Abstract ABSTRACT: This article evaluates students' perception of the presence of forecast data within traditional accounting statements. The data base utilized was generated by simulating financial statements containing varying forecast disclosure formats. The statements were generated under two conditions, unbiased versus intentionally biased forecasts and high versus low variation in the financial data over time. The data sets were distributed to students, and their responses were analyzed using analysis-of-variance. Disclosure of the point estimate of earnings per share, confidence interval and ratio of predictive ability were the most useful methods in this experiment for predicting future earnings per share.