Ground rules, Suggestions, Syllabus

Find a good study/homework partner work together regularly You will REQUIRE a 2-variable statistics graphing calculator, preferably the TI-83 or TI-84 (if you have a TI 89, it will do, but I am not as familiar with its usage). We will also use MINITAB a lot. Bring book and calculator to every class. Read materials ahead of time, and try (at least try) section EXERCISES prior to being covered in class. (I suggest trying pretty much every other odd number.) Homework assigned for GRADING will include work posted in BlackBoard and selections from the text. These will be representative of applications and QUIZZES/EXAMS. Bring your homework, problems, questions to class. After I cover the material, examples, and some homework problems, DO THEM AGAIN. Due dates for collection of graded homework are given in the schedule in the syllabus. There is either a Quiz or Homework assignment due each week. I am neither a cop nor an entertainer. I am your guide to this material, and you will get out of it what you put into it.

Syllabus
Quick review

1-1 The Engineering Method and Statistical Thinking

Engineers solve problems of interest to society by the efficient application of scientific principles The engineering or scientific method is the approach to formulating and solving these problems.

What is Statistics?
(Insides of the box on previous chart) Statistics is the making of inferences and decisions in the face of uncertainty. Terms and Concepts Individuals
form Are measured by

Populations
From which we take

Variables
-Numeric/quantitative -Categorical/qualitative

Samples
Dr. Sidik 5

1/22/2012

1-2 Collecting Engineering Data

Three basic methods for collecting data:

A retrospective or observational study A designed experiment

Simple Comparative Experiments Factorial experiments

Case Studies (papers and data in BlackBoard):

Michelson Speed of Light (Michelson)
Retrospective analysis of some observational data An observational study A Simple Comparative Experiment A Factorial Experiment examining/modeling multiple variables

Heart Rate Variation (Gelber, Shields) Bladder Stimulation (Dalmose) Electro-spun micro-fiber vascular grafts (Nottelet)

Michelson/Speed of Light
Exercise 2-68, pg 55, and associated document and web sites Detailed observational studies Illustrating several roles of theory, measurement methodology, and experimental methodology with respect to random error
BlackBoard: Case Studies & Readings folder. Michelsons 1879 Determination http://www.desy.de/user/projects/Physics/Relativity/SpeedOfLight/measure_c.html http://njsas.org/projects/speed_of_light/

History
Date 1676 1726 1849 1862 1879 1907 1926 1947 1958 1973 1983 Author Olaus Roemer James Bradley Armand Fizeau Leon Foucault Albert Michelson Rosa, Dorsay Albert Michelson Essen, GordenSmith K. D. Froome Evanson et al Method Jupiter's satellites Stellar Aberration Toothed Wheel Rotating Mirror Rotating Mirror Electromagneti c constants Rotating Mirror Cavity Resonator Radio Interferometer Lasers Adopted Value Result (km/s) 214,000 301,000 315,000 298,000 299,910 299,788 299,796 299,792 299,792.5 299,792.4574 299,792.458 +-500 +-50 +-30 +-4 +-3 +-0.1 +-0.001 Error

Background
Michelson: Rotating Mirrors approach
The data consist of 100 measurements, made over 28 days. Each measurement was an average of several replicates. The 100 measurements were made in 5 runs of 20 measurements each. Presumably something was done to the measurement process which distinguishes these runs. For now, we focus on descriptions of each set of 20.

Questions/Discussion
Type of data collection? Population of interest? Representativeness of sample? Adherence to Engineering Method? Variables and models possible?

Data Chapter 2, EX2-68

Loaded in Minitab
Tr1 850 900 930 950 980 1000 Tr2 960 960 880 850 900 830 Tr3 880 880 720 620 970 880 Tr4 890 810 800 760 750 910 Tr5 890 780 760 790 820 870

Graph Individual Value Plot Multiple Ys, simple, OK Select the Graph Variables and OK

 Histograms & Boxplots Minitab, TI, Technology tips Numerical descriptions SOCS
Shape Outliers Center Spread

Heart Rate Variation

What & Why

Some autonomous system neuropathic conditions can be identified readily and non-invasively from examination of an electrocardiogram (ECG) Time between successive heartbeats, the R-R interval, is easily read from an ECG. Its variation is called the Heart Rate Variation (HRV) To determine abnormal, one must first determine normal. D. Gelber et al. describe a study of 611 normal subjects observed across 63 hospitals, and also captured age, gender, bmi, and blood pressure for many. More info on this subject is also described by R. Shields of the Cleveland Clinic. The two articles and the Gelber dataset are in BlackBoard.

