted by the matching columns. JMP automatically sorts the data
table so that matching takes place properly. You do not need to sort tables before joining them with
matched columns.
Figure 6.22 Table Joined by Matching Columns
If you specify this option for one table, the first match value is joined with all matches in the other
table.
• When you select Include Non Matches for a data table, each row from that data table is included in
the new data table even when there is no matching value, as in the previous example. You can spec-
ify this option for either or both data tables being joined.
If you do not select either the Drop Multiples or the Include Non Matches option, a Cartesian join
operation is performed within each group of matching column values.
Cartesian Join
If you choose the Cartesian Join option, each row in the Join data table joins with each row in the
With data table. To illustrate this, suppose you want to construct a JMP table that has a row for each
combination of levels of experimental conditions in the popcorn example.
You can begin with the three small tables shown at the top in Figure 6.23. Each table has two rows and
one column. The values are the experimental categories for popcorn yield trials.
The Join command joins the active data table with any other open data table you choose. In this exam-
ple, you must use the Join command twice. The first join combines the oil amt.jmp data table with the
batch.jmp table using the Cartesian option. The lower-left table in Figure 6.23 has all columns from
the original tables. Each value in the oil amt table pairs with each value in the batch table, giving a new
table with four rows.
Next, Join the new table (the result of the Cartesian join of oil amt and batch) with popcorn. Again
use the Cartesian Join option, and click Join to produce the lower-right table in Figure 6.23.
This final data table has a row for each experiment condition and is ready for recording results of the
corn popping trials. Keep in mind that the number of rows produced by a Cartesian join is the product
of the number of rows in the original tables.
Note: Another way to produce a factorial design table is to use the Full Factorial command on the
DOE menu.
Figure 6.23 A Cartesian Join
6 The Tables Menu—The Join Command 141
6 Tables Menu
Update a Table
The Join command has an option to update the Join table with information you select from a table in
the With list. The tables at the top in Figure 6.24 could be an update example. Suppose you want to
change the values of oil amt to be “1 tablespoon” instead of “little” and “4 tablespoons” instead of
“lots”. To do this, complete the join dialog as shown, with the Update first table with data from
second table box checked.
When you click Join, a new table is created that has the information from the second table written over
the information in the first table for columns with the same names.
The other options in the join dialog are the same as other join options. If the column names in the
update table are different than those to be updated in the first table, change the Matching Specification
to By Matching Columns and complete the Match Columns dialog.
Figure 6.24 Example of Updating a table with the Join command
7
7 Edit Formula
Using the Formula Editor
This chapter describes the Formula Editor and shows how to build formulas. The next chapter, For-
mula Editor Functions, p. 163, gives definitions and examples of the functions supported by the For-
mula Editor.
Formulas can be simple assignments of numeric, character, or row state constants, or can contain com-
plex evaluations based on conditional clauses.
When you create a formula for a column, that formula becomes an integral part of the data table. The
formula is stored as part of a column’s information when you save the data table, and it is retrieved
when you reopen the data table. You can examine or change a column’s formula at any time by opening
its Formula Editor window.
A column whose values are computed by a formula is both linked and locked. It is linked to (or depen-
dent on) all other columns that are part of its formula. Its values are automatically recomputed when-
ever you edit the values in these columns. It is also locked so that its data values cannot be edited, which
would invalidate its formula.
The Formula Editor window operates like a pocket calculator with buttons, displays, and an extensive
list of easy-to-use features. This chapter begins with a simple example and shows how to build formu-
las.
7
Contents
A Quick Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Formula Editor Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
The Formula Work Panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Formula Element Browser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Keypad . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Function Browser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
The Formula Display Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
Keypad Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Formula Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Table Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Table Columns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Local Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Working with Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Computational Order of Precedence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Building a Formula in Order of Precedence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Constant Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Cutting and Pasting Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Selecting Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Dragging Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Tips on Editing a Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Table Templates and Scripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Grid Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Actuarial Life Table Template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Central Limit Theorem Template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Nonlinear and Loss Function Templates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Keyboard Shortcuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
7 Using the Formula Editor—A Quick Example 145
7 Edit Formula
A Quick Example
The following example gives you a quick look at the basic features of the Formula Editor. Keypad and
function browser definitions are in this chapter. The next chapter, Formula Editor Functions, p. 163,
documents the function library.
For this example, open the Students data table in the sample data folder. It has a column called weight.
Suppose you want a new column that computes standardized weight values. Begin by selecting New
Column from the Cols menu.
The New Column dialog (Figure 7.1) lets you set the new column’s characteristics. Type the new name,
Std. Weight, in the Column Name area. Then select Formula from the New Property popup menu,
which displays formula information on the New Column dialog, and the Formula Editor window
shown in Figure 7.2. The other column characteristics define a numeric continuous variable and are
correct for this example.
Figure 7.1 Create a New Column with the Formula Property
Click Edit Formula on the formula information panel to see the Formula Editor window, as shown in.
function browser
keypad
formula element
browser
formula editing
area
To build a formula, you first click the empty formula element in the formula editing area to select it.
When you click a formula and Show Boxing is checked, the selected portion of the formula is outlined
146 7 Using the Formula Editor—A Quick Example
with a thin red line. All terms within the smallest nesting box relative to the place you clicked become
selected, and the subsequent action applies to those combined elements. Boxing becomes very useful
when you are building complex formulas and want to modify specific terms within the formula, or
need to determine the order evaluation takes place.
Enter the formula that standardizes the weight values by following these steps.
1. While the initial missing term is selected, click on the column named weight in the column selector
list. Then Click on the minus button in the keypad. A new missing term appears after the minus
sign as shown in Figure 7.3.
2. While the new missing term is selected, click on weight again.
3. Click in the function browser topics list, and scroll to locate Statistical functions. Click on this
topic to see a list of statistical functions, and then select Col Mean in this list.
4. Click anywhere in the white space of the total boxed area to select the entire expression. The red box
should now enclose the whole formula.
5. While the entire expression is selected, click on the division button in the keypad. The result gives a
selected missing denominator for the whole expression.
6. Choose weight again from the column selector list.
7. While weight is still selected in the denominator, choose Col Std Dev from the Statistical func-
tions.
You have now entered your first formula. Close the Formula Editor window to see the new column fill
with values. When a weight value changes, the calculated std weight value automatically recalculates.
Figure 7.3 Building a Formula
The Column Info dialog for a computed column displays its formula (see Figure 7.4). If the column’s
formula is larger than four inches wide or five inches tall, it will be shrunk to fit within these limits. The
formula can be resized by clicking and dragging from the right or bottom border. To preserve the pro-
portions of the formula picture, use Shift-drag.
To see the Formula Editor again, click Edit Formula on the Column Info dialog. Or, highlight one or
more columns that have formulas you want to see and choose Formula from the Cols menu. You can
also Right-click (Control-click on the Macintosh) at the top of any column to bring up its Formula
Editor.
7 Using the Formula Editor—Formula Editor Terminology 147
7 Edit Formula
Figure 7.4 Column Info Dialog for Column with Formula
Text Mode
You also have the option of entering or changing any part of a formula in text mode. Text mode displays
the formula in JMP scripting language (JSL). The entire formula (or any of its terms) appears in text
mode when you double click in the white space inside the boxed formula area. The elements of that box
appear as plain text and you can then edit the formula as you would in any text editor. Any element of
a formula can be displayed as a scripting command and then edited. After editing formula scripting
commands, click the formula again to see its formatted form. For example, the text for the standardiza-
tion of weight is
(:weight – Col Mean(:weight)) / Col Std Dev(:weight)
argument
Any element or an entire expression (including mathematical operands) that is operated on by a func-
tion is an argument to that function.
term
An indivisible part of an expression such as Constants and variables are terms.
expression
An expression is any part of a formula that can be selected as a single unit, including terms, missing
terms, and functions grouped with their arguments, as well as the entire formula.
Clause
A complete segment in a conditional function.
function
A mathematical or logical operation that performs a specific action on one or more arguments. Func-
tions include most items in the function browser and all keypad operators.
missing term
Any empty place holder for an expression, represented by an empty box in the formula.
missing value
Excluded or null data consisting of the missing value mark (•) for numeric data, or null character
strings for character data.
Figure 7.5 Formula Terminology
7 Edit Formula
The following paragraphs briefly describe each component of the work panel. Formula elements are
covered in detail later in this chapter, and the next chapter, Formula Editor Functions, p. 163, which
documents all functions.
Keypad
The keypad is a set of buttons used to help build formulas. Some of the buttons, such as the arithmetic
operators, are familiar. Others have special functions, described in the next section. The popup menu
icon has the commands shown in Figure 7.6.
Figure 7.6 Keypad and Formula Editor Menu Commands
dialog appears when there is an error and asks if you want to ignore further errors. This has the
same effect as the Ignore Errors menu selection.
Larger Font and Smaller Font
incrementally increases or decreases the font used to display the formatted formula.
Derivative
finds and displays the derivative of a function you enter into the Formula Editor. The derivative is
found with respect to a function argument (a single variable name) you highlight. Thus, in order
to differentiate with respect to x, it must be one of the immediate arguments in an expression you
enter.
• enter function into column
• highlight variable
• select Derivative from menu
Function Browser
The function browser groups collections of functions and features in lists organized by topic. To enter a
function in a formula, select an expression and click any item in one of the function browser topics.
You can also see a list of ungrouped functions in alphabetical order by selecting Functions (All) from
the popup menu above the function list.
7 Edit Formula
Keypad Functions
The keypad is composed of common operators (referred to as functions in all Formula Editor docu-
mentation). Enter a keypad function by selecting an expression in the formula display and clicking on
the appropriate keypad buttons. The keypad buttons are shown in Figure 7.7, followed by a description
of each button function.
Figure 7.7 Keypad Keys and Explanations
In the example shown in Figure 7.8, Col 2 is first selected (step 2). Repeated clicks on the Delete
Expression key then produce the sequence of steps from 3 to 10.
Figure 7.8 Deleting Terms from a Formula
The Delete key also lets you peel functions from their arguments. To do this, select a function
and its arguments, then click the delete button. The function and all but its first argument are
deleted.
Formula Elements
The formula elements browser displays the categories of items to build formulas. The categories are
accessed from the menu at the upper-left of the calculator, as illustrated in Figure 7.9.
Figure 7.9 Formula Elements Browser
Constants
The Constants list includes commonly used numeric constants such as e and π, as well as some com-
monly used integers such as –1, 0, 1 and 2. To use a constant, select a term in the model and select an
item in the constant list. Also, when you double-click a term in a formula it becomes a text box and you
can enter numeric, character, or row state constants by typing them.
7 Using the Formula Editor—Formula Elements 153
7 Edit Formula
Table Variables
Table variables are character strings that are available to the entire table. Table Variable names are dis-
played in the Tables panel at the left of the data grid. The Formula Editor can refer to a table variable in
a formula. Many of the sample data files have a Table variable called Notes. It has as its value the table
notes carried over from previous versions of JMP. The Table Variables command in the Formula Ele-
ments popup menu lists all the Table variables that exist for a table. You can create additional Table
variables with the New Table Variables command in the Tables panel of the data table.
Table Columns
Table columns, the default element list, are the data table columns. To insert columns into a formula,
select a formula term and click the column name in the Table Columns list. If you delete a column used
in a formula, an empty term appears in its place. The Table Column list updates whenever you add or
delete a column, or edit a column name.
Local Variables
Local variables lets you define temporary numeric variables to use in expressions. Local variables exist
only for the column in which they are defined.
To create a new local variable, select Local Variables from the Formula Elements popup menu and
complete the dialog, as illustrated in Figure 7.10. By default, local variables have the names t1, t2, and
so on, and have missing values. Optionally, you can assign a starting value in the Local Variable dialog.
Local Variables appear in a formula as bold italic terms.
Local Variables are most often used with Assignment functions. Assignment functions can assign
expressions to local variables, which are then used in a complex equation. This technique can some-
times simplify building an equation and improve the efficiency of its evaluation.
As an example, suppose you have variable X and Y and you want to compute the slope in a simple lin-
ear regression of Y on X using the standard formula shown in Figure 7.10.
One way to do this is to create two local variables, call them XY and Xsqrd, as described in Figure 7.10.
Then assign them to the numerator and the denominator calculations of the slope formula. Delimit
each assignment with a semicolon. The slope computation is simplified to XY divided by Xsqrd.
154 7 Using the Formula Editor—Formula Elements
Note: Statements in the Formula Editor are actually JSL programming statements. Multiple statements
in a formula must be separated by a semicolons.
Note: Another way to create local variables is to use the local variable button (t=) on the keypad, which
automatically creates and displays a local variable with the default name, t1, in an expression and places
a semicolon after it.
Use Alt-click (Control-click on the Macintosh) on the name in the local variable selection list to see a
popup menu with commands to copy, edit or delete, as described next under Parameters, p. 154.
Parameters
A Parameter is a special type of numeric local variable created in the Formula Editor and can be used in
any formula. Numeric parameters are most useful in formulas created for nonlinear fitting. You can set
the initial value of a parameter when you create it, but its final value is determined by the formula com-
puting process that contains it.
To create a parameter, select Parameters from the formula element popup menu and click New
Parameter to bring up the dialog shown in Figure 7.11. You use the dialog to assign a name and value
to the new parameter. You can create as many parameters as you need, but they are only available to the
Formula Editor (column) that created them. Click OK when the dialog is complete or Cancel to exit
the dialog without creating a new parameter.
To change parameter settings or remove a parameter, Control-click (Option-click on the Macintosh)
the parameter name and chose Edit from the menu (see Figure 7.11).
Parameters are added to formulas in much the same way that variables are added. To insert parameters
into a formula, select a formula term and click the parameter name in the parameter list.
Parameters are easy to recognize in formulas because they are displayed in bold type. If you open the
US Population.jmp table from the Nonlinear Examples in the Sample Data folder and display
Parameters, you see the parameters shown in Figure 7.11, with the formula that uses them.
7 Using the Formula Editor—Working with Formulas 155
7 Edit Formula
When a parameter is used in a model for the Nonlinear platform, an initial value or starting value
should be given to the parameter. After completing a nonlinear fit or after using the Reset button in
the nonlinear control panel, the parameter’s value is the most recent value computed by the nonlinear
platform.
A new parameter shows in the parameter list with an initial value if there is one. When a computing
process changes the value of a parameter, the new value then appears in the parameter list.
When you paste a formula with parameters into a column, the parameters are automatically created for
that column unless it has existing parameters with the same names.
Figure 7.11 Using Parameters in a Formula
Note: When a function has an expression as its argument, the argument has a higher order of prece-
dence than it would if enclosed in parentheses outside the function.
Because order of precedence determines which arguments are affected by each function, order of prece-
dence also affects the grouping of expressions. Select functions in the formula to verify how the order of
precedence rules have been applied.
Constant Expressions
Once JMP has evaluated a formula, you can select an expression to see its value. This is true for both
parameters and expressions that evaluate to a constant value. To do this, select the expression you want
to know about and Right-click (Control-click on the Macintosh) on it. This displays a popup menu as
shown in Figure 7.12. When you select the Evaluate command and release the mouse click, the current
value of the selected expression shows until you move the cursor.
Figure 7.12 Context-Click to See the Value of an Expression
7 Using the Formula Editor—Working with Formulas 157
7 Edit Formula
Cutting and Pasting Formulas
You can Cut or Copy any expression or an entire formula and Paste it into another formula display.
The formula is saved on the clipboard as a JSL statement. When you paste it into another formula dis-
play, it appears in formatted form. If you copy it into other applications, it appears as a JSL statement.
Selecting Expressions
You can use the keyboard arrow keys to select expressions for editing or to view the order of precedence
within a formula when parentheses are not present or the boxing option is not in effect. You can click
on any single term in an expression to select it for editing. Clicking an operand (+, –, *, ÷) in an
expression selects the operand and its operator.
Once an operand is selected, the left and right arrow keys move the selection across other associative
operands having equal precedence within the expression. The left arrow extends the selection to include
the next operand and operator of higher precedence to the selection, until the entire formula is selected.
The up arrow extends the current selection by adding the operand and operator of higher precedence to
the selection. The down arrow reduces the current selection by removing an operand and operator from
the selection.
Dragging Expressions
You can drag any part of a formula that can be selected to any other location than can be selected.
When you place the arrow cursor inside an expression and click, the expression highlights and the cur-
sor changes to a hand cursor. As you drag across the formula, destination expressions are highlighted,
and then you can drag the selected expression to the new location you want. When you drag, the
selected expression is copied to the new location where it replaces the existing expression when the
mouse button is released.
After you complete formula changes, the new values fill the column automatically when you click
Apply or close the Formula Editor window.
Once you have created a formula, you can change values in columns that are referenced by your for-
mula. JMP automatically recalculates all affected values in the formula’s column.
Efficiency
Usually, it is not necessary to structure formulas with efficient evaluation in mind. Most formulas eval-
uate almost instantaneously regardless of their structure. This is because statistical functions and con-
stant expressions are evaluated only once when a column’s values are calculated.
However, when you are creating conditional expressions, keep in mind that Match evaluates faster and
uses less memory than an equivalent Condition function, If.
Consider the following two formulas for predicting a child’s height from his age. In each case there is a
base height of 58.125 inches to which a quantity is added depending on the value of the age variable.
