OUTLINE FOR CHAPTER 8

1. The Seven-Step Version of Hypothesis Testing
   1. Introduction:
      1. A brief review of the simplest version of hypothesis testing
      2. The 7th step: After rejecting Ho, determining the degree to which Ho was wrong
      3. The difference between "statistical significance" and "practical significance"
   2. Two ways researchers can do Step #7
      1. They can compute a measure of "effect size"
      2. They can conduct a post hoc "power analysis"
2. The Nine-Step Version of Hypothesis Testing
   1. A simple listing of the nine steps . . . with the 3 "new" steps located in positions 4, 5, and 6
   2. Step #4: Specification of the effect size (ES):
      1. ES: the point that separates cases where Ho is false by a small and trivial amount vs. cases where it's false by a big and noteworthy amount
      2. Two options for ES: "Raw" or "standardized" (and Cohen's "standards")
   3. Step #5: Specification of the desired level of "power":
      1. The notions of "statistical power" and a "beta error"
      2. Why researchers don¹t set power at .999
   4. Step #6: Determination of the needed sample size:
      1. Computing n from a formula or looking up the needed n in a chart
      2. What to do if there is a fixed n
   5. The primary advantage of using the 9-step version of hypothesis testing
3. Hypothesis Testing Using Confidence Intervals
   1. Using a confidence interval to test a null hypothesis about a single population
   2. Using a confidence interval to test a null hypothesis involving two populations
4. Adjusting for an Inflated Type I Error Rate
   1. The notion of an "inflated Type I error rate"
   2. Why the Type I error rate becomes "inflated" if a fixed alpha is used with multiple tests
   3. The Bonferroni adjustment technique
   4. The experimentwise error rate
   5. The Dunn-Sidak modification
5. A Few Cautions
   1. Two meaning of the term "effect size"
   2. "Small," "medium," and "large" effect sizes
   3. The simplistic nature of the 6-step version of hypothesis testing
   4. Inflated Type I error rates

Copyright © 2003 Schuyler W. Huck
All rights reserved.