Study data
rr_var valsal_1 valsal_2 age gender systol diastol bmi mbp val1sq val2sq val1val2 val_cat age_cat rr_cat Value Codes: gender
RR_VAR VALSAL_1 VALSAL_2 39.40 2.08 2.08 14.00 1.78 1.71 71.00 140.50 39.40

heart rate variation to deep breathing valsalva measurement 1 valsalva measurement 2 age of subject gender of subject systolic blood pressure diastolic blood pressure body mass index mean blood pressure square of valsalva 1 square of valsalva 2 product of valsalva 1 & valsalva 2 valsalva in categories age in categories rr-variation in categories

1 = men, 2 = women
AGE GENDER SYSTOL DIASTOL 102.00 67.00 112.00 81.00 22.00 19.00 38.00 BMI MBP VAL1SQ VAL2SQ 84.50 4.33 4.33 96.50 3.17 2.92 VAL1VA VAL_CA AGE_CA T T RR_CAT L2 4.33 9.00 35.00 3.04 6.00 10.00 25.00 75.00 15.00 100.00 35.00 35.00

Data in Minitab
Graph Individual Value Plot One Y, simple Select data, OK

Graph Scatterplot simple, OK Select RR_VAR for Y, AGE for X, OK

Questions/Discussion
Type of data collection? Population of interest? Representativeness of sample? Adherence to Engineering Method? Variables and models possible?

Bladder Control by Electrical Stimulation

Aims: To investigate the feasibility of conditional short duration electrical stimulation of the penile/clitoral nerve as treatment for detrusor hyper-reflexia, the present study was initiated. Methods: Ten patients with spinal cord injury, 4 women and 6 men, with lesions at different levels above the sacral micturition center had a standard cystometry performed. During a subsequent cystometry, conditional short duration electrical stimulation of the penile/clitoral nerve was performed as treatment for one or more detrusor hyper-reflexia contractions. Results: In all patients, at least one contraction (mean, 7.8, range, 1-16 contractions) was inhibited by the stimulations. The mean cystometric capacity was increased significantly by conditional electrical stimulation, from 210 mL in the control cystometries to 349 mL in the stimulation cystometries (P=0.016). The maximal detrusor pressure during the first contraction in the control cystometries was mean 51 cm H2O, whereas the maximal pressure of the first contraction in the stimulation cystometries was reduced to mean 33 cm H2O (P=0.045). Conclusions: The authors conclude that repeated conditional short duration electrical stimulation significantly increased cystometric capacity in patients with spinal cord injury. The increase was caused mainly by an inhibition of detrusor contractions. The need for a reliable technique for chronic bladder activity monitoring is emphasized, as it is a prerequisite for clinical application of this treatment modality.
Conditional Stimulation of the Dorsal Penile/Clitoral Nerve May Increase Cystometric Capacity in Patients With Spinal Cord Injury, Dalmos et.al., Neurology and Urodynamics, 2003

Recent Advances: completed studies (KJG)

Efferent Afferent

Animal model proof of concept

Stimulation of urethral afferent nerves generates bladder contractions and voiding.

Translational tool developed

Urethral afferents can be electrically stimulated minimally invasively in humans. (Gustafson, 2003)

Human studies
Stimulation generates bladder contractions in SCI humans. (new)
First data generating bladder contractions via intraurethral electrical stimulation in humans.
(Gustafson, 2004)

Genital nerve stimulation inhibits bladder contractions in SCI humans. (accepted)

1/22/2012

(Dalmose, 2003)
19

Simple Comparative Experiment

Sub 1 2 3 4 5 6 7 8 9 10 Ctrl 290 188 208 77 86 376 400 57 353 68 Stim 500 424 268 279 159 500 320 500 500 42 S-C 210 236 60 202 73 124 -80 443 147 -26

In people with Spinal Cord Injury or with certain neurological conditions, Functional Electrical Stimulation of appropriate nerves may enable subjects to regain some control over bladder activity. Capacity data (ml) from Dalmos et al (table at left) is used to recap the essentials of simple statistical comparison.

-100

100 Mean (S-C) = 138.9 S.D. (S-C) = 147.9 T = 2.97, P = .016

500

Conditional Stimulation of the Dorsal Penile/Clitoral Nerve May Increase Cystometric Capacity in Patients With Spinal Cord Injury, Dalmos et.al., Neurology and Urodynamics, 2003 A urethral afferent mediated excitatory bladder reflex exists in humans, Gustafson et.al., Neuroscience Letters, 2004

1/22/2012

20

Questions/Discussion
Type of data collection? Population of interest? Representativeness of sample? Adherence to Engineering Method? Variables and models possible?

