|
|
STAT 440 - Forecasting This activity uses the dataset “BodyFat.xlsx” from this website. These are real data, from a study conducted at Brigham Young University, that sought to develop a regression model for a (male) person’s percent body fat, given various body dimensions. I obtained these data from the Journal of Statistics Education’s online archive of data sets for classroom use. Working with your partner(s), use these data to answer the following research questions. A write-up is expected from each group member. Question 1 What is the correlation between abdomen measurement (in centimeters) and percent body fat? Give a 95% confidence interval for this quantity. Question 2 How would your answer to Question 1 change, if abdomen size were measured in inches rather than in centimeters? Explain. Question 3 What is the (sample) correlation between abdomen and body fat, for men age 39 or less? For men aged 50 and older? Conduct an appropriate hypothesis test to see whether these two population correlations are the same. Question 4 For the overall sample, what is the correlation between abdomen measurement and percent body fat? Between thigh measurement and percent body fat? Conduct an appropriate hypothesis test to see whether these two population correlations are the same. Question 5 What is the population to which these results may be generalized? (What might you infer, based upon your limited knowledge of the study? And what would you need to find out, to answer the question? How might you obtain that information?) |