The Match conditional evaluates faster than the If function because the age variable is evaluated only
once for each row in the data table. The If condition must evaluate the age variable at each If clause for
each row until a clause evaluates as true.
7 Edit Formula
In addition, most table templates also have JSL scripts in the Templates folder that accomplish the same
function as the template. See the JMP Scripting Guide for documentation of script commands.
Grid Templates
The grid templates provide a quick way to compute an evenly spaced x, y grid for three-dimensional
plotting. There are two grid templates in the Templates folder of the Sample Data:
• The template Grid Template (x=y) produces a grid in which x and y both range from columns called
Start Grid to End Grid.
• The template Grid Template (x !=y) allows x to range from columns called Start x to End x, while y
ranges from Start y to End y.
To use a grid template, first save it under a new name. Next, specify the grid endpoints in the first row
of the new table. Then, enter a function of x and y into the Surface Formula column. For example, you
might want to paste a prediction formula from a response surface analysis into the Surface Formula.
At this point you still have only one row in your table. To obtain the grid, choose Add Rows from the
Rows menu. Enter the number of rows you want minus 1 into the How Many Rows to add box. As
soon as you add rows, JMP evaluates all column formulas. To view the data in three dimensions, use the
Spinning Plot platform.
Note: To create an n-by-n square grid, you add n2 – 1 add rows to the data table because there is already
one row in the table.
If you need to alter the grid size or several axis ranges, remove all but the first row from your data table
before doing so. This allows JMP to reevaluate formula columns for only the first row until all changes
are made. See the Cowboy Hat Template and Odor Control Surface template data tables for examples
that use grids.
Note: Adding colors to your points based on the value of the column Surface Formula can enhance
your spinning plot and is simple to do. Use the Distribution of Y to create a histogram of Surface
Formula column. Get the hand tool and adjust the number of bins until there are at most twelve bins.
Choose Save Level Numbers from the Surface formula popup menu to save the bin levels with values
1 to 12 into the data table. Now create a row state column and give it the formula:
Hue State(Level Surface Formula)
After you enter or paste your data into the life table template, you can graph the distribution functions
using the Overlay Plot command in the Graph menu.
See the JMP data table called Angina Life Table Data table (Lee, 1980) for a life table example. It is in
the Example Data for Templates folder in the Templates folder of the Sample Data.
Note: The complete reference for (Lee, 1980) is in the appendix “References,” p. 679 of theJMP Statis-
tics and Graphics Guide.
Keyboard Shortcuts
You can use the mouse or the keyboard shortcuts shown in Table 7.2 Keyboard Shortcuts, to apply
Formula Editor features. Replace a selected expression with the function or argument by entering the
corresponding key command.
7 Edit Formula
Table 7.2 Keyboard Shortcuts
Item Shortcut
() set of parentheses (
() set of parentheses )
• * on keypad
+ + on keypad
– – on keypad
÷ / on keypad
+/– Shift-minus
And &
Not !
Or |
Power ^
Delete Delete
New argument ,
x<y <
x=y =
x>y >
8
8 Functions
Formula Editor Functions
Formulas can be simple assignments of numeric, character, or row state constants or can contain com-
plex evaluations based on conditional clauses.
When you create a formula for a column, that formula becomes an integral part of the data table. The
formula is stored as part of a column’s information when you save the data table, and it is retrieved
when you reopen the data table. You can examine or change a column’s formula at any time by opening
its Formula Editor window.
A column whose values are computed by a formula is both linked and locked. It is linked to (or depen-
dent on) all other columns that are part of its formula. Its values are automatically recomputed when-
ever you edit the values in these columns. It is also locked so that its data values cannot be edited, which
would invalidate its formula.
The Formula Editor window operates like a pocket calculator with buttons, displays, and an extensive
list of easy-to-use features. This chapter begins with a simple Formula Editor example and serves as a
reference guide for all Formula Editor functions and features.
8
Contents
The Function Browser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Row Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Numeric Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Transcendental Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Trigonometric Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Character Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Comparison Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Conditional Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Probability Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Statistical Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Random Number Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Date Time Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Row State Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Assignment Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
8 Formula Editor Functions—The Function Browser 165
8 Functions
The Function Browser
The function browser groups the Formula Editor functions by the topics as shown at the left in
Figure 8.1. When you click a function group, a popup menu lists the functions in that group. Alterna-
tively, you can select a choose Functions (All) from the menu above the function browser to see all func-
tions listed in alphabetical order.
Figure 8.1 The Function Browser
To use a function, select an expression in a formula by clicking on it and then click any item in the
function browser topics. When the formula appears in the editing area, highlight a term in its argument
and click on columns, other functions, or other formula elements to enter them into the argument.
Most formulas give hints about appropriate arguments, and show a small caret in the argument area if
additional arguments can be added.
The function categories are briefly described in the following list. The remaining sections in this chap-
ter give a detailed description and examples of each function in the function categories. They are pre-
sented in the order you find them in the function browser.
• Row functions lists miscellaneous functions such as Lag, Dif, Subscript, Row, and NRow
• Numeric functions are terms commonly used in formulas.
• Transcendental functions are functions such as natural log, common log, exponential, root, facto-
rial, combinatorial, beta, and gamma.
• Trigonometric functions are the standard trigonometric functions: sine, cosine, tangent, inverse
functions, and hyperbolic functions.
• Character functions operate on character arguments for trimming, finding the length of a string,
changing numbers to characters or characters to numbers, and more.
• Comparison functions are the standard logical comparisons such as less than, less than or equal to,
not equal to, and so forth.
166 8 Formula Editor Functions—Row Functions
• Conditional functions are programming-like functions such as If, Match, and Choose.
• Probability functions compute probabilities and quantiles for the normal, Student’s t, chi-square,
and F distributions.
• Statistics functions calculate standard statistical quantities such as the mean or standard deviation.
• Random functions generate random numbers based on predefined distributions such as the uni-
form, normal, Cauchy, and so on. There is also a function to randomize the order of table rows.
• Date Time functions require arguments with the date data type, which is interpreted as the number
of seconds since January 1, 1904. Date Time functions return values such as day, week, or month of
the year. They can also compute dates and can find data intervals.
• Row State functions assign or detect row state status of color, marker, label, hidden, excluded, or
selected.
• Assignment functions place the value on the right side of the assignment operator into the variable
on the left side of the operator.
Row Functions
Row Functions let you reference specific rows or cells within specific rows:
Lag
The Lag function returns the value of the first argument in the row defined by the current row less the
second argument. The default Lag is one, which you can change to any number. The value returned for
any lag that identifies a row number less than one is missing. Note that Lag(X, ßn) gives the same
result as the subscripted notation, XRow( )–n.
Dif
The Dif function returns the difference between the value of the first argument in the current row and
its value in the row defined by the current row less the second argument. The default Dif is one, which
8 Formula Editor Functions—Row Functions 167
8 Functions
you can change to any number. Note that Dif(X, n) gives the same result as XRow()–XRow()-n, or as
XRow()–Lag(X, n).
Subscript
The Subscript function enables you to use a column’s value from a row other than the current row. To
use a subscript, enter a column name in the formula display and click Subscript to display the col-
umn’s empty subscript argument. You can enter any numeric expression into the subscript argument.
Subscripts that evaluate to nonexistent row numbers produce missing values. A column name without a
subscript always refers to the current row.
To remove a subscript from a column, select the subscript and delete it. Then delete the missing box.
The formula CountRow() – CountRow()–1, where Row() is the row number as described below, uses
subscripts to calculate the difference between each pair of values from the column named count. This
result is the same as that given by the Dif() function. When Row() is 1, the computation produces a
missing value.
The formula
calculates a column called Fib, which contains the terms of the Fibonacci series (each value is the sum of
the two preceding values in the calculated column). It shows the use of subscripts to do recursive calcu-
lations. A recursive formula includes the name of the calculated column, subscripted such that it refer-
ences only previously evaluated rows (rows 1 through (i–1)).
The calculation of the Fibonacci series shown above includes a conditional expression and a compari-
son. See the sections Comparison Functions, p. 178, and Conditional Functions, p. 179, for more
information about these functions.
Row
The Row function is the current row number when an expression is evaluated for that row. You can use
Row() in any expression, including column name subscripts. The default subscript of a column name is
Row() unless otherwise specified.
NRow
The NRow function is the total number of rows in the active data table.
168 8 Formula Editor Functions—Numeric Functions
Numeric Functions
Abs
The Abs (absolute value) function returns a positive number of the same magnitude as the value
of its argument. For example, |5| and |–5| both result in 5.
Modulo
The Modulo function returns the remainder when the second argument is divided into the first. For
example, Modulo(6, 5) results in 1.
Ceiling
The Ceiling function returns the smallest integer greater than or equal to its argument. For example,
Ceiling(2.3) results in 3, while Ceiling(–2.3) results in –2.
Floor
The Floor function returns the largest integer less than or equal to its argument. For example,
Floor(2.7) results in 2, but Floor(–.5) results in –1.
Round
The Round function rounds the first argument to the number of decimal places given by the second
argument. For example, Round(3.554, 2) rounds to 3.55 and Round(3.555, 2) rounds to 3.56.
8 Formula Editor Functions—Numeric Functions 169
8 Functions
Count
The Count function creates a list of values beginning with the from value and ending with the to
value. The number of steps specifies the number of values in the list between and including the from
and to values. Each value determined by the first three arguments of the Count function occurs con-
secutively the number of times you specify with the times argument. When the to value is reached,
count starts over at the from value.
Optionally, you can add the times argument with the insert button on the keyboard. This argument
is on by default, but repeats the count process as many times as you specify, as illustrated by the Count4
column in the data table in Figure 8.2. To add any argument to the Count function, highlight the argu-
ment preceding the one you want to enter and either type a comma or use the insert button.
The columns in the data table shown in Figure 8.2 result from the following formulas:
• Count (1, 9, 2) gives Count 1
• Count (1, 9, 3) gives Count 2
• Count (1, 9, 9) gives Count 3
• Count (1, 9, 3, 3) gives Count 4
Figure 8.2 Example of the Count Function
The Count function is especially useful for generating a column of grid values. For example, the follow-
ing formulas create a square grid of increment NRow() (the Row function that gives the total number
of rows in the data table) and axes that range from –5 to 5. There are table templates in the Sample
Data folder for grids. See the previous chapter, Using the Formula Editor, p. 143, for details.
Count (–5, 5, Root(NRow()))
Count (–5, 5, Root(NRow(), Root(NRow()))
170 8 Formula Editor Functions—Transcendental Functions
Transcendental Functions
In the Transcendental functions category, the Formula Editor supports logarithmic functions for any
base, functions for combinatorial calculations, the Beta function, and several gamma functions.
Exp
The Exp function raises e to the power you specify. Thus, Exp(1) = e.
Log Log10
The Log function calculates the natural logarithm (base e). To change the default base, highlight the
argument and type a comma or click the insert key on the keypad. The base appears and is editable.
The Log argument can be any numeric expressions. The expression Log(e) evaluates as 1, and
Log2(32) is 5. The Log10 function calculates the logarithm of base 10 only.
Squish
The Squish function is an efficient computation of the function, 1 / (1+e-x), where x is any numeric
column, variable, or expression.
Root
The Root function calculates the root of its argument as specified by the index. Root initially shows
with an index of 2. To change the index, highlight the index argument and enter the value you want.
8 Formula Editor Functions—Transcendental Functions 171
8 Functions
Factorial
The Factorial function returns the product of all numbers 1 through the argument you specify. For
example, Factorial(5) evaluates as 120.
NChooseK
The NChooseK function returns the number of n things taken k at a time (n choose k). and is com-
puted in the standard way using factorials, as n! / (k!(n – k)!). For example, NChooseK(5,2) evaluates
as 10.
Beta(m, n)
The two parameter Beta function, B(m, n), is written terms of the Gamma function as
Γ ( m )Γ ( n )
B ( m, n ) ) = --------------------------
Γ(m + n)
See the discussion of the Gamma function below.
Gamma
LGamma
The LGamma function is the natural log of the result of the gamma function evaluation. You get the
same result using the natural log function in the Formula Editor with the Gamma function. However,
the LGamma function computes more efficiently than do the Log (natural log) and the Gamma func-
tions together.
Trigonometric Functions
The Formula Editor supports most Trigonometric functions. The following list shows how each func-
tion appears in the Formula Editor, and gives an example.
The Sine and Cosine functions calculate the sine and cosine of their respective arguments given in
radians. For example, the expression Sine(0) evaluates as 0, and Cosine(0) evaluates as 1. The
Tangent function calculates the tangent of an argument given in radians. The expression Tan(.25)
evaluates as 0.255342.
The ArcSine and ArcCosine functions return the inverse sine and inverse cosine of their respective
arguments. The returned value is measured in radians. For example, both expressions ArcSine(1) and
ArcCosine(0) evaluate as 1.57080. The ArcTan function returns the inverse tangent of its argument.
The returned value is measured in radians. The expression ArcTan(0.5) evaluates as 0.46364.
8 Formula Editor Functions—Character Functions 173
8 Functions
SinH, CosH, TanH
The SinH and CosH functions return the hyperbolic sine and hyperbolic cosine of their respective
arguments. The expression SinH(1) evaluates as 1.175201, and CosH(0) evaluates as 1.0. The TanH
function returns the hyperbolic tangent of its argument. The expression TanH(1) evaluates as
0.761594.
The ArcSinH and ArcCosH functions return the inverse hyperbolic sine and inverse hyperbolic cosine
of their respective arguments. The expression ArcSinH(1) evaluates as 0.881374, and ArcCosH(1) is
0. The ArcTanH function returns the inverse hyperbolic tangent of its argument. The expression
ArcTanH(0.5) evaluates as 0.549306.
Character Functions
Character functions accept character arguments or return character strings and can convert the data
type of a value from numeric to character, or character to numeric.
Note: Character functions can result in either character or numeric data. If you calculate a data type
different than the one specified, the data type of the computed column is automatically changed to
match the result.
Note: When character functions result in character values greater than the Field Width in the Column
Info dialog, the length of the character field is automatically increased to hold the computed result.
Note: Arguments which are literal character strings must be enclosed in quotation marks.
174 8 Formula Editor Functions—Character Functions
Char
The Char function produces a character string that corresponds to the digits in its numeric argument.
For example,
Char(1.123)
evaluates as “1.123”.
Concat ||
The Concat function (shown as two vertical bars) concatenates character strings to produce a new
string with the function’s second character argument appended to the first. For example,
"Dr." || " " || name
produces a new string consisting of the title “Dr.” followed by a space and the contents of the name
string.
Contains
The Contains function returns the numeric position within the first argument of the first instance of
the second argument, if it exists. The second argument can contain one ore more characters. If the sec-
ond argument does not exist, Contains returns a zero. For example,
• Contains("Veronica Layman", "ay") evaluates as 11.
• Contains("Lillie Layman", "L") evaluates as 1.
Munger
The Munger function computes new character strings from existing strings by inserting or deleting
characters. It can also produce substrings, calculate indices, and perform other tasks depending on how
you specify its arguments.
Note: The Munger function treats uppercase and lowercase letters as different characters.
Text is a character expression. Munger applies the other three arguments to this string to compute a
result.
Offset is a numeric expression indicating the starting position to search in the string. If offset is greater
than the position of the first instance of the find argument, the first instance is disregarded. If offset is
greater than the search string’s length, the string’s length is the offset.
Find/Length is a character or numeric expression. Use a character string as search criterion, or use a pos-
itive integer to return that number of consecutive characters starting from the offset position. If you
specify a negative integer as the Length value, Munger returns all characters from the offset through to
the end of the string.
8 Formula Editor Functions—Character Functions 175
8 Functions
Replace (optional argument) can be a string or unspecified. If it is a string and the Find/Offset value is
numeric, Munger replaces the search criterion with the Replace string to form the result. If the Find/
Offset value is numeric and no string is specified, Munger calculates a substring. If the Find/Length
value is a character string, Munger always returns the numeric offset, disregarding the Replace value if
it exists. To insert the Replace argument, click on any argument in the Munger function and then
click on the insert button. Use the delete button to remove the Replace argument.
The following examples show uses of the Munger function. Assume a character column of names with
“Veronica Layman” as one of its values. To simplify the examples, the literal name “Veronica Layman” is
the search string instead of a column name.
Insert Characters
This Munger example finds the blank between the first and last name, and inserts the middle initial
“J.”. The formula
Munger("Veronica Layman", 1, " ", " J. ")
inserts the middle initial J., and evaluates as Veronica J. Layman.
Note: Double quotes are required by the Munger function for literal strings, including strings that con-
sists of a blank or when leading or trailing blanks are part of a string.
Delete Characters
To delete one or more characters from a string, designate the characters to delete as the find string in the
Munger function, and enter a replace string that is two quotation marks to indicate a null string. For
example, the function,
Munger("Veronica Layman", 1, "onic", "")
removes the “onic” from Veronica and evaluates as Vera Layman.
Note: A replace field with a null (no value) string enclosed in quotation marks is different from a
replace field with no value. If you delete the replace string altogether, Munger shows the argument
name ("replace") in the Formula Editor window and behaves as if that optional argument doesn't
exist, as in the next example. The resulting data type may also change from character to numeric,
depending on the value of the Find/Replace argument.
Find a substring
Munger can extract substrings. For example, to see the first name only
Munger("Veronica Layman", 1, 8,)
176 8 Formula Editor Functions—Character Functions
starts at position 1, reads through position 8, and ignores the remaining characters because the replace
function is not defined. This yields “Veronica.” This produces the same result as the Substring,
described later in this section.
