TESTING PSYCHOLOGICAL HYPOTHESES ADDRESSING 2 INDEPENDENT VARIABLES AND ONE DEPENDENT VARIABLE

Psychological hypotheses addressing the simultaneous influence of two independent variables on one dependent variable lead to the prediction of certain data patterns usually looked upon from the perspective of statistical interaction in two-way analysis of variance. With respect to certain types of psychological hypotheses statistical interaction m ay best be interpreted as a contrast of contrasts or a difference betw een two differences. Three types of patterns derived from three differ ent psychological hypotheses are analyzed from this point of view and with respect to statistical interaction. Our focus lies in discussing predicted patterns derived from some psychological hypotheses; are do not discuss subsequent problems such as choice of robust alternatives to the classical parametric tests we apply nor do we consider how to d eal with the problem of possibly cumulating error probabilities. In ad dition, we stress the difference between examining psychological hypot heses adequately and exhaustively and tests of data to gain more infor mation that may be interesting although not relevant to a particular h ypothesis.