Math 215 Report 1

Due at class time, Friday, 10/8/04. You may work with a partner on this report.

The data Bloodalc are obtained from a project by Ohio State University students. Sixteen students were randomly assigned to drink from 1 to 9 beers. 30 minutes after their last beer, a state patrol officer gave them a breathalyzer test. The variable beers gives the number of beers an individual consumed, and bac measures blood alcohol content. Also, the patrol officer administered a road sobriety test before (test1) and after (test2) drinking. The test measured coordination skills such as balancing on one foot and touching the tip of your nose with your eyes closed (10 is a perfect score).

Write a report addressing the questions below.

  1. Create a scatter plot of bac against beers. Describe the relationship and point out any unusual features.

  2. Now, find the regression line (save the residuals) and interpret the slope. What is the R^2 coefficient in this case (and what does it mean?) What bac would this line predict for a person who had had 4.5 beers? Plot the residuals against beers. Are there any unusual features?

    Note: Change the variable name and label so you'll remember the residuals come from this regression.

  3. Add a horizontal reference line to the original scatter plot at y = 0.08 (the legal limit in many states for driving). If an officer's breathlyzer machine was broken, should he/she use the number of beers as a predictor for blood alcohol content? Why or why not?

  4. Create a scatter plot of bac against test2, the score on the sobriety test after drinking. Describe the relationship and point out any unusual features, if any.

  5. Now, find the regression line and interpret the slope. What bac would this line predict for a person who scored a 7 on the sobriety test?

  6. Add a horizontal reference line at y = 0.08 (the legal limit in many states for driving) to the original scatter plot. If an officer's breathlyzer machine was broken, should he/she use the sobriety test as a predictor for bac? Why or why not?

  7. Repeat (4-6) but use the difference in test scores (test1 - test2) as the explanatory variable (for the prediction part, use a difference of 3).

Now, return to bac against beers:

  1. Recreate the scatter plot of bac against beers but now, add gender as markers. In addition, add the regression lines. Describe what you see.

  2. Repeat the scatter plot of the residuals (from part 2) against beers and add the gender markers. Describe what you see.

  3. Now, create a scatter plot of the residuals against weight and describe what you see (as weight increases, does the regression line under or over predict bac?)

  4. Split the data file Data > Split File by gender and then perform the regression (bac against beers) and save the residuals. What are the two regression lines? Interpret the slopes, R^2, and comment on the residual plots. Why do you suppose the slope for females is larger than the one for males?

  5. Summarize your findings and discuss the reliability of using the number of beers or the sobriety test as a predictor of blood alcohol content.

Report guidelines

Through this report, you will be making arguments in favor of certain conclusions about the data you are studying. The effectiveness of your arguments will depend on several things, including some (proper spelling, for example) that are not strictly relevant to the content of your report. Here's how I will break down the points for this assignment.

Start early, and have fun.




Jeff Ondich, Department of Mathematics and Computer Science, Carleton College, Northfield, MN 55057, (507) 646-4364, jondich@carleton.edu