1. Read Case Study 6.1.1.
2. Produce a boxplot for the 5 groups, plot individual qqplots for each group. Here is a script file to make all the qqplots at once. Take a look at the code; it may be useful later. Make preliminary conclusions about the adequacy of the ANOVA assumptions.
3. Run a 1-way ANOVA, and "Save Model Object" to the name of your choice. Confirm the p-value given in "Summary of Statistical Findings."
4. Make a plot of residuals vs. fitted values. What can you say about the assumptions that the errors have mean zero and constant variance across the 5 groups?
5. Reproduce the results of Display 6.6. Here you will be examining all possible pairwise comparisons using the multiple comparison procedures in Splus
6. Problem 13, page 169.

1. What would happen to the p-value in an ANOVA if 2 was added to each data point? Justify your answer.
2. What happens to the p-value in an ANOVA if the dataset is doubled by duplicating the data? (That is, you create a new dataset made of two copies of your data, thus doubling the number of observations, but keeping the same number of treatments or groups.) Justify your answer.

One-page writeup to turn in [15 points]: Vegetarians and Zinc (p. 147, Sleuth). Data located here.

Recall the format: Use Times New Roman font, no smaller than 11 point, with 1 inch margins all around for the 1-page summary. You can attach a one-page appendix, but the writeup should stand on its own. (Don't bury important information in the appendix. We will use this section to give partial credit in cases of calculation errors.) Your summary should have sections which give the research questions and briefly describe the data, exploratory data analysis, statistical analysis and scope of inference. Be sure to interpret any important confidence intervals or other important summary measures. If you transform the data, make sure that you include interpretations back in the original units.Explicitly state your assumptions, and use any diagnostic tools to ensure that the assumptions you make are appropriate. You will be working with real data, so be careful as you decide what assumptions are appropriate. As with any real data problem, there may not be a single "correct" answer. If there are outliers or influential cases that you have identified, mention them in your report. Rerun your tests without these cases to determine the sensitivity of your results to these observations.

Questions to address:

• Is there a difference in zinc levels between the 3 groups?
• What evidence is there that pregnant vegetarians tend to have lower zinc levels than pregnant nonvegetarians?
• Compare all pairwise differences using an appropriate multiple comparison procedure.