Congruence between actual and retrospective reports for pre- and postcompetition emotional states was investigated separately and together. Fifty-two members of four university sport teams participated in one or more of three experimental conditions. The first condition consisted of actual measurement of precompetition emotional states and retrospective measurement of the same situation following a 72-hr delay. Actual and retrospective measurement of postcompetition emotional states comprised the second condition. The third condition included actual measurement of pre- and post-states and retrospective measurement of both states after a 72-hr delay. RM-MANOVA procedures revealed that athletes could report and differentiate between their pre- and postcompetition emotional experiences, and that retrospective report was not affected by the pre/post interference after a 72-hour delay. However, athletes underestimated the intensity of postcompetition unpleasant emotions. Correlations between the structured actual and retrospective measures of emotions were moderate to strong, and thus congruent. However, thoughts and feelings that were openly expressed after 72 hours were not fully congruent with thoughts and feelings reported in real time. These findings are discussed in relation to Ericsson and Simon’s (1980, 1984) conceptualization of verbal reports as data, and Ross’ (1989) implicit theory of stability and change.