An alternative way to find a substring is with a start value, any negative find value, and a no replace
argument. This example returns “Layman.”
Munger("Veronica Layman", 9, –1,)
Lowercase, Uppercase
The Lowercase function converts any upper case character found in its argument to the equivalent
lowercase character. For example, Lowercase("VERONICA LAYMAN") evaluates as “veronica layman”.
The Uppercase function converts any lower case character found in its argument to the equivalent
uppercase character. For example, Uppercase("Veronica Layman") evaluates as “VERONICA
LAYMAN”.
Length
The Length function calculates the length of its argument. For example, Length("Veronica") eval-
uates as 8.
Num
The Num function produces a numeric value that corresponds to its character string argument when
the character string consists of numbers only. If a character string contains a non-numeric value, the
result is a missing value. For example, Num(“1.123”) evaluates as the number 1.123.
Substr
The Substr function extracts the characters that are the portion of the first argument beginning at the
position given by the second argument and ending based on the number of characters specified in the
third argument. The first argument can be either a character column or a literal value. The starting
argument and the length argument can be numbers of expressions that evaluate to numbers.
For example, to show the first name only, Substr("Veronica Layman", 10, 6) starts at position
11 and reads through position 16, which yields “Layman.”
Trim
The Trim function produces a new character string from its argument, removing any trailing blanks.
For example, Trim("john ") evaluates as “john”.
8 Formula Editor Functions—Character Functions 177
8 Functions
Word
The Word function extracts the nth word from a character string. One or more spaces define where
each word begins and ends unless the optional delimiters argument is specified.
For example, Word(2, "Veronica Layman") returns "Layman".
To insert the delimiters argument, click on any argument in the Word function and then click on
the insert button on the keypad. Use the delete button to remove the delimiters argument.
Most special characters act as single delimiters. You can enter any character or set of characters to act as
an word delimiter. For example, to extract the last name in the following example, use a comma and
blank together as the delimiting characters and ask for the first word. The following function returns
the word “Layman”.
Word(1, "Layman, Veronica", ", ")
Item
The Item function is different than the Word function because of the way it treats word delimiters. If a
delimiter is found multiple times, or you enter a delimiter with multiple characters, the Word function
treats them as a single delimiter. The Item function uses each delimiter to define a new word position.
To compare, suppose a name is of the form lastname, firstname. The delimiter is a comma followed by
a blank.
Item(2, "Layman, Veronica", ", ")
Word(2, "Layman, Veronica", ", ")
The Item function returns a missing value because it treats the comma and blank separately and finds
nothing between them. The Word function treats the comma and blank as a single delimiter and finds
"Veronica" as the second word.
Hex
The Hex function returns the hex representation of its argument. If the argument is character (in
quotes), then the result is a character string twice as long containing the hexadecimal codes for the char-
acter values. For example, Hex("A") returns the string “41”.
If the argument is numeric, the Hex function returns an 8-hex-character representation of the integer returned. For
example, Hex(12) returns the string "0000000C" (Mac) or "0C000000"(PC).
178 8 Formula Editor Functions—Comparison Functions
Repeat
The Repeat function creates a string that is the first argument repeated the number of times specified
by the second argument. The first argument can be either a character literal, a character variable, or a
character expression. For example, Repeat(“Katie”, 3) creates “KatieKatieKatie”.
Note: The third argument applies when Repeat is used in a JSL script to repeat a matrix. When the
first argument is a matrix, the second argument is the row-wise repeat and the third argument is the
column-wise repeat.
Comparison Functions
Comparison functions compare the values of their arguments. Each comparison relationship evaluates
as true or false based on numeric magnitudes or character rankings.
Comparisons are most useful when you include them in conditional expressions, but they can also
stand alone as numeric expressions if neither term in comparison is missing. A true relationship evalu-
ates as 1, and false evaluates as 0.
A relational symbol’s arguments can be any two expressions. However, both arguments in a
Comparison function must be of the same data type. The relational symbols used in comparison
expressions are:
Note: JMP displays an error if you use a singe “=” in a conditional where “==” is expected.
8 Formula Editor Functions—Conditional Functions 179
8 Functions
The Formula Editor uses the International Utilities package when comparing character strings. This
package contains different rankings for each international character set and takes diacritical marks into
consideration.
Note: Do not use comparison operators to specifically compare to a missing value. If either term in a
comparison is missing, the result of the comparison is missing (not 0 or 1). Use the Is Missing func-
tion to detect a missing value.
Is Missing
The Is Missing function returns a one (1) if the value of the argument for the current row is missing,
and a zero if the value is not missing. The Formula Editor excludes missing numeric values from its sta-
tistical calculations.
Note: You can write a comparison function using the == function (see Comparison Functions, p. 178
earlier in this chapter) and compare a numeric variable to the numeric missing value (•), or a character
variable to a null string. The result is missing if either of the comparison arguments are missing.
Conditional Functions
JMP offers conditional expressions (called conditionals for short) in the Conditional functions list.
These expressions let you build a sequence of clauses paired with result expressions. Constructing a
sequence of clauses is the way you conditionally assign values to cells in a calculated column. With If and
Match, the Formula Editor searches down from the top of the sequence for the first true clause and
evaluates the corresponding result expression. Subsequent true clauses are ignored. For maximum effi-
ciency, list the most frequently evaluated clause/result pairs first in the sequence. When no clause is
true, the Formula Editor evaluates the result expression that accompanies the else clause. With Choose
the Formula Editor goes directly to the correct choice clause and evaluates the result expression.
To build a conditional expression, first select a type of conditional. Then, use the insert and delete
clause buttons on the Formula Editor panel to expand the expression. Each feature is described below.
Note: All result expressions in a conditional expression must evaluate to the same data type. A missing
term matches any data type. By definition, expressions that evaluate as zero are false. If an expression
180 8 Formula Editor Functions—Conditional Functions
evaluates as missing, no clauses are executed and missing is returned. All other numeric expressions are
true.
The following examples show how to use conditionals.
If
When you select If from the Conditions list, the Formula Editor shows a single If condition with a
missing expression and a missing then clause, as shown above. You can highlight either the expr or then
clause to enter an expression. When you highlight an expression in any formula and click the insert
button on the keypad, the Formula Editor creates a new conditional expression with the original high-
lighted expression as expr and a missing else clause.
To add a new condition to the If conditional, highlight then clause and click the insert button on the
keypad. Initially this changes the existing else condition to an expr clause. Click the insert button
again to add an else clause. Repetitively clicking then or else clauses alternately changes the else to
expr or adds a new expr clause. To delete a clause, select the then clause above it and click the delete
button on the keypad or the delete key on your keyboard.
A conditional expression is usually a comparison. However, any expression that evaluates as a numeric
value can be a conditional expression.
Note: By definition, expressions that evaluate as zero are false. If an expression evaluates as missing, no
clauses are executed and missing is returned. All other numeric expressions are true.
To enter a conditional expression, fill the missing terms with expressions. For example, to calculate
count as a percentage of total when total is not 0, enter the conditional expression (using columns
called count and total)
8 Functions
Match
When you select Match, the Formula Editor shows a single Match condition with a missing expression
and a missing then clause. You can highlight either expr, value, or the then clause to enter an expres-
sion. When you highlight an expression in any formula and click Match, the Formula Editor creates a
new Match conditional with the original highlighted expression as expr and nothing for the value and
else clause.
You add and delete clauses in a Match conditional the same way as in the If conditional described pre-
viously. To delete a clause, select the then clause above it and click the delete button on the keypad.
Match compares an expression to a list of clauses and returns the value of the result expression for the
first matching clause encountered. With Match, you provide the matching expression only once and
then give a match value for each clause. The Match expression is often just a column name, but can be
any valid formula expression.
For example, to assign character literals to age values in the Big Class data table, enter the conditional
expression
Note: Shift-click is a shortcut for building Match conditionals. Select the Match column in the column
selector list. Then hold down the Shift-key and select Match in the Conditional functions list. For
example, suppose you want a Match conditional for the nominal variable Type from the Hot dogs data
table, with values Beef, Meat, and Poultry. First choose the column Type from the column selector list.
Then with Type selected in the formula display, Shift-click on Match to see the following formula:
.
182 8 Formula Editor Functions—Conditional Functions
Note: Match evaluates faster and uses less memory than an equivalent If. See Efficiency, p. 158 in the
“Using the Formula Editor” chapter, for a comparison of Match and If conditionals.
Choose
Choose is a special case of Match in which the arguments of the condition are a sequence of integers
starting at 1. The value of clause replaces the match condition.
When you highlight an expression and click Choose, the Formula Editor creates a new conditional
expression with one clause. Use the insert and delete buttons on the keypad to add new clauses or
remove unwanted clauses, as described previously for the If conditional.
Choose evaluates the choose expression and goes immediately to the corresponding result expression
to generate the returned value. With Choose, you provide a choosing expression that yields sequential
integers starting at 1 only once, then you give a choice for each integer in the sequence.
For example, to assign character class values to protein to fat ratios in the Hot dogs table, which has rat-
ings 1, 2, and 3, enter the conditional expression
The logical functions (And, Or and Not) evaluate as either true or false, and are usually used in con-
junction with conditional statements. The next examples show how the logical functions evaluate.
And &
The And function evaluates as 1 when both of its arguments are true. Otherwise it evaluates as 0 (see
Figure 8.3). The formula
Or |
The Or function evaluates as 1 when either of its arguments is true. If both of its arguments are false,
then the Or expression evaluates as 0 (see Figure 8.3). The formula
8 Formula Editor Functions—Conditional Functions 183
8 Functions
assigns males and all participants over 13 to cabin 1.
Not !
When its argument is false, the Not function evaluates as 1. Otherwise, Not evaluates as 0.
When you apply the Not function, use parentheses where necessary to avoid ambiguity. For example,
!(weight==64) can be either true or false (either 1 or 0), but (!weight)==64 is always false (0)
because Not can only return 0 or 1. Expressions such as !(weight==64) can also be entered as
weight != 64.
The truth tables on the left in Figure 8.3 illustrate the results of the And ( &) and Or (|) functions when
both arguments have nonmissing values that evaluate to true or false. The table on the right illustrates
the result when either the left or right expression (call them a and b) or both have missing values.
Figure 8.3 Evaluations of And and Or Expressions
Note: The following statements (Interpolate, Step, For, and While) are most often used in conjunc-
tion with other commands to build a JSL script. You can use the column Formula Editor to create and
execute a script in that column, but this is not recommended because of dependencies and ambiguities
that can result. Most often, scripts are stored as .JSL files, and can be saved with a data table as a Table
Property. See the JMP Scripting Guide for documentation of all scripting commands.
Interpolate Step
The Interpolate function linearly interpolates the y value between two points, x1, y1 and x2, y2 that
corresponds to the arguments you give. You can insert additional pairs of X, Y arguments with the
insert key. Interpolate finds the pair of X, Y points that correspond to the X value and completes the
interpolation.
The Step function is like Interpolate except that it returns the Y value corresponding to the greatest X
value less than or equal to the X, Y arguments. That is, it finds the corresponding Y for a given X from
a step function rather than a linear fit between points. Like Interpolate, you can have as many X, Y
argument pairs as you want.
184 8 Formula Editor Functions—Probability Functions
For
For repeats the statements in the body argument as long as the while condition is true. The init and
next control the iterations.
while
While repeatedly tests the expr condition and executes the body until expr is no longer true.
Probability Functions
Probability functions calculate quantiles and probabilities for beta, chi-square, F, gamma, normal, Stu-
dent’s t, and Weibull distributions, and quantiles for the Tukey HSD.
The Beta Density function requires a quantile argument, and optionally accepts a shape parameter
alpha > 0, and a scale parameter, beta > 0. It returns the value of the Beta probability density function
(pdf ) for the given arguments. The beta density is sometimes useful for modeling the probabilistic
behavior of random variables such as proportions constrained to fall in the interval [0, 1]. Examples of
densities for several combinations of α and β are shown in Figure 8.4.
The Beta Distribution has a positive density only for an X interval of finite length, unlike normal and
gamma which have positive density over an infinite interval. The theoretical beta distribution has a
shape parameter, α > 0 and a scale parameter, β > 0, and constants A ≤ x ≤ B that define the interval for
which the distribution has values. The Beta Distribution function accepts the response variable argu-
ment x, whose range defines the interval for the distribution. The standard beta distribution occurs in
the interval [0, 1]. The Beta Distribution function is the inverse of the Beta Quantile function.
8 Formula Editor Functions—Probability Functions 185
8 Functions
The Beta Quantile function accepts a probability argument p, and shape and scale parameters, α > 0
and β > 0. It returns the pth quantile from the standard beta distribution. The Beta Quantile function
is the inverse of the Beta Distribution function.
Figure 8.4 Overlay Plot of Three Beta Density Curves
The ChiSquare Density function accepts a quantile argument from the range of values for the Chi
Square distribution, a degrees-of-freedom argument, and optionally, a noncentrality parameter. It
returns the value of the Chi Square density function (pdf ) for the arguments. Figure 8.5 shows the
shape of three Chi Square curves, with degrees of freedom 4, 6, and 10.
186 8 Formula Editor Functions—Probability Functions
Figure 8.5 Overlay Plot of Three Chi-Square Density Curves, DF=4, DF=6, DF=10
DF=4
DF=6
DF=10
The ChiSquare Distribution function accepts a response argument (range of X values) and three
parameter arguments: a quantile, a degrees-of-freedom, and a noncentrality parameter. It returns the
probability that an observation from the chi-square distribution with the specified noncentrality
parameter and degrees-of-freedom is less than or equal to the given quantile. For example, the expres-
sion
ChiSquare Distribution(11.264, 5)
returns the probability that an observation from the chi-square distribution centered at 0 with 5
degrees-of-freedom is less than or equal to 11.264. The expression evaluates as 0.95361.
The ChiSquare Distribution function accepts integer and noninteger degrees-of-freedom. It is centered
at 0 by default. The ChiSquare Distribution function is the inverse of the ChiSquare Quantile func-
tion.
The ChiSquare Quantile function accepts three arguments: a probability p, a degrees-of-freedom, and
a noncentrality parameter. It returns the pth quantile from the chi-square distribution with the speci-
fied noncentrality parameter and degrees-of-freedom. For example, the expression
ChiSquare Quantile(.95, 3.5, 4.5)
returns the 95% quantile from the chi-square distribution centered at 4.5 with 3.5 degrees-of-freedom.
The expression evaluates as 17.50458.
The ChiSquare Quantile function accepts integer and noninteger degrees-of-freedom. It is centered at
0 by default. The ChiSquare Quantile function is the inverse of the ChiSquare Distribution function.
The F Density function accepts a quantile argument from the range of values for the F distribution,
numerator and denominator degrees-of-freedom arguments , and optionally, a noncentrality parameter.
8 Formula Editor Functions—Probability Functions 187
8 Functions
It returns the value of the F density function (pdf ) for the arguments. Figure 8.6 shows the shape of
three F density curves, with degrees of freedom (5, 10), (10, 20), and (20, 50).
Figure 8.6 Overlay Plot of Three F Density Curves
F(20, 50)
F(10, 20)
F(5, 10)
The F Distribution function accepts four arguments: a quantile, a numerator and denominator
degrees-of-freedom, and a noncentrality parameter. It returns the probability that an observation from
the F distribution with the specified noncentrality parameter and degrees-of-freedom is less than or
equal to the given quantile. For example, the expression
F Distribution(3.32, 2, 3)
returns the probability that an observation from the central F distribution with 2 degrees-of-freedom in
the numerator and 3 degrees-of-freedom in the denominator is less than or equal to 3.32. The expres-
sion evaluates as 0.82639.
The F Distribution function accepts integer and noninteger degrees-of-freedom. By default it is cen-
tered at 0 and has 1 numerator degree of freedom. The F Distribution function is the inverse of the F
Quantile function.
The F Quantile function accepts four arguments: a probability p, a numerator and denominator
degrees-of-freedom, and a noncentrality parameter. It returns the pth quantile from the F distribution
with the specified noncentrality parameter and degrees-of-freedom. For example, the expression
F Quantile(0.95, 2, 10, 0)
returns the 95% quantile from the F distribution centered at 0 with 2 degrees-of-freedom in the
numerator and 10 degrees-of-freedom in the denominator. The expression evaluates as 4.1028.
The F Quantile function accepts integer and noninteger degrees-of-freedom. By default, it is centered
at 0 and has 1 numerator degree of freedom. The F Quantile function is the inverse of the F
Distribution function.
188 8 Formula Editor Functions—Probability Functions
Shape=1
Shape=3
Shape=5
8 Formula Editor Functions—Probability Functions 189
8 Functions
Normal Density, Normal Distribution, Normal Quantile
The Normal Density function accepts a quantile argument from the range of values for the standard
normal distribution. It returns the value of the standard normal probability density function (pdf ) for
the argument. For example, you can create a column of quantile values (X) with the formula
count(-3, 3, nrow()), and a second column computed as Normal Density(X) to generate den-
sity values. Then use Graph > Overlay to plot the normal density by X. Figure 8.8 shows an overlay
plot of normal density curves with various means and standard deviations.
