This paper reexamines the explanatory power of beta, firm size, book-to-market equity, and the earnings-price ratio for average stock returns, correcting two currently controversial biases: selection bias in COMPUSTAT and the errors-in-variables (EIV) bias. After filling in the missing data on COMPUSTAT with the Moody's sample, I do not find any significantly different results for book-to-market equity from using the COMPUSTAT sample only. After correcting for the EIV bias, I find stronger support for the beta pricing theory than previous studies. Regardless of the presence of firm size, book-to-market equity, and earnings-price ratios, betas have significant explanatory power for average stock returns. In particular, firm size is barely significant using monthly returns, but no longer significant using quarterly returns. However, book-to-market equity still has significant explanatory power for average stock returns, even though the EIV bias is corrected. I. Introduction The primary implication ofthe capital asset pricing model by Sharpe (1964), Lintner (1965), and Black (1972) is the mean-variance efficiency of the market portfolio. Put differently, there exists a positive linear relation between ex ante expected returns and market betas, and variables other than beta should not have power in explaining the cross-section of expected returns. Contrary to the predictions of the CAPM model, previous empirical studies have found that idiosyncratic factors have significant explanatory power for average stock returns, while beta has little power. The most prominent idiosyncratic factors are firm size, book-to-market equity (B/V), and earnings-price (E/P) ratio. Banz (1981), Reinganum (1981), and Keim (1983) find that small (large) firms have greater (smaller) returns than those predicted by the CAPM model.1 Jegadeesh