Sloan (1996)'s result of return predictability based on accrual information has generated a large stream of literature. Sloan's (1996) explanation for the phenomenon is that investors fixate on earnings without taking into account accruals' tendency to reverse. Thus, the returns to an accrual strategy are related to accrual reversals. In this study, I directly examine this explanation. I use two properties of the accrual-fixation hypothesis - reversals and overreaction - to generate testable empirical predictions. First, I find that extreme accrual firms tend to remain in extreme deciles in two consecutive years. Further, I find that these sticky firms are associated with future abnormal returns, suggesting that the returns are not a result of accrual reversals, as implied by the accrual-fixation hypothesis. Second, I do not find any evidence of overreaction to accrual information based on the relation between the returns around earnings announcements in successive quarters. Overall, I conclude that my results are not consistent with the accrual-fixation hypothesis. A more conservative interpretation of my results is that the accrual-fixation hypothesis is not the major explanation for the accrual anomaly.