The Normal Distribution function accepts a quantile argument from the range of values for the stan-
dard normal distribution with mean 0 and standard deviation 1. It returns the probability that an
observation from the standard normal distribution is less than or equal to the specified quantile. For
example, the expression Normal Distribution(1.96) returns 0.975, the probability that an observation
from the standard normal distribution is less than or equal to the 1.96th quantile. Optionally, you can
specify mean and standard deviation parameters to obtain probabilities from nonstandard normal dis-
tributions. The Normal Distribution function is the inverse of the Normal Quantile function.
The Normal Quantile (Probit) function accepts a probability argument p, and returns the pth quantile
from the standard normal distribution. For example, the expression Normal Quantile(0.975) returns
the 97.5% quantile from the standard normal distribution, which evaluates as 1.96. Optionally, you
can specify parameter values for the mean and standard deviation to obtain quantiles from nonstandard
normal distributions. The Normal Quantile function is the inverse of the Normal Distribution function.
Figure 8.8 Overlay Plots of Normal Density Curves with Mean and Std (0,1) (1,1) and (0.5,0.75)
Normal(1,1)
Normal(0,1) Normal(0.5,0.75)
190 8 Formula Editor Functions—Probability Functions
The t Density function accepts a quantile argument from the range of values for the t distribution, a
degrees-of-freedom argument, and optionally, a noncentrality parameter. It returns the value of the t
density function (pdf ) for the arguments. To compare a t density with 5 df with a standard normal dis-
tribution you can create a column of quantile values (X) with the formula count(-3, 3, nrow()), a
second column computed as t Density(X), and a third column computed as Normal Density(X). Then
use Graph > Overlay to plot the t density and the normal density by X to see the plot shown in
Figure 8.9. You can see that the t density has slightly more spread than the normal.
The t Distribution function accepts three arguments: a quantile, a degrees-of-freedom, and a noncen-
trality parameter. It returns the probability that an observation from the Student’s t distribution with
the specified noncentrality parameter and degrees-of-freedom is less than or equal to the given quantile.
For example, the expression t Distribution(.9, 5) returns the probability that an observation
from the Student’s t distribution centered at 0 with 5 degrees-of-freedom is less than or equal to 0.9.
The expression evaluates as 0.79531.
The t Distribution function accepts integer and noninteger degrees-of-freedom. It is centered at 0 by
default, but you can enter a value for the noncentrality parameter. The t Quantile function is the
inverse of the t Distribution function.
The t Quantile function accepts three arguments: a probability p, a degrees-of-freedom, and a noncen-
trality parameter. It returns the pth quantile from the Student’s t distribution with the specified non-
centrality parameter and degrees-of-freedom. For example, the expression Student’s t Quantile(.95, 2.5)
returns the 95% quantile from the Student’s t distribution centered at 0 with 2.5 degrees-of-freedom.
The expression evaluates as 2.558219. The t Quantile function is the inverse of the t Distribution
function.
The t Quantile function accepts integer and noninteger degrees-of-freedom. It is centered at 0 by
default, but you have the option to enter a value for the noncentrality parameter. The t Distribution
function is the inverse of the t Quantile function.
8 Formula Editor Functions—Probability Functions 191
8 Functions
Figure 8.9 Comparison of Normal Density and t Density
normal density
t density
The Weibull Density function accepts a quantile argument from a range of values for the Weibull dis-
tribution. It returns the value of the Weibull probability density function (pdf ), which is the probabil-
ity that an observation from a Weibull distribution is less than or equal to the specified quantile
argument.
The Weibull distribution has different shapes depending on the values of α (a scale parameter that
affects the x direction) and β (a shape parameter). It often provides a good model for estimating the
length of life, especially for mechanical devices and in biology. The two-parameter Weibull is the same
as the three-parameter Weibull with a threshold of zero. Figure 8.10 shows examples of several combi-
nations of shape and scale parameter values, with the threshold parameter equal to zero. The Extreme
Value distribution is a two-parameter Weibull (α, β) distribution with the transformed parameters:
δ = 1 / α and γ = ln(β).
The Weibull distribution has two parameters, α > 0 and β > 0. When α = 1 the pdf reduces to the
exponential distribution (with γ = 1 / β). The exponential distribution is a special case of both the
gamma and Weibull distributions. However, there are Weibull distributions that are not exponential
distributions.
The Weibull Distribution function uses an argument with a quantile value and, optionally, a value for
the scale parameter α and the shape parameter β. It returns the probability that an observation is less
than or equal to the specified X for Weibull distribution with the shape and scale parameters you speci-
fied The Weibull Distribution function is the inverse of Weibull Quantile function.
192 8 Formula Editor Functions—Probability Functions
The Weibull Quantile function accepts a probability argument p, and returns the pth quantile from the
Weibull distribution with the shape and scale parameters you specify. The Weibull Quantile function is
the inverse of the Weibull Distribution function.
Figure 8.10 Comparison of Weibull Curves
Weibull(2,1)
Weibull(1,1)
Weibull(1,2)
The binomial functions accept arguments that are the probability of success p (the event of interest),
the number of trials n, and the number of successes k.
The Binomial Distribution function returns the probability that an observation from a binomial distri-
bution with parameters p and n is less than or equal to k.
The Binomial Probability function computes the probability that a random variable from a binomial
distribution is equal to p.
The negative binomial functions accept an argument that is the probability of success p, a number of
trials n, and the count of interest k, which is the number of failures that precede the ith success.
The NegBinomial Distribution function computes the probability that a random variable from a nega-
tive binomial distribution is less than or equal to p.
The NegBinomial Probability function computes the probability that a random variable from a nega-
tive binomial distribution is equal to p.
8 Formula Editor Functions—Probability Functions 193
8 Functions
Hypergeometric Distribution, Hypergeometric Probability
The hypergeometric distribution models the total number of successes in a fixed sample drawn without
replacement from a finite population. The hypergeometric functions accept as arguments the size of the
population N, the total number of items with the desired characteristic in the population, K, the num-
ber of samples drawn n, and the number of successes in the sample x.
Hypergeometric Distribution computes the probability that a random variable from a hypergeometric
distribution is less than or equal to x.
Hypergeometric Probability computes the probability that a random variable from a hypergeometric
distribution is equal to x.
The Poisson functions accept an argument that is the count of interest, and lambda, the mean param-
eter.
Poisson Distribution computes the probability that a random variable from a Poisson distribution
with mean lambda is less than or equal to the count of interest.
Poisson Probability computes the probability that a random variable from a Poisson distribution with
mean lambda is equal to the count of interest.
The Tukey HSD Quantile function accepts a probability argument p, and returns the pth quantile from
Tukey’s HSD test for the parameters you specify. The alpha argument is the significance level you want,
nGroups is the number of groups in a study, and dfe is the error degrees-of-freedom (based on the total
study sample). This is the quantile used to calculate least significant difference in Tukey’s multiple com-
parisons test.
F Power calculates the power from a given situation that involves an F test or t test and F Sample
Size computes the sample size. The arguments are the values you specify for computation of a prospec-
tive power analysis.
alpha
is the significance level you are willing to tolerate (often 0.05).
dfh
is the hypothesis degrees-of-freedom. It is one (1) for a t test.
dfm
194 8 Formula Editor Functions—Statistical Functions
n
(in the F Power function) is the total number of observations (runs, experimental units, or samples)
you expect to have.
(in the F Sample Size function) is the probability you want to have of declaring a significant result.
These functions perform the same computations as the Sample Size, Power command found in the
DOE main menu. See the chapter “Prospective Power and Sample Size,” p. 139 of JMP Design of Exper-
iments, for a discussion of power and sample size.
Statistical Functions
8 Functions
• The Mean, Std Dev, Number, Sum, Quantile, Maximum, Minimum, and N Missing functions
evaluate across columns or arguments. The statistic is computed for each row across the series of
arguments. You can use the insert key on the Formula Editor panel, or type a comma to add argu-
ments to the functions that accept multiple arguments. When there are multiple contiguous argu-
ments, select the function and the first argument, then Shift-click the last argument in the group.
These functions then automatically shows with the complete list.
The following paragraphs give descriptions and examples of the Statistical functions.
Col Mean
The Col Mean function calculates the mean (or arithmetic average) of the numeric values identified by
its argument. The formula Mean(age) calculates the average of all nonmissing values in the age col-
umn.
The Col Std Dev function measures the spread around the mean of the distribution identified by its
argument. In the normal distribution, about 68% of the distribution is within one standard deviation
of the mean, 95% of the distribution is within two standard deviations of the mean, and 99% of the
distribution is within three standard deviations of the mean.
Col Number
The Col Number function counts the number of nonmissing values in the column you specify. A miss-
ing numeric value occurs when a cell has no assigned value or as the result of an invalid operation (such
as division by zero). Missing values show on the spreadsheet as a missing value mark (•). Missing char-
acter values are null character strings.
In formulas for row state columns, an excluded row state characteristic is treated as a missing value. The
Formula Editor interprets other missing values according to their data types.
Col NMissing
The Col Nmissing function counts the number of missing values in the column you specify. A missing
numeric value occurs when a cell has no assigned value or as the result of an invalid operation (such as
division by zero). Missing values show in the data grid with a missing value character •. Missing charac-
ter values are null character strings.
196 8 Formula Editor Functions—Statistical Functions
Col Sum
The Col Sum function computes the sum of the values in its numeric argument. Missing values are
ignored.
The Col Minimum function takes the minimum of its numeric arguments. Col Minimum ignores miss-
ing values.
The Col Maximum function takes the maximum of a numeric column argument, and ignores missing
values.
Col Quantile
The Col Quantile function computes the value at which a specific percentage of the values is less than
or equal to that value. For example, the value calculated as the 50% quantile, also called the median, is
greater than or equal to 50% of the data. Half of the data values are less than the 50th quantile.
The Col Quantile function’s quantile argument represents the quantile percentage divided by 100. The
25% quantile, also called the lower quartile, corresponds to p = 0.025, and the 75% quantile, called the
upper quartile, corresponds to p = 0.075. The default value of p is 0.5 (the median).
The Formula Editor computes a quantile for a column of N nonmissing values by arranging the values
in ascending order. The subscripts of the sorted column values, y1, y2,...,yn, represent the ranks in
ascending order.
The pth quantile value is calculated using the formula p(N + 1) where p is the quantile and N is the
total number of nonmissing values. If I is an integer, then the quantile value is yp = yi. If I is not an inte-
ger, then the value is interpolated by assigning the integer part of the result to i, and the fractional part
to f and by applying the formula qp = (1 – f )yi + (f )yi+1.
For example, suppose a column has values 2, 4, 6, 8, 10, 12, 14, 16, 18, and 20. The 50% quantile is
calculated as 0.5(10 + 1) = 5.5.
Because the result is fractional, the 50% quantile value is interpolated as
(1 – 0.5) x 10 + (0.5) x 12 = (0.5)10 + (0.5)12 = 6 + 5 = 11
The following are example Quantile formulas:
Quantile(age, 1) calculates the maximum age.
Quantile(age, 0.75) calculates the upper quartile age.
Quantile(age, 0.5) calculates the median age.
Quantile(age, 0.25) calculates the lower quartile age.
Quantile(age, 0) calculates the minimum age.
8 Formula Editor Functions—Statistical Functions 197
8 Functions
The Quantile argument can be any expression that evaluates to a value between (and including) 0 and
1. For example, the first formula Figure 8.11 calculates quantile values of age in axcending order for
each row. The column then contains the interpolated values of age in ascending order in the calculated
column. The second formula lists the interpolated values of age in descending order.
Figure 8.11 Examples of the Quantile Function
Col Standardize
Col Standardize performs the usual standardization on its numeric expression. For each row i,
Col Standardize(height) is (HeightRow()–Col Mean(Height))/Col Standardize(Height).
Mean
Mean calculates the arithmetic average of the nonmissing arguments you specify. The arguments can be
constants, numbers, or expressions. The Mean function initially shows with a single argument. You add
arguments with the insert button or by typing a comma.
Std Dev
Std Dev computes standard deviation of the list of arguments you specify. The arguments can be con-
stants, numbers, or expressions. The Std Dev function initially shows with a single argument. You add
arguments with the insert button or by typing a comma.
Number
Number counts the number of nonmissing values in the list of arguments you specify.
Sum
Quantile
m
Quantile calculates the quantile given by its first argument for all the following arguments given.
198 8 Formula Editor Functions—Statistical Functions
The Summation and Product functions always evaluate for an explicit range of values in a column, as
given by the summation indices. All other statistical functions always evaluate on every row.
The Summation function uses the summation notation shown in Figure 8.12. To calculate a sum,
replace the missing body term with an expression containing the index variable i, or an index variable
you assign. Summation repeatedly evaluates the expression for i = 1, i = 2, through i = NRow() and then
adds the nonmissing results together to determine the final result.
You can replace NRow(), the number of rows in the active spreadsheet, and the index constant, i, with
any expression appropriate for your formula. For example, the second summation formula in
Figure 8.12 computes the total for each row of all revenue values for rows 1 through the current row
number, filling the calculated column with the cumulative totals of the revenue column.
The Product function uses the notation shown in the formulas on the right in Figure 8.12. To calculate
a product, replace the missing body term with an expression containing the index variable j. Product
repeatedly evaluates the expression for i = 1, i = 2, through i = n and multiplies the nonmissing results
together to determine the final result.
You can replace NRow(), the number of rows in the active spreadsheet and the index constant, i, with
any expression appropriate for your formula.
For example, the expression second product example in Figure 8.12 calculates i! (each row number’s
factorial).
Figure 8.12 Examples of Summation and Product Functions
Minimum and Maximum return the minimum and maximum value, respectively, from the list of non-
missing arguments you specify.
N Missing
The N Missing function counts the number of missing values in the list of arguments you specify.
8 Formula Editor Functions—Random Number Functions 199
8 Functions
Random Number Functions
Random Number functions generate real numbers by effectively “rolling the dice” within the con-
straints of the specified distribution. Each time you click Apply in the Formula Editor window, these
functions produce a new set of random numbers. This section describes the random functions.
Note: Random numbers are generated using the Mersenne-Twister technique. This technique has a
period length of 219937-1 (as opposed to 231-1 for the generator used in MP 4.03 and earlier). Details
on the generators is found in Matsumoto and Nishimura (1998). The new generators are verified to
pass all the DIEHARD tests as documented in Marsaglia (1996).
The previously used routines are still available as RandomSeededUniform() and
RandomSeededNormal() and work with the RandomSeed() function call. For compatibility, Version
3 data tables with Random Uniform() and Random Normal() functions are converted to the new
names for the old functions, RandomSeededUniform() and RandomSeededNormal().
The seeded functions are important then you want the same series of random numbers generated
repeatedly for testing.
Random Uniform
Random Uniform generates random numbers uniformly between 0 and 1. This means that any num-
ber between 0 and 1 is as likely to be generated as any other. The result is an approximately even distri-
bution. You can shift the distribution and change its range with constants. For example,
5 + Random Uniform()*20 generates uniform random numbers between 5 and 25.
Random Normal
Random Normal generates random numbers that approximate a normal distribution with a mean of 0
and variance of 1. The normal distribution is bell shaped and symmetrical. You can modify the Normal
function with constants to specify a normal distribution with a different mean and standard deviation.
For example, the formula Random Normal()*5 + 30 generates a random normal variable with a
mean of 30 and a standard deviation of 5.
200 8 Formula Editor Functions—Random Number Functions
Random Exp
Random Exp generates a single parameter exponential distribution for the distribution parameter
lambda=1. You can modify the Exponential function to use a different lambda. For example,
0.1*Random Exp()-0.1 generates an exponential distribution for lambda=0.1. The exponential dis-
tribution is often used to model simple failure time data, where lambda is the failure rate.
Random Gamma
Random Gamma gives a gamma distribution for the parameter, alpha, you enter as the function
argument. The gamma distribution describes the time until the kth occurrence of an event. The gamma
distribution can also have a scale parameter, beta. A gamma variate with shape parameter alpha and
scale beta can be generated with the formula beta*Random Gamma(alpha). If 2*alpha is an integer,
a chi-square variate with 2*alpha degrees-of-freedom is generated with the formula
2*Random Gamma(alpha).
Random Cauchy
Random Cauchy generates a Cauchy distribution with location parameter 0 and scale parameter 1.
The Cauchy distribution is bell shaped and symmetric but has heavier tails than the normal distribu-
tion. A Cauchy variate with location parameter alpha and scale parameter beta can be generated with
the formula alpha+beta*Random Cauchy().
Random Triangular
Random Triangular generates a triangular distribution of numbers between 0 and 1, with the midpoint
you enter as the function argument. You can add a constant to the function to shift the distribution,
and multiply to change its span.
Random Integer
Random Integer generates a uniform distribution of integers between 1 and the argument you enter as
n1, if nothing is entered for n2. If you enter both n1 and n2 (n1<n2), Random Integer generates a uni-
form distribution of the integers between and including n1 and n2.
8 Formula Editor Functions—Random Number Functions 201
8 Functions
Random Binomial
Random Binomial generates random numbers from a binomial distribution with parameters you enter
as function arguments. The first argument is n, the number of trials in a binomial experiment. The sec-
ond argument is p, the probability that the event of interest occurs. When n is 1, the binomial function
generates a distribution of Bernoulli trials. For example, n =1 and p = 0.5, give the distribution of toss-
ing a fair coin. The mean of the binomial distribution is np, and variance is np(1 – np).
