Sie sind auf Seite 1von 3

Minitab: ANOVA

Objectives: 1. Perform an ANOVA for one response variable and one factor with any number of levels. 2. Perform an ANOVA using stacked and unstacked data sets. Outcomes: 1. Recognise situations for which ANOVA is the appropriate analysis. 2. Explain the contents of an ANOVA table. 3. Interpret a test of the null hypothesis that the true means of several independent groups are equal. 4. Assess whether conditions and assumptions are met for a test. 5. Generalize from samples to populations.

Minitab Lab Activity: 1) Analysis of variance; unstacked data. (a) Generate 5 rows of random data from a Normal distribution with 60 and 3 . Store in C1-C4. You have four samples of size 5. Round off each column to integers using the Calculator command; store back in original column. Run a One-Way ANOVA (unstacked) on the data in C1-C4. Select the Graphs button in the ANOVA command window and choose Boxplots of data. (b) Generate 5 rows of random data from a Normal distribution and store in c5c8. For each column, use the following means and use 3 for all. c5 : = 20 c6: = 25 c7: = 30 c8: = 20 Remember to label your columns with appropriate titles. (eg C5 could be mean=20). Round off each column to integers using the Calculator command; store back in original column. Run a One-Way ANOVA (unstacked) on the data in C5-C8 . Select the Graphs button in the ANOVA command window and choose Boxplots of data.5

Minitab/ANOVA

Page 1

2)

Analysis of variance; stacked data. Retrieve the Pulse.mtw worksheet from the Data Directory in Minitab. This data set contains personal data on 92 randomly-selected students in a statistics class. Use Sample from Columns to randomly sample 87 rows from columns C1-C8 and store back into C1-C8. (a) Run a One-Way ANOVA to test for a difference in Pulse 1 between the smokers and non-smokers. Select the Graphs button in the ANOVA command window and choose Boxplots of data. (b) Run a One-Way ANOVA on Pulse1 versus Activity. Select the Graphs button in the ANOVA command window and choose Boxplots of data. (c) Run a One-Way ANOVA on Pulse1 versus Weight. Select the Graphs button in the ANOVA command window and choose Boxplots of data.

3)

Textbook exercises from Chapter 28: #21 and #22. Data sets can be found on the publishers website for your textbook, listed by chapter and exercise number: www.aw.com/deveaux-velleman Choose: Statistics:Data and Models Datasets Minitab Files Run a One-Way ANOVA on the data. Select the Graphs button in the ANOVA command window and choose Boxplots of data and Histogram of Residuals.

Analysis and Interpretation: 4) For each part of step (1): (a) Bold the p-value in the output. Using Graph Probability Distribution Plot View Probability F, provide a plot that illustrates the location of the area under the t-distribution represented by each p-value. (b) State your decision on Ho at the 5% level of significance and write a concluding statement immediately under the output. Compare your results from step (1a) to your results from step (1b). Why was the test significant in step (1b)? Put your answer directly under the output for step (1). For each part of step (2), do the following: (a) Based on the Boxplot, describe the distribution of differences with respect to centre and spread. Put your answer immediately under the Boxplot. (b) Bold the p-value in the output. (c) State your decision on Ho at the 5% level of significance and write a concluding statement immediately under the output. (d) Are the conditions for ANOVA satisfied?

5)

6)

Minitab/ANOVA

Page 2

7)

For both textbook exercises in step (3), do the following: (a) Based on the Boxplot, describe the distribution of differences with respect to centre and spread. Put your answer immediately under the Boxplot. (b) Bold the p-value in the output. (c) State your decision on Ho at the 5% level of significance and write a concluding statement immediately under the output. (d) Have the conditions for ANOVA been satisfied? Discuss. Write your answer under the output. A study compared saturated fat content (g) of pizzas sold by three national chains. For each chain, eight pizzas were chemically analyzed. The calculated F-ratio (the test statistic) was 4.56. (a) What are the hypotheses for this test? (b) What is the p-value for the test? Use the Graph Probability Distribution Plot View Probability F command. Remembera p-value is an area under the model curve based on a test statistic value. (c) What does the p-value measure in the context of this problem?

8)

New Minitab Commands: Calc Random Date Normal Calc Calculator

[use Round(c#,0) in Expression box where # is the column being rounded]

Stat ANOVA One-Way (unstacked) [Graphs: Boxplots of Data] [Graphs: Histogram of residuals] [Graphs: Boxplots of Data] [Graphs: Histogram of residuals] Graph Probability Distribution Plot View Probability F Stat ANOVA One-Way

Minitab/ANOVA

Page 3

Das könnte Ihnen auch gefallen