# Statistics

This chapter allows you to discuss a variety of statistical issues. In addition to covering the independent groups t test (see Lecture 10.2,), you could talk about effect size (see box 10.2),  the controversy about significance testing (to give a balanced view of both sides of the debate, see this link for an elegant defense of significance testing or see Appendix D), the relative value of one-tailed versus two-tailed tests, or about nonparametric alternatives to the traditional t test. Generally, you will want to cover the following statistical issues:

### Null hypothesis

Emphasize the fact that one either rejects the null hypothesis or fails to reject it: The null hypothesis is never supported. The inability to find a difference doesn't mean that a difference doesn't exist. This fact affects what hypotheses researchers can test (they can't prove that something has no effect or that two manipulations have identical effects) and how researchers interpret null results. (Table 10-2 will help you make these points). You could even talk about the controversy about the null hypothesis. In discussing the controversy, you may wish to refer to the the home page for the Journal of Articles in Support of the Null Hypothesis and to APA's preliminary report on hypothesis testing.

### Type I and Type II errors

To help students understand the statistical significance decision, what can go wrong (Type I and Type II errors), and the tradeoffs made between Type I and Type II errors, review Table 10-3. Then, have students generate examples of each kind of error and present them with different situations, asking them which error would be more serious for that situation. To illustrate that these errors do not occur solely in psychological investigations, you might ask students what the implications are of the different errors for jury decision making, polygraph tests, pregnancy tests, AIDS testing of individuals, AIDS testing of donated blood, car alarms, drug tests (for example, according to one report, Type 2 errors in drug testing were 33%, Type 1 66.5%.).