Random Negative Binomial generates a negative binomial distribution for the parameters you enter as
function arguments. The first parameter is the number of successes of interest ( r) and the second argu-
ment is the probability of success (p). The random variable of interest is the number of failures that pre-
cede the ith success. In contrast to the binomial variate where the number of trials is fixed and the
number of successes is variable, the negative binomial variate is for a fixed number of successes and a
random number of trials. The mean of the negative binomial distribution is (r(1 – p))/p and the vari-
ance is (r(1 – p))/p2.
Random Geometric
Random Geometric returns random numbers from the geometric distribution with the parameter you
enter as the function argument. The parameter, p, is the probability that a specific event occurs at any
one trial. The number of trials until a specific event occurs for the first time is described by the geomet-
ric distribution. The mean of the geometric distribution is 1 / p, and the variance is (1 – p) / p2.
Random Poisson
Random Poisson generates a Poisson variate based on the value of the parameter, lambda, you enter as
the function argument. Lambda is often a rate of events occurring per unit time or unit of area.
Lambda is both the mean and the variance of the Poisson distribution.
Random Seed
This function lets you start a random number sequence with a seed you specify. To use the Random
Seed function, assign its argument and use it in an If conditional function. Then use the random
number function you want as the else clause. This example uses the number 1234567 as the seed to
generate a sequence of uniform random numbers.
202 8 Formula Editor Functions—Random Number Functions
These functions generate uniform and normal random numbers with the random number generator
used in JMP 3. However, for compatibility, JMP 3 data tables with Random Uniform() and Random
Normal() functions are converted to the new names for the old functions, RandomSeededUniform()
and RandomSeededNormal. The random numbers in JMP 5 are generated using the
Mersenne-Twister technique. See Matsumoto and Nishimura (1998) in the reference section of the
JMP Statistics and Graphics Guide for details.
Col Shuffle
Col Shuffle selects a row number at random from the current data table. Each row number is selected
only once. When Col Shuffle is used as a subscript, it returns a value selected at random from the col-
umn that serves as its argument. Each value from the original column is assigned only once as
Shuffle’s result.
For example, to identify a 50% random sample without replacement, use the following formula:
This formula chooses half the values (n / 2) from the column X and assigns them to the first half of the
rows in the computed column. The remaining rows of the computed column fill with missing values.
8 Formula Editor Functions—Date Time Functions 203
8 Functions
Date Time Functions
JMP stores dates and times in numeric columns, using the Macintosh standard of the number of seconds
since Jan 1, 1904. When a column has date values, you can assign a date format to that column using
the Date Time format popup menu in the Column Info dialog so that they display in a familiar form.
See the chapter JMP Data Tables, p. 93, for more details about using dates.
The Formula Editor supports JMP dates with the following Date Time functions:
Each of the functions listed above converts from the units of the function name to the equivalent num-
ber of seconds for the argument. The argument must be a number or numeric expression. For example,
In Minutes(2) yields 120, and In Years(1) yields 31,557,600 (60 seconds * 60 minutes * 24
hours * 365.25 days).
These Date DMY and Date MDY functions accepts numeric expressions for day, month, and year and
return the associated JMP date. For example, Date DMY (20, 3, 1991) and Date MDY(3, 20,
1991) evaluate to 2,752,272,000.
Today
The Today function returns the number of seconds between January 1, 1904 and the current date. For
example, at midnight on March 20, 1991 (a Wednesday), the Today function returns 2752272000,
(2,752,272,000 seconds) and continues counting. If you evaluate the Today function later in the day, it
reflects the additional seconds.
204 8 Formula Editor Functions—Date Time Functions
The argument for the Day, Month, and Year functions is interpreted as a JMP date. These functions
returns the day of the month, the month as a number from 1 to 12, and a four-digit year, respectively.
For Example, on March 20, 1991,
• Day(2752272000) returns the number 20.
• Month(2752272000) returns the number 3.
• Year(2752272000) returns the nubmer 1991.
The argument for Day Of Week, Day Of Year, and Week Of Year functions is a JMP date. Day Of
Week returns a number from 1 to 7 where 1 represents Sunday, Day Of Year returns the number of
days from the beginning of the year, and Week Of Year returns a number from 1 to 52.
For example, on Wednesday, March 20, 1991,
• Day Of Week(2752272000) returns the number 4.
• Day Of Year(2752272000) returns the number 79.
• Week Of Year(2752272000) returns the number 12.
Informat
The argument for the Informat function is a date character string. For example, Informat("03/20/
1991") returns the appropriate JMP date value, 2752272000. JMP can read all the date formats
except for Abbrev Date and Long Date.
The argument for these date functions is a JMP date. They return character strings that are the format-
ted representation of the argument.
For example,
• Abbrev Date(2752272000) returns “Wed, Mar 20,1991”,
• Long Date(2752272000) returns “Wednesday, March 20, 1991”,
• Short Date(2752272000) returns “3/20/91”.
8 Formula Editor Functions—Row State Functions 205
8 Functions
Format
The first argument in the Format function is a JMP date. This function returns the character string
representation of the date by the date format you specify in the second argument. If you apply this for-
mula to a numeric column, JMP automatically changes the column’s data type to character.
MDYHMS
The argument of MDYHMS is a JMP date. This function shows all date and time fields, appending
zeros as time fields if no time information is present. This can be useful if a date column is formulated
such that not all date information is displayed. The MDYHMS function can be used to see all available
date and time information.
Formulas process row state data just as they process character and numeric data. There are six row state
conditions: selected, hidden, excluded, labeled, colored, and marked. You assign row states to rows
using commands in the Rows menu. You can also create row state data table columns and use Row
State functions to compute and save row state conditions. Row states identify rows in the following
ways:
• Select rows to identify data applicable to JMP commands.
• Hide rows to hide points in report displays.
• Exclude rows to eliminate data from analysis calculations and displays.
• Label rows to identify points in plots.
• Use colors to distinguish points in report displays.
• Use markers to distinguish points in report displays.
206 8 Formula Editor Functions—Row State Functions
Note: A row can be assigned any combination of row states; a row state column can have multiple row
states as a value. Table 8.2 “Summary of State Functions,” p. 209 at the end of this section summarizes
the Row State functions by listing the type of argument each requires, and what each function returns.
The Function Browser has the following Row State functions.
Row State
The Row State function returns the active row state condition of the current row as true or false. You
can use this function to conveniently write conditional clauses that depend on the status of the current
row. For example,
assigns a 1 to rows that are currently selected and labeled and a zero otherwise.
As Row State
The As Row State function converts a numeric argument to a row state or set of row state conditions.
Row states are stored internally in JMP as a 16-bit number, with each bit assigned to represent one of
the possible row states. For example, the binary representation of 1327 is 0000010100101111. As
Row State(1 3 2 7) would therefore set the row state as Selected, Excluded, Hidden, Labeled, with
marker 2 and color 10.
Combine States
The Combine States function generates a row state combination with two or more arguments. Use
the Insert button on the keypad or type a comma to add arguments to the Combine States function.
The currently selected expression becomes the first argument when you choose Combine States.
Replace each argument with an expression that evaluates to a row state.
8 Formula Editor Functions—Row State Functions 207
8 Functions
The formula
Combine States(
Selected State(Modulo(Row(),2),
Labeled State(Modulo(Row()+1,2))
alternately labels or selects each row in the calculated row state column. The Selected State and
Labeled State functions are defined later in this section.
Use the insert and delete buttons in the Formula Editor panel to add more arguments or remove
unwanted arguments.
Note: If you include conflicting row states in a combination, the results are unpredictable.
These Row State functions are used in a column that has Row State as its data type. They interpret a
numeric argument as true or false.
• When an argument evaluates as true, the Excluded State function assigns the excluded condition as
the value of the column for that row. For example, Excluded State(Modulo(Row(),2), assigns
the excluded row state as the value of the row state column for each odd numbered row.
• The Hidden State function assigns the hidden row state condition when its argument is greater
than zero. If the argument is zero the value in the column for that row is not hidden.
• The Labeled State function gives the labeled row state condition when its argument is greater than
zero. If the argument is zero the row value in the column for that row is not labeled.
• The Selected State function gives the selected row state condition when its argument is greater
than zero. If the argument is zero, the value in the column for that row is not selected.
Color State
The Color State function returns the color from the JMP Color Map that corresponds to its integer
argument. JMP colors are numbered 0 through 12, but larger integers map to color indices are treated
as modulo 12. Zero maps to black.
Marker State
The Marker State function returns markers from the JMP Marker Map that correspond to its integer
argument. JMP markers are numbered 0 through 16, but larger integers that map to marker indices are
treated as modulo 8.
The formula
Marker State(Row())
assigns all the row state markers in a repeating sequence determined by the current row number to the
calculated row state column.
208 8 Formula Editor Functions—Row State Functions
Note: A row state column can have multiple row states as a value.
The Hue State function is used with a Row State data type column. The Hue State function returns
the color from the JMP hue map that corresponds to its integer argument. JMP hues are numbered 0
through 11 but larger integers are treated as modulo 12. The Hue State function does not map to
black, gray, or white. A Hue of 0 maps to red and Hue of 11 maps to magenta. The formula on the left
in Figure 8.14 assigns row state colors in a chromatic spread based on the value of z.
The Shade State function assigns 5 shade levels to a color or hue. A shade of –2 is darkest and shade of
+2 is lightest. A shade of zero is a pure color. The formula on the right in Figure 8.14 assigns shade val-
ues based on the value of z.
Figure 8.14 Examples of Hue and Shade Functions
However, to assign all shades of all the colors in the colors palette, you need to use the Hue State and
Shade State assignments together. The formula in Figure 8.15 uses the Combine States function
described at the beginning of this section. The first argument in the Combine State function is the
Hue State formula shown previously, and the second argument is the Shade State formula. In addi-
tion, the Marker State function with an argument of 2 assigns the X marker to each row, and the
Selected State function with an argument of 1 selects each row.
Figure 8.15 Combine States Example For Using Both Hue State and Row State
The Excluded, Hidden, Labeled, and Selected functions accept a row state expression argument
(row state column or row state constant) that evaluates as either 1 or 0 (true or false). These characteris-
tics are inactive by default. Often, the Row() function is the argument, which detects the active row
8 Formula Editor Functions—Row State Functions 209
8 Functions
state condition of each row. For example, in a numeric column this formula assigns 99 whenever a row
is actively selected, and 0 otherwise.
Another example assigns row state conditions to a row state column. The formula for the row state col-
umn (in the columns called row state col) in Figure 8.16 checks to see if the active row state is either
Hidden or Excluded, and if so, assigns the Labeled row state.
Figure 8.16 Calculate Row state Information in a Row State Column
The Color Of function accepts any row state expression or column, or the Row State() function as its
argument and returns a number from the JMP color map that corresponds to the active color state, or a
missing value if there is no assigned color.
The Marker Of function accepts any row state expression or column, or the Row State() function as
its argument and returns a number from the JMP marker map that corresponds to the active marker or
a missing value if there is no assigned marker.
Assignment Functions
assign
add to
subtract to
multiply to
divide to
post increment
post decrement
Assignment functions work in place. That is, the result returned by the operation (on the right of the
operator) is stored in the argument on the left of the operator and replaces its current value. Assignment
statements are most often used in conjunction with other commands to build a JSL script. You can use
the column Formula Editor to create and execute a script in that column, but this is not recommended
because of dependencies and ambiguities that can result. Most often, scripts are stored as .JSL files, and
can be saved with a data table as a Table Property. See the JMP Scripting Guide for documentation of all
scripting commands.
Note: The first argument of an assignment function must be capable of being assigned. This means you
cannot have an assignment such as 3+=4, because 3 is a constant value that cannot be reassigned. You
must first create a variable (a table variable or column) whose value is 3. Then use that variable as the
left-hand argument of the assignment function.
assign function (a=b) puts the value of b into a.
add to function (a+=b) adds the value of b to a and puts the result back into a.
8 Formula Editor Functions—Assignment Functions 211
8 Functions
subtract to function (a–=b) subtracts the value of b and puts the result back into a.
multiply to function (a*=b) multiplies b with a and puts the result back into a.
divide to function (a/=b) divides b into a and puts the result back into a.
post increment function (a++) adds one (1) to a, in place. For example, if the initial value of a is 4, the
expression a++ changes a to 5.
post decrement function (a– – ) subtracts one (1) from a, in place. For example, if the initial value of
a is 4, the expression a– – changes a to 3.
212 8 Formula Editor Functions—Assignment Functions
9
9 JMP Reports
Report Windows and Surface Features
This chapter describes features common to all report windows in JMP. These include standard window
controls and special additional JMP features.
To make the most of each analysis, all report windows have button controls and popup menus. In addi-
tion, you can select from tools that change the function of the mouse. This puts report features literally
at your fingertips.
You can customize the appearance of report windows by changing text styles, adding colors and labels
to plots, and more. Once you’ve explored the report windows on your monitor, you can print them
directly from JMP. If you want to include analyses in your own comprehensive report, you can journal
report windows to word-processing files or copy any part of a report window into another application.
9
Contents
Report Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Standard Window Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Report Disclosure Icons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Popup Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Cursor Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Features and Formatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Formatting Analysis Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Context Menu Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Resizing Plots and Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Selecting Points in Plots and Using Row States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Customizing Axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Using the Annotate Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
Using the Draw Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Copy, Paste, Drag and Drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
Journal and Layout Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
The JMP Journal Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
The JMP Layout Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
9 Report Windows and Surface Features—Report Windows 215
9 JMP Reports
Report Windows
All analysis platforms produce report windows. These windows contain both text reports and graphical
displays. They are designed to be convenient in several ways:
• Although report windows consist of complex plots and calculations, they appear on the screen
almost instantaneously most of the time.
• Individual plots can be resized and text reports can be closed to optimize the use of space on the
monitor.
• Report windows always print each individual plot and text report completely on one page whenever
possible.
• Many areas of report windows have context-sensitive help.
• Report windows can be displayed in a journal window or a layout window. A layout window lets
you ungroup reports and graphs and rearrange them.
All or part of any report window can be copied to other applications or printed from JMP.
vertical
scroll bar
report
disclosure
buttons
horizontal
resize
vertical resize
horizontal scroll bar window resize box
216 9 Report Windows and Surface Features—Report Windows
open or
close right-click disclosure
reports button to see menu
JMP reports are organized in a hierarchical outline. Each level of the outline has a diamond-shaped dis-
closure button like those shown above. Click a disclosure button to alternately open and close the corre-
sponding graphs or reports for that level of the report outline. When you close an outline level, all its
sublevels are also hidden.
If you right-click on a disclosure button (open or closed), the menu shown above appears. These com-
mands let you rearrange the analysis report and give you control over report outline levels. You can
open or close multiple levels instead of managing them one at a time.
Horizontal
By default, most reports are arranged with the outline levels showing vertically with indented outline
levels. The Horizontal command attempts to horizontally align the next lower outline levels.
Open All Below, Close All Below
Open and Close All Below open or close all outline levels beneath the level where this command is
selected, including that level.
Open All Like This, Close All Like This
If you analyze several variables at a time, or select multiple options on a single analysis, you often have
many of the same type report tables. You can select the Open or Close All Like This command to open
or close all of the same type of report present in the analysis window. For example, a distribution for
several continuous variables produces a Moments table and a Quantiles table for each variable. Select-
ing the Close All Like This command on any Moments table disclosure button closes all the Moments
tables.
Close Where No Outlines
This command is usually used at the top level of the report outline. It closes all parts of the report that
do not have sublevels. This is a quick way to see a nesting structure overview of a report.
DisplayBox
DisplayBox has the submenu shown above, with commands that affect all reports at the outline level
where they are used.
• Select and Deselect select and deselect all reports for that outline level. You can perform the same
function with the area selection tool (large plus) in the Tools menu.
9 Report Windows and Surface Features—Cursor Tools 217
9 JMP Reports
• Journal creates a journal window of all reports at that outline level, whether or not they are selected.
The Journal command in the Edit menu does the same thing, but only on selected report areas.
• Copy Picture copies the entire report at that outline level to the clipboard as a single picture.
• Page Break inserts a page break for printing purposes. The page break appears as a gray line on the
report.
• Set Report Title displays a text box to type a new title for the entire report window. Under Win-
dows, you can do the same thing with the Set Title command in the Window menu.
Popup Menus
Many report outline levels have popup menus indicated by a red downward pointing triangle. Each
popup menu icon displays a list of options that apply for that particular report.
Popup menus have several modes that give the same commands in different ways:
• Click the menu icon and select a command to affect a single analysis.
• If you Alt-click (Option-click on the Macintosh) the popup icon, a panel of all commands and
options appears with check boxes so that you can select multiple actions at the same time.
• Control-click (Command-click on the Macintosh) the popup icon on any analysis to broadcast the
command you select to all like reports in the analysis window. For example, for a bivariate analysis,
if you Control-click and select Means/Anova/t Test, an analysis of variance is performed for all
bivariate analyses in the active report window.
Chapters in the Statistics and Graphics Guide describe the popup menus for each statistics and graphics
platform.
Cursor Tools
To expand the capabilities of the mouse in report windows, you can choose special mouse cursors from
the Tools menu or palette. Mouse actions have different effects depending on which tool you choose
and where you use it.
There are general tools and specific graphics tools that affect points in plots and histogram bars. The
default tool is the arrow. The tools are briefly described here, with more discussion and examples in the
section Features and Formatting, p. 220, later in this chapter.
