Alcohol and marital interactions
An experiment was conducted to examine the effects of alcohol on the marital interactions of husbands and wives (Journal of Abnormal Psychology, November 1998). A total of 135 couples participated in the experiment. The husband in each couple was classified as aggressive (60 husbands) or nonaggressive (75 husbands), based on an interview and his response to a questionnaire.
Before the marital interactions of the couples were observed, each husband was randomly assigned to three groups: receive no alcohol, receive several alcoholic mixed drinks, or receive placebos (nonalcoholic drinks disguised as mixed drinks). Consequently, a 2 × 3 factorial design was employed, with husband's aggressiveness at two levels (aggressive or nonaggressive) and husband's alcohol condition at three levels (no alcohol, alcohol, and placebo). The response variable observed during the marital interaction was severity of conflict (measured on a 100-point scale).
(a) A partial ANOVA table is shown below. Fill in the missing degrees of freedom.
(b) Interpret the p-value of the F-test for Aggressiveness.
(c) Interpret the p-value of the F-test for Alcohol Condition.
(d) The F-test for interaction was omitted from the article. Discuss the dangers of making inferences based on the tests, parts a and b, without knowing the result of the interaction test.
