Emotions and moods are affective states that sometimes influence evaluative judgments. The most common research finding is that affect colors evaluations to be affect congruent. For example, people in positive affective states generally make more positive evaluations than people in negative states. The literature on affect and judgment is so consistent that one might assume that the influence of affect on judgment is always affect congruent, so that people in positive affective states evaluate everything more positively. This dissertation, however, asks what happens when a person in a positive state is confronted with a negative judgment object. Do people in happy moods see lying, cheating, or stealing as less bad than they otherwise would? Despite the consistency in the literature on affective congruence, I propose that affect often does not lead to affect-congruent evaluation, but rather that the direction of affective influences on evaluations often depends on participants' initial opinion of the target of evaluation. In this dissertation I will explore situations that do not lead to congruence, but instead lead to affective incongruence (Study 1), to influences on the extremity of responses (Study 2), and to affective enhancement of existing beliefs (Studies 3 and 4). In a final study, I tested the effect of mood on experimentally manipulated positive and negative targets. The goal of the dissertation is to examine the judgment conditions under which affect incongruence, rather than affect congruence, may occur and to propose an explanation to account for both traditional and new patterns of results.