This study examines whether market participants’ overpricing on previous-period accruals (i.e., accruals anomaly) is distinguishable depending on different patterns of earnings announcement. Sloan (1996) provided an explanation for the accruals anomaly that it is caused by a investors’ function fixation on the bottom-line earnings without proper understanding of mean-reverting properties of accruals. It means that investors could not recognize the information about accrual and cash flow components of earnings. Earnings announcement patterns could affect on investors' cognition about accrual informations. Particularly, the investors’ function fixation on earnings could differ depending on whether the accrual information is disclosed on earnings announcement days. Under disclosure requirement in Korea, the preliminarily disclosed earnings do not include detailed information on earnings components. Investors do not have information on earnings components until the financial statements disclosure date. Some investors who form their portfolios based on financial statements have information on earnings components while others who invest based on preliminary earnings do not have information on earnings components. Those investors who make their investment decisions based on preliminary earnings are likely to rely on an anchor implied in preliminary earnings. Also, such a bias toward the anchor is (not) likely to be adjusted subsequently when the financial statements are disclosed with preliminary earnings information adjusted (unadjusted). Thus, if the accruals anomaly is related to the anchoring and adjustment effect, we expect that the degree of the accruals anomaly differ across the earnings announcement patterns. The empirical results show that the accruals anomaly is significantly different depending on earnings announcement patterns in consistent directions with our expectations. Specifically, we find that the accruals anomaly is more prominent for the subsample with preliminary earnings announcements than the subsample without preliminary earnings announcements. Further, we show that the accruals anomaly is more prominent for the subsample with preliminary earnings announcements to be subsequently adjusted than for the subsample with preliminary earnings announcements to be subsequently not adjusted. In this regard, we believe that our findings provide policy implications that investors practical implications that the requiring firms to report cash-based and accrual-based earnings in preliminary earnings announcement leads to attenuate the anomaly. Also, the results in this study provide policy implications how disclosure affects investors’ behavior.