Hypothesized that later occurrences in a series of events tend to evoke counterfactual alternatives more strongly and, hence, tend to be blamed more for ensuing negative outcomes than do earlier occurrences. In Study 1, Ss played the role of students whose task it was to read an article and then to identify the questions they thought a teacher might include on a test of it. Consistent with the hypothesis, Ss were less critical of a teacher whose test questions did not match their own when the teacher generated his or her questions before they did than when he or she generated them after they did. In Study 2, Ss played the role of teachers whose task it was to select questions to be answered by a student. Presumably, because of a greater fear of being blamed, Ss selected easier questions when their selection of questions occurred after the student had finished studying than when it occurred before the student began studying.