Day 23: Statistics for Business Applications

Last Time: Testing the difference in two means (the case of unequal variances.)
- Example handout
- Heights of NKU students: males versus females.
  - use of linear regression on the data that four people provided.
  - Men's distribution
  - Women's distribution
  - Both distributions
  - T-test for difference in means of heights - reject H_o of no difference
  - T-test for difference in means of "distances" - fail to reject H_o of no difference
- In the case where the standard deviations are unequal,
  - the change is slight to the standard error
  - the change is great to the degrees of freedom (we lose additional degrees of freedom)
  - At this point is hard to say when to say they're unequal (just eyeball it!)
Return homework 12.4
Next time (Thursday): Collect problems 13.1-2
Questions about old stuff?
- One more example of difference in mean for independent samples:
  - Lou Gehrig's disease: Yesterday's Cinci Post (4/7/2003)
  - Example: #13.50, p. 411 and #13.52, p. 411
Today: The difference between two means (paired samples)
- Paired (matched) samples
  - The variable is the difference between a matched pair
  - Reduces to a standard t-test on the variable X_D, the difference in the pair
  - Example: 13.68, p. 425
  - Example: 13.69, p. 425
Next time: More on paired samples and inference about the difference of two means

Website maintained by Andy Long. Comments appreciated.