Comparing Two Means (Chapter 24)

independent t-test

• the means of two unpaired/independent groups are compared

• each group is composed of an independent set of subjects

effect size

the difference (variations) between group means

variance

measure tells us about dispersion of scores within groups

group differences

differences between groups / variability within groups

What assumption is made when comparing means?

homogeneity (or equality) of variance among groups (Levene’s test)

What do confidence intervals estimate in group comparisons?

the difference between group means that exist in the population

paired t-test

• repeated/dependent measures

• test analyzes difference scores within each pair, so that subjects are only compared with themselves

  • concept of using subjects as own control

• tend to be more powerful than independent t-tests because most of the extraneous sources of variation will be the same across the conditions

• reduces total error variance because most unknown error will be the same across conditions