218 9 Report Windows and Surface Features—Cursor Tools
Help Tool
The question mark icon accesses JMP Help. Select the help tool and then click graphs, plots, or tables
to access specific help topics in the JMP Help system. JMP uses the standard Windows or Macintosh
Help systems. For details about the Windows Help system, consult a Microsoft Windows User’s Guide.
Macintosh users should consult the Macintosh User’s Guide for the Macintosh OS 8.6 or later operat-
ing system.
The fat plus is for selecting rows and columns in the data table, or components of analysis tables or
graphs. When you click on a section of a plot or report, that section highlights and is selected. To
extend the selection, drag the selection tool or Shift-click on other report elements. Clicking near the
edge of the report window selects the entire report. To deselect, click a second time in a highlighted
area. The Copy command in the Edit menu copies selected display areas to the clipboard.
Annotate Tool
The annotate tool places a text box wherever you click in a JMP window. You can key in notes and
remove them at a later time, draw lines to make a special point, or use the annotate tool to enhance a
JMP graphical display. The annotate tool is selected for only a single click and then reverts to the arrow
cursor. Use Shift-select to cause the annotate tool to persist. See the section Using the Annotate Tool,
p. 229 in this chapter for details on the annotate tool.
Scroller Tool
Dragging the scroller tool scrolls reports up or down to display only the results you want to see. The
scroller is a precise way to perform the same function as the vertical and horizontal scroll bars. The
scroller reverts to the arrow when you release the mouse button. Use Shift-select to cause the scroller
tool to persist.
The hand tool (or grabber) is for direct manipulation or grabbing in plots and charts. It works like the
arrow on all other parts of report displays.
9 Report Windows and Surface Features—Cursor Tools 219
9 JMP Reports
Here are some examples of the way the hand behaves in graphs and plots:
• On axes in all plots, you can use the hand tool to change the range of the axis values, which then
changes the part of the plot that shows. Using the hand tool in a plot frame changes the X and Y
axes together.
• On histograms, use the hand tool to change the number of bars in a histogram or to shift the
boundaries of the bars on the axis.
• On spinning plots, the hand tool spins the plot. To spin a plot, grab the plot with the hand by hold-
ing down the mouse button and then move the hand about.
• In a multivariate scatterplot matrix, the hand can drag the position of a column of scatterplots to a
new position in the matrix by dragging along the diagonal.
• In a CUSUM control chart the hand relocates the origin of the V-mask.
• In all reports, use the hand tool and drag columns to rearrange the report.
The brush tool and lasso tool are for selecting points in plots. When you click the brush tool on a plot,
a rectangle appears. Move the rectangle over points to select and highlight them in the plot and in the
active data table. Shift-click to extend the selection. Alt-click (Option-click on the Macintosh) to
change the size of the selection rectangle. If the brush tool is not in a plot area, it behaves the same as
the arrow tool.
The lasso tool lets you select points in an irregular area of a plot. Drag the lasso around any set of
points. When you release the lasso it automatically closes and highlights the points within the enclosed
area. Use Shift-lasso to extend a selection.
There are more details about selecting points with the cursor, the brush tool, and the lasso tool in the
section Selecting Points in Plots and Using Row States, p. 223.
Magnifier Tool
The magnifier tool automatically zooms in on any area of a plot. When you click the center of a new
view of the data. The scale of the new view is enlarged approximately 25%, giving you a closer look at
interesting points or patterns. Click and drag the magnifier to focus in on a particular region of the
plot. You can zoom repeatedly to look closer at the data. Use Alt-click under Windows (Option-click
on the Macintosh) at any time to restore the original plot.
Crosshair Tool
The crosshair makes movable set of axes used to measure points and distances in graphical displays. For
example, on a fitted line or curve the crosshair identifies the response value for any predicted value. The
220 9 Report Windows and Surface Features—Features and Formatting
values where the crosshair intersects the vertical and horizontal axis appear automatically as you drag
the crosshair within a plot. On a ternary plot, this tool displays small triangular crosshair lines and dis-
plays the three axis values.
9 JMP Reports
Context Commands for Report Tables
By default, an analysis table is plain, with no lines or other style to separate rows and columns. Context
menu items for report tables let you tailor the appearance and content of the tables as follows:
Table Style
enhances the appearance of a table by drawing borders or other visual style to the table rows and
columns (Figure 9.3).
Figure 9.3 Table Style Context Menu and Beveled Table
Columns
lets you specify which columns you want to show in the analysis table. Analysis tables often have
many columns. The leftmost table in Figure 9.4 is a Parameter Estimates table showing only the
estimate name, the estimate itself, and the probability associated with the estimate. The standard
error and Chi-square values are hidden.
Make into Data Table
creates a JMP data table from an analysis table. The rightmost data table in Figure 9.4, is the
JMP data table result of sorted Parameter Estimates table with hidden columns.
Sort by Column
rearranges the rows of a report table. This command displays a list of visible columns in a report
and you choose one or more columns as sort variables. The middle table in Figure 9.4 is the
Parameter Estimates table sorted by Prob>ChiSq.
Figure 9.4 Commands for Analysis Tables
Hidden Report Columns Sorted Columns Report Report as JMP Data Table
legend with colors or markers based on the values of a variable you choose, or add information to the
plot with JSL commands.
The example in Figure 9.5 shows a bivariate analysis of variables height and weight. Context menu
commands added a plot legend (Row Legend) that identified gender with different colored markers.
The Add Graphics Script (or Edit Graphics Script) has commands that draw dashed horizontal lines
on the plot at the maximum and minimum of height.
Figure 9.5 Analysis Window Modified with JSL Statements
Special options are available to manipulate plots and graphs in each analysis window. To resize a plot or
graph, place the cursor on the right edge, bottom edge, or lower-right corner of the plot frame. The
cursor changes to a small double-arrow pointer. Drag to change the size of the plot horizontally or ver-
tically or both at the same time. When you resize, the height and width of all plots in that frame adjust
independently of other frames in the same report window.
When you use resize, the Alt, Control, and Shift keys (Option, Command, and Shift on the Macin-
tosh) have the following special effects:
• Alt-resize under Windows or Option-resize on the Macintosh adjusts a plot in 8-pixel increments.
• Control-resize under Windows (Command-resize on the Macintosh) adjusts all like plots in that
window simultaneously. For example, you can have scatterplots and mosaic plots in the same win-
dow. If you use Control-resize (Command-resize) in one scatterplot, all scatterplots resize together.
The mosaic plots are unchanged.
• Shift-resize adjusts the plot frame and preserves the proportions (aspect ratio).
9 Report Windows and Surface Features—Features and Formatting 223
9 JMP Reports
Selecting Points in Plots and Using Row States
Most graphs display columns from the data table. Points in scatterplots have different appearances
depending on the active row state characteristics. Row state conditions include select, exclude, hide,
label, marker, and color. There are 16 JMP markers and 65 JMP colors. Row states are described in the
chapter “Characteristics of Data” on page 75. To apply a row state characteristic you must first select
rows or points in plots.
Selecting Points
Clicking a point in a plot with the arrow cursor selects (highlights) the point as well as the correspond-
ing row in the current data table. The label for that row appears on the display and persists as long as
you hold down the mouse button. (The label also shows if you move the mouse and linger briefly over
a point, then disappears when you move)
When you click the arrow in a display, previously selected data are no longer selected. You can either
select a new point or no points depending on where you click. To extend a selection, hold down the
shift key while you click new points.
The brush tool in JMP selects a group of points delineated by a rectangle. As you drag the brush over a
plot, points that fall within the rectangle are temporarily selected. As the brush tool passes over points,
selected points are highlighted and appear larger. When you release the mouse button, the points
within the rectangle remain selected.
You can Shift-drag the brush tool to extend the current selection. The new selection includes the previ-
ous selection and all points that pass within the rectangle while the Shift key is pressed. The additional
selected areas need not be contiguous.
You can Alt-drag (Option-drag on the Macintosh) the brush tool to resize its rectangular selection area.
This feature lets you drag a rectangle around the area containing all the points you want to select. You
can also use Alt-drag to form a thin vertical or horizontal rectangle. This shape acts like a slicing tool
that can traverse and highlight slices of points across either axis.
If you hold down the Command key and drag the brush tool, the selection status of points within the
rectangle continuously inverts. This causes the points to blink on all representations of the data.
If you Control-Alt-drag (Command-Option-drag on the Macintosh) with the brush tool and then let
go (push the rectangle), the rectangle bounces around in the frame on its own. The speed and direction
you push (or drag) the mouse determines the speed and direction of the bouncing rectangle. Click the
mouse button in the report window to stop the rectangle from bouncing.
The lasso tool lets you select an irregular area of points in plots. To select points, drag the lasso around
any set of points. When you release the lasso, it automatically closes and highlights the selected points
within the enclosed area. Use Shift-lasso to extend a selection.
Labeling Points
Label is a special row state that signals plots to show a label for selected rows. Row numbers are the
default labels when there is no column assigned as a label column.
224 9 Report Windows and Surface Features—Features and Formatting
To designate a label using values in a column, select Label/Unlabel from the Cols menu for the
selected columns whose values you want to show as labels.
To apply a row state, select rows in any representation of the data table and choose the row state you
want from the Rows menu. The data table and plots automatically show the effect of the row state.
You can save row states permanently in a row state column. The popup menu that shows with a row
state column in the Columns panel to the left of the data grid has these commands.
Copy to Row States
Copy the row state condition from the row state column to the row number area of the data grid,
which activates the row states. This command writes over any existing active row state informa-
tion.
Add to Row States
Add the row state conditions from the row state column to the row number area of the data grid,
but preserves all active row state information.
Copy from Row States
Copy the row state condition showing in the row number area of the data grid to the row state
column. This command writes over any row state information already in the row state column.
Add from Row States
Add the row state conditions showing in the row number area of the data grid to the row state
column, leaving any row state information in the row state column intact.
Customizing Axes
You can double-click in the numeric axis areas of most plots or charts to see an Axis Specification dia-
log. Customization features on the dialog depend on the data type of the axis and with the specific plat-
form. Figure 9.6 shows a typical Axis Specification dialog for numeric axes.
If an axis is a JMP date value, the Format popup menu gives you options for the date display and an
additional popup lists date increments for tick marks. JMP dates are numeric values that correspond to
the number of seconds since January 1, 1904. JMP dates and the date display formats are described in
the chapter “Characteristics of Data” on page 75.
9 Report Windows and Surface Features—Features and Formatting 225
9 JMP Reports
Figure 9.6 The Axis Specification Dialog
divides the JMP date (number of seconds) into the appropriate units to give the plot scale you want for
your data. The date axis must be a column with a JMP data value and appears in the axis specification
dialog in the date format found in the Column Info dialog. However, you can use the Axis Specifica-
tion dialog to format the date any way you want it to be displayed in the plot.
Figure 9.8 Format Axis with Dates
9 JMP Reports
Figure 9.9 Context Menus for Nominal and Continuous Axes
options to modify
label text of
numeric columns
Scrolling an Axis
The hand tool (grabber tool) is another way to change the axes of a plot, and the view of the plot. You
can drag the hand back and forth on the X or Y axis of a plot and slide the axis range accordingly. Drag-
ging the hand tool in a plot frame changes the X and Y axes together and changes the position of the
plot in the plot frame.
Note: The grab-and-drag axis feature of the hand tool is locked in a journal if the axis is dependent on
output that is no longer active. For example, when a distribution is journaled, the histogram is no
longer active. The histogram axis must be locked to correspond with the histogram bars.
Zooming
The magnifying glass tool lets you automatically zoom in on any area of a plot. When you click the
magnifier, the point or area where you click becomes the center of a new view of the data. The plot
enlarges approximately 40-45%, focusing on the area in to give you a closer look at interesting points or
patterns. When you drag the magnifier to form a rectangular area, this area enlarges and occupies the
whole axis.
Option-click on the Macintosh or Alt-click under Windows at any time anywhere in the plot frame to
restore the original plot.
The Pollen.jmp sample data table dramatically illustrates the magnifier. The data were devised with a
hidden pattern (message) buried in over 3,800 points. The Fit Y by X platform displays the whole table
as the dense cloud shown to the left in Figure 9.10. Clicking the magnifier twice produces the plot in
the middle, and clicking twice more reveals the surprise message shown in the plot on the right.
Figure 9.10 Example of Zooming with the Magnifier Tool
9 Report Windows and Surface Features—Features and Formatting 229
9 JMP Reports
Using the Annotate Tool
The annotate tool creates an editable note area wherever you click and drag it in a JMP report, journal,
or layout. You can key in notes and remove them at a later time, or use the annotate tool to enhance a
JMP graphical display.
To use the annotate tool, select it from the Tools menu or toolbar, and click or drag a rectangle in a dis-
play window to create a white editable text box. Enter text and click outside the note to see a yellow
sticky note. By default, the position of the note is fixed (or anchored) to its place on the report. It
moves when you scroll the report, resize a graph, or shift an axis.
The annotation can be manipulated in the following ways:
• Click inside the text box at any time to add or edit its text.
• When you click outside the text box, it is no longer editable.
• Click and drag inside the annotation box or on its edge to move it. When an annotation is moved, it
becomes selected, as indicated by a double blue line with handles around the perimeter of the note
(see Figure 9.11)
• Place the cursor on the handle of a note (showing in the middle of the edges and in the corners) and
it appears as a single crossed arrow; drag to resize the note.
• Place the cursor on the edge of a note, but not on a handle, and it appears as a double crossed arrow;
drag to move the note.
Right-click (Control-click on the Macintosh) on a note for a menu of options (Figure 9.11) to tailor
the appearance of the annotation.
Background Color and Text Color
display the JMP color palette to change the background color of the note and the color of the
Font used for the message. The color of the font also defines the color of the note outline. If you
choose the same color for both the background and the font, the font does not show except for
black and white. A black font on a black background changes to white, and a white font on a
white background appears black.
Font
displays the standard dialog for choosing a font type, size, and style.
Tagline
draws a line from the nearest side, initially the left side, of the note to the report surface as illus-
trated in Figure 9.11. The note moves slightly down and to the right, and the line is positioned
from where the upper left corner of the note was to the edge of the note in its new position. By
default, the end of the line that is not attached to the note stays attached to its original position
when you move the note. To move the line with the note to a new position, hold down the Con-
trol key (Command key on the Macintosh) and drag the note.
Filled
gives the option of having the note filled with the background color you chose, or of having a
transparent note. A transparent note is handy for putting titles and footnotes on a graph.
230 9 Report Windows and Surface Features—Features and Formatting
The Line, Polygon, and Simple Shape (oval or rectangle) tools work much like the Annotate tool. They
behave in a similar fashion to those in many draw packages.
Line Tool
Select the Line tool and drag anywhere on a report, journal, or layout window. The line appears
selected, showing handles on both ends. Click and drag on the line to move it. Click on a handle and
drag to rotate the line. Click on the line to alternately select and deselect it.
Right-click (Control-click on the Macintosh) on a line for a menu of options to tailor the appearance of
the line:
Point to and Point from
place arrows on either end (or both ends) of the line.
Thick
alternately displays the width of a line as thick or thin. A line is thin by default.
9 Report Windows and Surface Features—Features and Formatting 231
9 JMP Reports
Dashed
alternately displays a line as dashed or solid. A line is solid by default.
Color
displays the JMP color palette to change the color of the line.
Reanchor and Unanchor
fixes(Reanchor) or releases(Unanchor) the line to or from an object in a layout window. These
two commands work as described above for the Annotate tool.
Delete
removes the line from the report surface. You can also remove the line by pressing the Delete (or
Backspace) key.
If you want to copy part of a report window, use the area selection tool in the Tools menu. Click and
drag across the area you want to copy to select and highlight it. Use Shift-click to extend the selection.
The example below shows the selection of a histogram. They are two report elements that can be
selected separately, or together, as illustrated in Figure 9.12.
9 Report Windows and Surface Features—Journal and Layout Windows 233
9 JMP Reports
Figure 9.12 Selection of Graph Elements
If nothing is selected, the Copy command copies the entire window to the clipboard. If you right-click
(Control-click on the Macintosh) when this tool is active, a popup menu gives you commands to
Copy, Paste, or Select All.
Any element in a JMP report window that can be selected can be dragged. When you drag report ele-
ments within the report frame, they are copied to the destination area where you drop them. As you
drag, an outline of the element shows and the edge of the report frame shows if you attempt to drag
outside it.
To delete a copied report element, select it and click the Delete key.
You can copy and paste any report element to other applications, and drag-and-drop JMP reports and
graphs to any other application that supports drag-and-drop operations.
The format used when pasting depends on the application you paste into. If the application has a
Paste Special command, you can select a paste format such as: text (RTF), unformatted text (TXT),
picture (PICT or WMF), bitmap (BMP), and enhanced picture (EMF).
