Chapter 10: "t for 2" Most important concepts

1. Independent measures t is used to infer whether the µ's of two different unknown populations are the same or different

2. With two population means to estimate, we need two samples, and we lose two degrees of freedom, one for each sample mean.

3. This t test assumes that (a) both populations are normally distributed and (b) the populations have the same variance

4. Violations of (a) are not serious if the sample size is large; violations of (b) are least serious if the samples are the same size. Homogeneity of variance (b) can be tested using Hartley's F-max test (p. 253)

5. Independent means uses a new null hypothesis trick: We hypothesize that the mean difference is 0.

Key skills from Chapter 10:

1. Compute the t-statistic. This has many steps!!! Especially tricky is the pooling of standard errors.

To find the numerator,

(1) Subtract one sample mean from the other

(2) Subtract this from the hypothesized difference between population means (0 for null hypothesis).

To get the denominator,

(3) Find the "pooled" variance

(4) add the estimated errors from both samples to get the estimated standard error

2. Compute the degrees of freedom for the t statistic (n-2, lose one df for each x-bar).

3. Conduct test on SPSS and interpret the output, summarizing results in English and in APA form (p.249)