The congnitive dissonance theory and the social judgment theory predict that the boomerang effect occurs when the discrepancy is large. In the studies of Sakaki (1980a, b), however, the boomerang effect took place when the discrepancy was small, so that four additional studies were conducted in order to clarify the relationship of the two variables. The method was classroom experimentation based on the before after design. The results always showed that the boomerang effect and little opinion change occured when the discrepancy was small, whereas more opinion change took place at the large discrepancy levels, regardless of various conditions. Furthermore, in contradiction to the prediction of the dissonance theory, the Ss derogated the communicator when the discrepancy was small, while they raised the evaluation of the source credibility when it was large. A dual-factor hypothesis is proposed for these findings in the present study, and the hypothesis is herein supported.