9 JMP Reports
Table 9.1 Available Save As file Formats
Suffix Format Use
.PNG Portable Network Graphics compressed bitmap picture, successor to GIF
.JPG Joint Photographics Expert compressed bitmap pictures, standard for photographs
Group
.WMF Window Metafile pictures stored as commands, resolution independent
.JRN JMP Journal file allows JMP to read and show as originally created
.TXT Plain Text plain text, no pictures
.HTM Hypertext Markup Lan- world wide web format, marked up text that refers to separate
guage picture files
.RTF rich text format word processing, mixture of pictures, text, and tables
To begin using the layout window to rearrange results, click inside the layout window and choose
Ungroup from the Layout menu. This performs the first stage of ungrouping report elements, which
ungroups (or disconnects) the topmost title bar from its reports and subreports. Choose Ungroup again
to ungroup the next level of the report outline. You can continue to select specific report elements (or
all elements) and ungroup them until each title, and each piece of a report or plot is an object. Then
click on an object to select it and move it anywhere in the layout window. The layout window has as
many pages as you want, outlined with gray boundary lines.
236 9 Report Windows and Surface Features—Journal and Layout Windows
You can ungroup a report from its title bar, and ungroup all its major pieces, but you cannot ungroup a
column in a report table from its column header. However, you can double-click on a title bar or report
table column header and edit it.
Note: As with most drawing software packages, keyboard shortcuts are very handy, especially for
ungrouping, To quickly ungroup a layout to its smallest objects, use Control-U (Command-U on the
Macintosh) repeatedly.
Each time you ungroup a report outline level title from its reports, the disclosure icon for that level dis-
appears and you can no longer close it. However, you can do many of the surface operations available in
the Analysis window and in the journal window, use the context menu in plots, tailor axes, and so
forth. Also, the popup menu on the title bar has two active commands: Rerun in a new window and
Edit Script.
You can use the area selection tool as described above for the layout window, but the Layout menu is
dimmed unless you are using the arrow cursor.
The example in Figure 9.13 shows results extracted, combined, and organized in a compact form from
a bivariate analysis (scatterplot and polynomial fit), a distribution analysis (histogram and frequency
count), and a one-way analysis (means and t-test). The result gives a neat-looking summary of oxygen
uptake as a function of run time in an exercise experiment. This was done by forming a layout window
for one analysis, then dragging the other desired analysis elements into the layout window, arranging
the desired pieces, and deleting unwanted elements if there are any.
Figure 9.13 Multiple Platform Results in a Layout Window
10
10 Get Personal
Personalizing JMP
This chapter describes ways to tailor JMP to give a comfortable working environment.
• JMP Preferences enable you to specify general and specific settings and save the settings so they are
associated with your JMP session.
• Under Microsoft Windows, complete customization of menus and toolbars give the flexibility to
show only the commands you need, or set up JMP for groups with special interests.
10
Contents
Preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Background Color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Fonts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Graphic Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Communications Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
File Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Preferences Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Platform Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Text Import/Export . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Tool Bars and Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
The Show Toolbars Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
The Customize Command in the Edit Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Customizing Menus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
Customizing Toolbars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
10 Personalizing JMP—Preferences 239
10 Get Personal
Preferences
The Preferences command in the File menu (Edit menu on Macintosh OS 9 and the JMP application
menu on OS X) displays the tab dialog in Figure 10.1. Most of the tab items are straightforward. You
can turn them on and off to see the effect. This section describes the preferences, and explains those
that might not be obvious. The following descriptions are for a Windows environment. Some of these
are Windows-specific preferences, but most are also available on the Macintosh.
To change preferences, click a tab page and check the items you want or select from a menu of items.
Click Apply to see the results, and then click OK.
Figure 10.1 The General Panel Preferences Dialog
General
When you select Preferences, the dialog appears with the General preferences tab (Figure 10.1) show-
ing. Most General preferences customize your JMP session environment as follows:
• Initial Splash Window shows or hides the initial splash window at start up.
• Initial JMP Starter Window shows or hides the JMP Starter window at start up. The default prefer-
ence shows the JMP Starter window. If you don’t use this tool, turn it off. The JMP Starter com-
mand in the View menu (Window menu on the Macintosh) can be used at any time in a JMP
session.
• Suppress Formula Eval On Open suppresses evaluation of formulas when a data table is opened.
• Suppress OnOpen Script Eval suppresses any table property that automatically executes a JSL
script when the table opens.
240 10 Personalizing JMP—Preferences
• Auto match brackets in script editor automatically includes a closed bracket whenever you type
an open bracket in a script.
• Enter Key moves down causes the data entry cursor to move to the next cell down instead of the
next cell to the right.
• Print Data Grid as is prints the JMP data grid as it appears on the screen instead of listing the data
without lines separating the rows and columns.
• Date Title on Output displays the date and time the analysis occurred at the top of the output page.
• Data Table Title on Output places the name of the data table and notes, if there are any, at the top of
its output.
• Report Table Style has the submenu shown below that sets the output table style. This is the same
menu displayed by a context click in a report window.
Background Color
Under Microsoft Window, the Report window and data table can have any color background. When
you check the Background Color tab, you see the JMP color palette (Figure 10.2) and a sample of the
current background color. To set the background color, click a color in the palette, or choose Other and
create a custom background color. Click the Apply button on the Preferences dialog to preview the
result.
Note: The background of the plot frame is white by default, and is not affected by the window back-
ground color. You can change the plot frame background by using the context popup menu in the
frame, or set a default graph background color using the Graphs radio button in the preferences dialog.
10 Personalizing JMP—Preferences 241
10 Get Personal
Figure 10.2 Background Color Preferences Tab
Fonts
One way you can customize the appearance of reports is to change text styles using the Fonts prefer-
ences tab. This is the standard Windows or Macintosh font preference dialog, which lists the fonts and
sizes available in your system, and the regular, bold, italic, and bold italic styles. You can specify font,
style and size characteristic for the items listed on the Fonts tab.
The following areas in a JMP can be given a specific font, font size, and font style.
Text is the text portion of a JMP analysis report
Heading is the heading of columns in an analysis report
Title is the title shown in all title bars
Small sets the font for small text such as that showing in the upper left corner of the data grid if when
you have closed the side panels of the data table.
Mono is for the font used in the JSL editor for script commands
Calculator is for the expressions entered into the Formula Editor
Annotation sets the font information for the first annotation (note) you create in a JMP session. In a
single JMP session, the appearance of the first annotation carries to all following annotations you cre-
ate. Thus, to set a default using this preference, you must set the preference before creating any anota-
tions.
Graphic Formats
The check boxes on the Graphics Formats tab (see Figure 10.3) tell the Copy command what formats
to use when copying the contents of the selected area in a JMP report to the clipboard. You can check
any or all of these formats. When you paste from the clipboard, the receiving application uses the for-
242 10 Personalizing JMP—Preferences
mat the clipboard has as its default. If the application does not recognize any of the formats on the clip-
board, a dialog appears and prompts you to select a format.
Also, you can save a JMP document as an RTF file or as an HTML file, and choose the format you
want (PNG, JPG, or Metafile) from the menus at the bottom of the Graphics Formats tab page.
Figure 10.3 The Graphics Format Tab for Preferences
Communications Settings
Under Microsoft Windows, the Communications Settings only need to be specified if you are using an
instrument to read data from an external source. Refer to the documentation for your instrument to
find the appropriate settings.
File Location
Under Microsoft Windows, the File Location tab lets you change the default locations of JMP system
files. Usually, the files can stay where JMP installs them. However, you might want to move large files,
such as the Help files (jmp.chm), to somewhere other than your hard drive. Information in the Installa-
tion Directory is sometimes needed by technical support. You should leave most installed files in the
default installation directory.
Note: If you have more than one version of JMP installed, the Reset Association button lets you assign
JMP-specific file extensions (.JMP, .JSL, and so on) to the JMP application.
Preferences Path
Under Microsoft Windows, you can designate a directory where the preferences file (JMP.pfs) should
be stored. To do so, select the File Location tab from the Preferences panel. Then, select Preferences
file directory from the listbox, and click the Browse button. Navigate into a directory where the
JMP.pfs should be stored and click the Select button. You can point to a directory even if there is no
10 Personalizing JMP—Preferences 243
10 Get Personal
JMP.pfs file already in it. When JMP saves the preferences, it will create the JMP.pfs file in that direc-
tory.
Figure 10.4 The File Location Tab for Preferences
Platform Settings
Each Analysis and Graph platform has a variety of different plots and tables in its report. Some of these
results are shown by default. You can set the default output by using the check boxes in the Options
checkbox list that displays for each platform. As an example, a bivariate analysis shows points by
default. If you always want to see a linear fit each time the platform is launched, check the Fit Line pref-
erence, as shown in Figure 10.5. All bivariate analyses will then automatically appear with a linear fit.
Figure 10.5 Platform Settings Preferences Tab
Text Import/Export
The Open command displays a specialized open file dialog that lets you locate the file you want to
open, and tell JMP the format of the incoming file.
244 10 Personalizing JMP—Tool Bars and Menus
For text files, you can use options to identify the field and line delimiters and strip enclosing quotes
from character strings. The defaults for these options are in the Text Import/Export Preferences dialog
shown in Figure 10.6. You can change the defaults to suit the most common type of text file you read
or write. The field and line delimiters can be any combination of the characters indicated by check
boxes. Of particular importance at this time is the option to pick the 2-digit year rule you want to use.
The default 2-digit year interpretation appends 20 to the front of 2-digit years indicated by 00–10, and
19 otherwise.
Figure 10.6 Text Import/Export Preferences Tab
10 Get Personal
Figure 10.7 Show Toolbars Check Box Options
In addition to tool bars for main menu items, there are also check boxes for the following options:
• The Data_Tables_List option displays all open data tables. This is especially useful when you have
multiple data tables open. You select a data table in this list to make it the current table. The current
table is the one to which analyses apply.
Note: The current table is not necessarily the front window. To bring a table (or any window) to the
front, select its name from the list of open windows in the Window menu.
• The Browser option displays standard browser buttons such as forward, backward, and home, to
use when an internet page is open in a browser within JMP. When this toolbar item is open, you can
type a URL into the text area to open a new page in the JMP browser. Also, the Internet Open com-
mand on the File menu prompts you for a URL and opens an internet browser in the JMP session
for the location you enter.
• The URL_List option lists the browser pages you have requested during the current JMP session (see
Figure 10.8). You can also use the Internet Open command on the File menu, which prompts you
for a URL and opens an internet browser in the JMP session for the location you enter, or opens a
file at the location as a data table.
Figure 10.8 Toolbars for Tools, Browser, URL List, Data Table List, and Search List Control
• The Search_List_Control option keeps track of all the data table searches performed on the current
data table. When you enter a value in this text area and click Enter, the action is the same as using
the Edit > Search > Find command. You can select from this list to do the same as
Edit > Search > Find Next.
246 10 Personalizing JMP—The Customize Command in the Edit Menu
(Windows only)
The Customize command in the Edit menu lets you completely tailor all menus and toolbars to make
your JMP environment best fit your work activities.
Customizing Menus
All main menus and their submenus can be rearranged or eliminated from the menu bar with the
Edit > Customize > Menus window. When you select Menus from the Customize menu, the window
in Figure 10.9 appears. It contains drawings of all the main menus. Using drag-and-drop actions and
context menu commands, you can reposition or eliminate any menu or menu item, or create new
menus and menu items. Each menu or menu item has properties that you assign using its Properties
dialog. To see a properties dialog, right-click on the menu or menu item and select the Properties com-
mand from the context menu that appears. Customized menus can be saved and used anytime during a
JMP session. You can restore the default menus at any time with Edit > Customize > Revert to
Factory Defaults.
10 Personalizing JMP—The Customize Command in the Edit Menu 247
10 Get Personal
Figure 10.9 The Customize Menus Palette
drag to
new position
10 Get Personal
• If you want the menu item to execute a JSL script, either give the path to the file containing the JSL
(Run JSL in this file), or key the JSL into the text area called Execute this JSL.
The other tabs on the Menu Item Properties dialog are for optionally assigning your own icon to the
menu item, or assigning a shortcut key.
Figure 10.11 Menu Item Properties Dialog
Customizing Toolbars
Customizing tool bars is similar to customizing menus. Choose Edit > Customize > Toolbars to see
the Customize Toolbars palette. You manipulate the toolbars and individual tools the same way as the
menus and menu items, described in the previous section. You can rearrange existing toolbars, elimi-
nate tools, or create new toolbars and new tools.
The customize tools palette has context menus accessible from the palette surface with commands to
create a New Toolbar, a New Item (tool), Paste, which duplicates a tool you have copied with
Edit > Copy, or Apply Changes, and import a JMP Version 4 toolbar (described below).
250 10 Personalizing JMP—The Customize Command in the Edit Menu
There are also context menus for the whole toolbar, accessible by right-clicking on the toolbar name
area above the tools, and for individual tools.
The menu for the whole toolbar has a Delete Toolbar command and a Properties command. The
Properties command lets you name the toolbar, and use radio buttons to designate whether the tool-
bar will initially be visible or not, and whether it is to be docked or floating when visible. If it is not vis-
ible, its name appears unchecked in the View > Show Toolbars list.
The menu for individual tools has commands for positioning the tool, Copy, Paste, and Delete com-
mands, and a Properties command that displays the Toolbar Button Properties dialog, similar to the
one shown previously in Figure 10.11. The Toolbar Button Properties dialog has these features:
• On the General tab, use radio buttons to specify whether the new menu item is a Command or a
Separator line.
• Give the new item a Caption (name).
• Optionally, give the new tool a Tooltip.
• If you want a tool to access an existing JMP command, select it from the Builtin command list. This
list has all platform level commands in JMP, including those that appear in submenus, such as
Nonlinear or Partition.
• If you want the menu item to execute a JSL script, either give the path to the file containing the JSL
(Run JSL in this file), or enter the JSL statements directly.
• The Toolbar Button Properties dialog is for optionally assigning your own icon to the tool, or giving
it a keyboard shortcut.
10 Get Personal
Figure 10.12 Import JMP Version 4 Toolbars
right-mouse
click to Import
V4 Toolbar File
Appendix A
A What’s New
What’s New in JMP Version 5
JMP Version 5 builds on the improved interface of Version 4, adding new data access, statistics, and
graphics commands. JMP’s scripting language has been enhanced. The menu structure and tool bars
are now fully customizable, and new drawing tools have been added to aid in presentation-style results.
This chapter provides an overview of the enhancements. For complete details of each command, see the
appropriate entry in the main documentation.
A
Contents
Data Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Data Manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
New Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Statistical Platform Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Design of Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Internal Optimizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
JMP Scripting Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
OLE Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Appendix A 255
A What’s New
Data Access
Column Selection
Under Microsoft Windows, you can select which columns are imported from other JMP or SAS files.
Text Export
There is now an option to have no delimiter in exported text files. When selected, the variables are
exported with no quotation marks or spaces between them.
Data Manipulation
Column Formulas
Save Formula and Suppress Formula Evaluation commands have been added to JMP.
Row Editor
The Row Editor brings up a window useful for browsing the columns of a data table one row at a time.
Random Selection
You can select rows of a JMP data table at random by using the new Select Randomly command. This
command allows selection of a random number or random proportion of rows.
Appendix A 257
A What’s New
New Platforms
Discriminant Analysis
Although the Manova platform has always contained discriminant analysis in previous versions of JMP,
it didn’t produce any displays and had few features. JMP 5 includes a new platform for discriminant
analysis and has features for stepwise selection, canonical plots, and identification of rows by scoring
profiles. See “Discriminant Analysis,” p. 423 in the JMP Statistics and Graphics Guide for details about
the Discriminant platform.
Partition
The Partition platform recursively partitions data according to a relationship between the X and Y val-
ues, creating a tree of partitions. Variations of this technique go by many names: CARTTM, CHAIDTM,
C4.5, C5, and others. See “Recursive Partitioning,” p. 483 in the JMP Statistics and Graphics Guide for
details about the Partition platform.
Neural Net
The Neural Net platform allows for simple neural net modeling using a single hidden layer. See “Neural
Nets,” p. 413 in the JMP Statistics and Graphics Guide for details about the Neural Net platform.
Diagram
The Diagram platform’s main intended use is to do Ishikawa charts, also called fishbone charts, or
cause-and-effect diagrams. See “Ishikawa Diagrams,” p. 625 in the JMP Statistics and Graphics Guide for
details about the Diagram platform.
258 Appendix A
Capability Analysis
Capability analysis can now use four different options for estimating σ:
• Long-term
• Specified
• Short-term, grouped by fixed subgroup size
• Short-term, grouped by column.
A What’s New
• CV (coefficient of variation) has been added as a hidden column in REML and EMS results.
Clustering
The following new commands have been added in JMP Version 5.
• An Ordering column can now (optionally) be specified in the Hierarchical Clustering launch dialog.
In the ordering column, clusters are sorted by their mean value.
• Color Map is an option to add a color map showing the values of all the data colored across it's value
range. There are several color theme choices in a submenu.
• Geometric X Scale is an option useful when there are many clusters and you want the clusters near
the top of the tree to be more visible than those at the bottom. This option is the default for more
than 256 rows.
• Save Cluster Hierarchy saves information needed if you want to create a custom dendrogram with
scripting.
• Orientation allows you to rotate the dendrogram.
• Biplots are now available for k-means clusters.
• Self-Organizing Maps (SOMs) have been implemented in k-means clustering.
When there are a large number of clusters and the frame is squeezed down, Cluster now adapts the
cluster labels to smaller fonts, as needed.
ROC Curves
The positive level in a ROC curve in now specified by the user.
Design of Experiments
• You can add effects to the model that aren’t required to be estimable. JMP 5 allows these effects to
only be estimated if possible.
• Supersaturated designs are now practical to obtain.