1-2 Collecting Engineering Data

1-2.3 Designed Experiments

Simple Comparative Experiments Factorial experiments Replicates Interaction

1-2 Collecting Engineering Data

1-2 Collecting Engineering Data

1-2 Collecting Engineering Data

1-2.4 Random Samples

1-2 Collecting Engineering Data

1-2.4 Random Samples

Populations and Samples

Physical vs. Conceptual
Individuals physically exist and all available Hypothetical set of all possible individuals that could exist, possibly in the future

Enumerative vs. Analytic

Sample used to enumerate (describe physical population of existing individuals) Sample used to analytically describe some future population (which may or may not yet exist)

Random Sampling
From a representative sample of individuals, we want to infer properties of the population. Simple random sample (of size n) Every sample of size n has equal chance of being selected Every member of the population has equal chance of being selected Best way to insure the sample is representative. Probability laws

Population

Sample

Statistical Inference Need to develop basic probability concepts, then use them for inference.
1/22/2012 Dr. Sidik 28

Using Random Numbers for Sampling

How to get a Simple Random Sample (from a small population) Construct a list (sampling frame) of every population member Number each member from 1 to N (population size) Use the TI 83/84 function: Math PRB 5:randInt randInt( lower,upper[,numtrials]) Simulates the act of blindly drawing a slip of paper from a box with numbered slips of paper between lower and upper and replacing the selected slip into the box. To simulate multiple draws, WITH replacement, provide optional argument numtrials. To simulate multiple draws, withOUT replacement, simply discard any repeats as they occur. You would want to set numtrials to something somewhat larger than the sample size you actually need. To save the numbers to a List, STATEDIT 1:Edit, ENTER Cursor up to the List name desired, then Math PRB 5:randInt with desired parameters and Enter

You will want to generate random numbers for project 1.

1/22/2012 29

1-3 Mechanistic and Empirical Models

A mechanistic model is built from our underlying knowledge of the basic physical mechanism that relates several variables. Example: Ohms Law Current = voltage/resistance I = E/R I = E/R +

1-3 Mechanistic and Empirical Models

An empirical model is built from our engineering and scientific knowledge of the phenomenon, but is not directly developed from our theoretical or firstprinciples understanding of the underlying mechanism.

1-3 Mechanistic and Empirical Models

Example of an Empirical Model
Suppose we are interested in the number average molecular weight (Mn) of a polymer. Now we know that Mn is related to the viscosity of the material (V), and it also depends on the amount of catalyst (C) and the temperature (T ) in the polymerization reactor when the material is manufactured. The relationship between Mn and these variables is

Mn = f(V,C,T)
say, where the form of the function f is unknown. where the s are unknown parameters.

Descriptions
SOCS
Shape Outliers Center Spread

Use Minitab or TI calculators

Minitab: Text Example 2-4: Compressive Strength BlackBoard, text datasets, chapter 2 Minitab StatBasic StatisticsGraphical Summary Variables: EX-24, OK TI: Exercise 2-38: sewage discharge Put data in L1 StatCalc1-Var Stats 2nd 1 2nd Stat plot, plot 1, turn on, select histogram or boxplot

2-1 Data Summary and Display

2-1 Data Summary and Display

2-1 Data Summary and Display

2-1 Data Summary and Display

Population Mean For a finite population with N measurements, the mean is

The sample mean is a reasonable estimate of the population mean.

2-1 Data Summary and Display

Sample Variance and Sample Standard Deviation

2-1 Data Summary and Display

2-1 Data Summary and Display

2-1 Data Summary and Display

The sample variance is

The sample standard deviation is

2-1 Data Summary and Display

Computational formula for s2

2-1 Data Summary and Display

Population Variance When the population is finite and consists of N values, we may define the population variance as

The sample variance is a reasonable estimate of the population variance.

2-2 Stem-and-Leaf Diagram

Steps for Constructing a Stem-and-Leaf Diagram

2-2 Stem-and-Leaf Diagram

2-2 Stem-and-Leaf Diagram

2-2 Stem-and-Leaf Diagram

2-2 Stem-and-Leaf Diagram

2-2 Stem-and-Leaf Diagram

2-2 Stem-and-Leaf Diagram

2-3 Histograms
A histogram is a more compact summary of data than a stem-and-leaf diagram. To construct a histogram for continuous data, we must divide the range of the data into intervals, which are usually called class intervals, cells, or bins. If possible, the bins should be of equal width to enhance the visual information in the histogram.

2-3 Histograms

2-3 Histograms

2-3 Histograms

2-3 Histograms

2-3 Histograms
An important variation of the histogram is the Pareto chart. This chart is widely used in quality and process improvement studies where the data usually represent different types of defects, failure modes, or other categories of interest to the analyst. The categories are ordered so that the category with the largest number of frequencies is on the left, followed by the category with the second largest number of frequencies, and so forth.

2-3 Histograms

2-4 Box Plots

The box plot is a graphical display that simultaneously describes several important features of a data set, such as center, spread, departure from symmetry, and identification of observations that lie unusually far from the bulk of the data. Whisker Outlier Extreme outlier

2-4 Box Plots

2-4 Box Plots

2-4 Box Plots