STAT 301 - Business Statistics
Lecture Review #19 - Analyzing Paired Data

Review questions:

1) When do we have paired data?

2) What is our basic procedure for analyzing paired data?

Computational exercises:

1) Balph Snerdwell, for his (decidedly weak) senior research project, wants to know whether sleeping with the textbook under one's pillow improves grades on a "knowledge festival." He recruits six volunteers. He has them put the book under their pillow for one exam, and not do so on another, and then asks about their test scores. Data are as follows:

Conduct all steps of an appropriate hypothesis test to determine whether sleeping with the book under your pillow affects test grades.

2) Clorinda Cragdingle believes that her pet wombat, Muffy, can select stocks that will perform well. (We won't go into details about how Muffy goes about picking the stocks. Suffice it to say that it involves putting a newspaper with stock listings down on the floor and having Muffy perform a biological function.) Clorinda has Muffy select one stock per month. She compares stock performance in that month with the performance of the S&P 500 (a measure of the overall stock market) for that month. Data are below:

a) Find a 95% confidence interval for the mean performance of Muffy's stocks, relative to the overall market.

b) May we conclude (with 95% confidence) that Muffy can beat the market?

SOLUTIONS:
1) t=-0.63; 5 df; p=0.278; don't reject null hypothesis
2) [Muffy - S&P]: -3.4% + 16.4%