• I-optimality is now an alternative to D-optimality in appropriate situations.
260 Appendix A
• For certain situations with covariate factors, the row exchange algorithm using a candidate set is
available in the custom designer.
• Desirability Functions are much improved. New desirability functions have been implemented for
both maximizing targets and minimizing targets.
• JMP now (optionally) shows a table with D, G, and A efficiencies for a custom design.
• The design matrix can be stored in a data table.
• Random starts have been implemented when searching for optimal designs.
The following scripting commands that let you predefine guidelines for the design search by the cus-
tom designer. See the JMP Design of Experiments guide for details on the new features in Version 5.
DOE Mixture Sum
DOE Starts
DOE Starting Design
DOE Search Points Per Factor
DOE K Exchange Value
DOE Bayes Diagonal
Graphics
Charting
Bar charts that show means can have standard error bars attached to them, as shown below.
Overlay Plot
• The Overlay Plot platform can use character variables as X variables.
• Overlay plots can have two y-axes, as shown below.
Appendix A 261
A What’s New
Control Chart
Capability analyses are included as an option in control charts.
Pareto Plot
Pareto Plot includes an Ungroup Plots option to allow a group of Pareto charts to be split up into sep-
arate plots.
Drawing Tools
In addition to JMP’s traditional annotation tool, JMP 5 has tools for drawing lines, ovals, rectangles,
and polygons. These graphics can appear on reports, in journals, and in layouts.
User Interface
Menu Customization
Under Microsoft windows, the way in which the user customizes menus and toolbars has been
changed. At the bottom of the Edit menu, there is a new submenu named Customize. This brings up a
drag-and-drop editor for customization of menus and toolbars.
Preferences Path
Under Microsoft Windows, the user can now designate a directory where the preferences file (JMP.pfs)
should be stored.
Formulas
Formulas now provide a visual clue (a caret) when optional parameters can be included.
Pictures of formulas now appear in the Column Info dialog.
Other Improvements
To move columns in a report around, drag them with the hand tool.
If you have nominal axes and ask for dividers, JMP now automatically wraps the text across several
lines, if needed.
Pictures that you have dragged or copied into JMP reports can now be resized. (To restore a picture to
normal size, Alt-click on it.)
Internal Optimizations
Although they are not feature enhancements, it is worth noting the optimizations that took place
between JMP 4 and JMP 5.
• The executable has been virtually halved in size. It now takes up approximately 7 megabytes of disk
space.
• The help system was reworked to reduce image sizes and to include a more useful navigation struc-
ture.
• Formulas that calculated across the rows of a data table (using, for example, subscripted variables or
the Lag function) stressed the formula dispatcher and dependency system for large data tables. This
resulted in a marked decrease in performance. The formula dependency system has been rewritten
completely and now performs quickly.
• The Cluster platform is much faster than before, making it now practical to do problems with sev-
eral thousand rows. One large test situation involved 4700 rows. Although requiring 125 meg of
2
memory because of the large n distance matrix, the calculations completed in about two minutes.
• ODBC is much faster for importing large database tables, including Excel files.
• The JSL Try function now intercepts errors better.
A What’s New
found in Matsumoto and Nishimura (1998). The new generators are verified to pass all the DIEHARD
tests as documented in Marsaglia (1996).
The old routines are still available as RandomSeededUniform and RandomSeededNormal and work
with the RandomSeed function call. For compatibility, Version 3 data tables with Random Uniform
and Random Normal functions are converted to the new names for the old functions,
RandomSeededUniform and RandomSeededNormal.
Formulas
New functions make setting column values possible without the use of a formula.
column<<Set Each Value(expression) sets each value in a column equal to the value of
expression. The expression is not stored in the column.
Invisible Reports
Reports can now be generated without displaying them. This is useful when a large number of reports
are generated that a script user does not need to see.
Foreground Analyses
Analyses initiated from Nonlinear or Stepwise commands can be forced to complete their execution
in the foreground before proceeding to the rest of the script using the Finish command.
An EvalFormula command can now be (optionally) added in a New Column function to cause the
formula in the column to be evaluated in the foreground.
dt<<Run Formulas runs all the formulas in the data table in the foreground until completion.
264 Appendix A
Spline Functions
Several new spline functions allow access to the spline features in the bivariate platform directly.
Pictures
JSL has a new data type Picture that holds pictures of generated reports.
Matrices
The following new matrix commands have been implemented in JSL.
• LocMin and LocMax
• Contains
• VecQuadratic
• InvUpdt
• GInverse
• QR
In addition,
• The Set Matrix command has been added to easily convert a matrix to a data table.
• Interpolate and Step now support x- and y-values that are matrices.
• The Mail command supports attachments (Windows only).
Data Table
Column attributes, locking, and formatting are accessible through JSL.
You can revert to a saved version of a data table through the File > Revert command.
Appendix A 265
A What’s New
DLLs
JSL can now pass a single parameter to and from functions embedded in DLLs.
Other Commands
The following other commands have been implemented.
• AsDate for date conversion
• AsName for string conversion
• Find and Window for window manipulation
• a graphical GradientFunction
• Text positioning
• Web for displaying web pages in JMP
• Quit and Exit for interactively shutting down JMP
• Journal and Layout messages are supported in JSL
• Squish, a logistic function
• Derivative now directly reports symbolic derivatives
• Type(expression) returns a string naming the type of expression.
Other Improvements
• Any structure of a list can be used in scalar operations.
• Evaluation of names of columns has been improved.
• Multiple left-side subscripts (e.g. a[i][j]) have been implemented.
• JSL references to the Distribution and Fit Model platforms can be subscripted by number or the
name of the y variable.
• It is now possible to launch Nonlinear when there is only a loss function.
OLE Automation
A vast number of new automation methods have been added to JMP 5, corresponding to other
improvements in JMP.
Index
Index
JMP User’s Guide
Index
columnwise statistics functions 194–198 Custom Design 20, 49
combinatoric functions 170–172 customize axes 224
combine columns into one 121 customize JMP 239–243
Combine Row States 206 customize menus 246
Combine States 208 customize results 235–236
Communications Settings 30, 242 customize toolbars 244, 249
Comparison functions 178–179 Cut 33, 157, 232
compositional data see Ternary Plot
compressed 235
computed-values column 143–160
D
Concat 174 dashed rectangle tool 64, 219
Concatenate 22, 40, 122, 134–135 data
Conditional functions 179–184 add/delete column 104–105
efficiency 158 create 93–118
conjunction 179–184 enter and edit 105
Connections 114 export 112
constant 152, 156 import 108–109
Contains 174 navigate 93–118
Contents 71 data characteristics 75–91
context menus 220–222 data properties 89–91
Context-click menus 221–222 data table
Contingency 11 select and deselect rows/columns 101–102
continuous 78 data table template 158–160
Contour Plot 17, 60 data types 77
Contour Profiler 18, 62–63 Database 29, 103, 113
Control Charts 18, 60 Database Open Table dialog 114
Control Limits column property 90 datafeed 242
Control-click menus 220–222 Date DMY 203
conventions 4 Date MDY 203
Copy 33, 116, 157, 218, 232 Date Time functions 203–205
Copy As Text 34 date-time data 224
Copy from Row States 86–87, 224 formats 82–84
Copy Picture 217 functions 203–205
copy pictures 234 text import/export 244
copy text 234 Day 204
Copy to Row States 86 Day Of Week 204
Copy to Row States 224 Day Of Year 204
Copy with Labels 33 dBase 108–112
copy/paste results 233–234 Decimal 220
correlation see Multivariate default path 30, 242
Cosine 172 Delete button 151
CosineH 173 delete column 104–105
Count 167, 169 Delete Columns 49, 105
Create Columns 104 delete formula 46
create data table 93–118 Delete key 152
create formula 143–160 delete results 233–234
create new data table 103 Delete Rows 45, 104
270 Index
Index
Fixed Decimal format 82 Group 73
flip data table 121 grouped statistics 121
Floor 168 grouping of expressions 156
foldover design see Augment Design
Fonts preferences 30, 241
For 183–184
H
Format 205 hand tool 51, 157, 159, 228
Format 224 in formula 147
format analysis table 220 hand tool (grabber tool) 218
formatting dates as axis values 225 Help 3, 71–72, 215, 218
Formula 46, 90, 104, 157 Help Center 72
formula Hex 177
drag and drop 157 Hidden 208
elements 152–155 Hide 70
functions 163–211 Hide Tool Palette 66
keyboard shortcuts 160 Hide/Unhide 42, 47, 85
store 158–160 hide/unhide results 216–217
tips 157–158 Hierarchical clustering 14
Formula Boxing 240 higher dimensions see Spinning Plot
formula display area 150 histogram
formula editor 143, 145–160 change bars with hand 219
formula work panel 148–150 Horizontal 216
four-digit year 244 HTM 235
Freq role 79–81 HTML file 242
Full Factorial Design 21, 50 Hue State 208
function 148, 156, 163–211 Hypergeometric Distribution 193
function browser 146, 150, 165 Hypergeometric Probability 193
Hypertext Markup Language 235
G
Gage R&R see Variability
I
Gamma 171 I-beam cursor 100, 106
Gamma Density 188 range check 100
Gamma Distribution 188 I-beam cursor (range check) 88
Gamma Quantile 188 ID column 130
General preferences 239 idf functions 184–194
generate random numbers 199 If 179–180, 201
GIF 235 Ignore Errors 149
Go to Row 102 Import as Data command 111
grabber tool (hand tool) 218, 228 import data 108–109
graph file formats 235 Importing
Graph Marker Size 240 Instant Text Importing 255
Graph menu 59–62 In Days 203
Graph tab of JMP Starter 16–18 In Hours 203
Graphic 241 In Minutes 203
Graphic Formats preferences 30 In Weeks 203
graphics formats 241 In Years 203
272 Index
match values of two data tables see Join NegBinomial Probability 192
Matched Pairs 12, 53 nesting box in formula editor 146
Matching Cols 138 Neural Net 55
Max 125 Neural Net 13, 55
Index
Maximum 198 New Column 46, 77–84, 104, 145
MDYHMS 205 New Column vs. Add Columns 105
Mean 125 New Data Table 28
Mean 197 new data table 103
Median 126 New Data View 67–68
menu bar 25 New Files 31
merge horizontally see Join New Property 104
merge vertically see Concatenate New Script 28
Mersenne-Twister technique 199, 202 New Table Property 96
metadata 89–91 New Table Variable 96
Min 125 Next Selected 43
Minimum 198–201 node controls 216–217
minor tick marks 225 nominal 78
missing term 148 Nominal Logistic 54
missing value noninteger weight 81
comparison 179 Nonlinear Fit 13, 54, 160
Mixture column property 90 nonlinear fit parameter 154–155
Mixture Design 21, 50 nonlinear table templates 160
Modeling tab of JMP Starter 12 normal curve 160
modeling types 77 Normal Density 189
Modulo 168 Normal Distribution 189
Month 204 normal distribution 160
Move Backward 73 Normal Quantile 189
Move Forward 73 Not 183
Move Rows 45 notation in manual 4
Move To Back 69 Notes 96
Move to Back 73 Notes column property 90
Move to Front 73 NRow 167
multiple actions 217 Num 176
Multiply To 211 Number 197
Multivariate 56 numeric 78
Multivariate Methods 14 numeric column 78
Multivariate tab of JMP Starter 14 numeric formats ??–82
Munger 174–176 Numeric functions 168–169
N O
N 125 OC Curves 20
N Missing 125 ODBC import 108–112
N Missing 198 Oneway 11
navigate data table 93–118 Open 108
NChooseK 171 Open 28, 103, 243
needle plot see Overlay Plot Open
negation 179–184 preferences 244
negative frequency 81 Open All Below 216
negative weight 81 Open All Like This 216
NegBinomial Distribution 192 open cross cursor 100
274 Index
Open JMP File (internet open option) 116 Poisson Distribution 193
operating characteristic curves see OC Curves Poisson Probability 193
operator 163–211 Pollen.jmp 228
optimize use space in report windows 215 polygon (spline) tool 65, 231
Or 182 polygon tool options 231
order of operations 155–157 popup menus 217, 220–222
ordinal 78 popup pointer cursor 100
Ordinal Logistic 54 Portable Network Graphics 235
orthogonal arrays see Taguchi Arrays Post Decrement 211
Other column property 91 Post Increment 211
Outline Connecting Lines 240 power see exponential, Sample Size and Power
outline controls 216–217 precedence 155–157
oval tool 65, 232 prediction profiler see Profiler
oval tool (simple shape tool) 230 prediction traces see Profiler
Overlay Plot 16, 59, 160 Preferences 30, 239–243
overview of JMP 1 preferences path 242
prerequisites 3–5
Preselect Roles 80
P Previous Selected 44
Page Breaks 215, 217 principal components see Spinning Plot
Page Setup 31 Print 30, 234
paintbrush tool 64 Print Preview 31
panels see Column, Row, Tables panel print report 233–234
parameter 147, 149, 154–155 Probability functions 184–194
Parameters 154 Probit 189
Parametric Regression 15, 58 Product 198
parentheses 155–156 Profiler 17, 62
Pareto Plot 19, 61 properties 89–91, 97
Partition 13, 55 Proportional Hazards 15, 54, 58
Paste 103, 117, 157 publish on WWW 235
paste results 233–234 PValue format 82
Paste Special 233
Paste with Labels 34, 117
path preference 30 Q
pdf functions 184–194 QC tab of JMP Starter 18–20
Pearson correlation 81 quality improvement see QC
peel function 152, 157 Quantile 197
personalize JMP 239–243 quantile functions 184–194
photographs 235 question mark tool 63, 218
PICT 233 Quit 31
picture file formats 235 quotation marks 109
plain text 235
plain text file 108, 112, 233
platform preferences 30, 243
R
plot statistics 122 R&R see Variability
PLS (partial least squares) 15, 57 Random Binomial 201
plus icon 97 Random Cauchy 200
PNG 235 Random Exp 199–201
PNG graphics format 242 Random Gamma 200
pocket calculator see formula Random Geometric 201
point chart see Chart Random Integer 200–201
Index 275
Index
Random Poisson 201 summary of 209
random sample 128 Rows 36, 41–45
Random Seed 201 Rows panel 98
Random Triangular 200 rowwise statistics functions 194–198
Random Uniform 199 RTF 233, 235, 242
Random Uniform() 202 Run Script 35
RandomSeededNormal 202
RandomSeededNormal() 199
RandomSeededUniform() 199, 202
S
Range 126 Sample Size and Power 21, 51
Range Check ??–88 SAS data 108–112
Range Check 47, 88, 100 date-time formats 84
Range Check column property 90 SAS Datasets 110
range plot see Overlay Plot SAS files, writing 110
real-time data 30 SAS Transport 108, 112
rearrange results 234–236 SAS Variable Names 110
rearrange terms in formula 147 Save 29, 104
rectangle tool (simple shape tool) 230 Save As 29, 36, 112
Recurrence 15, 58 Save As
Redraw 67, 69 preferences 244
reference of functions 163–211 Save Level Numbers 159
relative frequency see Pareto save report 233–234
Remove Property 46, 89 Save Selection As 34
Reorder Columns 48–49 scatterplot matrix
reorder rows see Sort rearrange 219
Replace 35, 107 Screening 54
Replace Table 130 Screening Design 21, 50
report windows script see JSL
customize 235–236 scroll axis 228
disclosure button 216–217 Scroll bar arrow style 240
features 215–217 Scroll bar thumb style 240
resize plots/graphs 222, 234 Scroll Lock/Unlock 47
Response Limits column property 91 scroller tool 64, 218
Response Surface Design 21, 50 Search 35, 71, 106
response surfaces see Contour Profiler search also see Find
Revert 29 Select 87, 216
rich text format 233, 235 Select All 34, 233
right arrow 157 Select All Rows 87, 102
Right-click menus 220–222 select bars of histogram 128
roles 79–80 select block of cells 33
Root 170 Select Excluded 87, 102
rotated tick marks 225 select expression 157
Round 168 select graph 218
Row 167 Select Hidden 87, 102
Row Editor 44, 100 Select Labeled 87, 102
Row functions 166–167 select multiple rows/columns 33
Row Order Levels column property 90 select points 128, 219, 223–224
276 Index
Index
time functions 203–205 window panel 101
Time Series 13, 15, 56 Windows Metafile 233
title bar 236 WMF 233, 235
Today 203 Word 177
Tool Elements Window 67 world wide web authoring 235
Toolbars 63, 217, 244 X role 79
Toolbars Y role 79
customize 244 Y2K date-handling 244
Tools 63, 216–217, 229, 232, 244 Year 204
Transcendental functions 170–172 yx tool 219
Transpose 22, 40, 121, 133 zoom tool 219, 228
triangle control 216–217
Trigonometric functions 172–173
Trim 176
truth table 183
Tukey HSD Quantile 193
two-digit year 244
TXT 233, 235
type 104
U
Unanchor Tool Palette 66
Undo 33, 49, 104–105, 108, 117, 147
Ungroup 72, 235
Unit column property 91
univariate statistics 122
up arrow 157
Uppercase 176
URL_List toolbar 115
V
Validation 47
validation 88
Variability 19, 61
vector graphics 235
View menu 65–66
V-mask origin 219
W-Z
Week Of Year 204
Weibull Density 191
Weibull Distribution 191
Weibull Quantile 192
Weight role 79, 81