The examination of psychological hypotheses by planned contrasts refering to two-factor interactions in fixed-effects ANOVA

When interested in examining psychological hypotheses preceding data gathering especially in designs with more than two groups the standard ANOVA approach does not represent a very satisfying testing strategy. In these cases the statistical hypotheses actually tested are linked only loosely to the psychological hypothesis under scrutiny. In order to link both kinds of hypotheses more closely, the statistical hypotheses should be derived from the psychological one. These derived statistical hypotheses refer to contrasts a priori, which are constructed in order to conform with the predictions. Accordingly, the method of planned contrasts as a versatile alternative avoiding the shortcomings of the common ANOVA approach is chosen, and its advantages for testing a priori hypotheses are demonstrated. Examples are provided to show how to examine psychological hypotheses by means of the proposed testing strategy where these hypotheses refer to two independent (e.g., age and an experimental variable) and one (or more) dependent variable(s) and enabling predictions within and across age groups.