We demonstrate empirically that P/Es calculated from accounting measures of EPS are understated because of the presence of potentially dilutive securities. This result persisted after we controlled for the effects of risk and growth, and we found it to hold for both generally accepted accounting measures of EPS (i.e., basic and diluted EPS). It is not surprising that P/Es calculated from basic EPS are understated. Basic EPS tends to be overstated because it is calculated without consideration of the effect of potentially dilutive securities. It is surprising that P/Es computed from diluted EPS, which is generally believed to be an extremely conservative measure that reflects maximum dilution, are also understated. Because of the importance of P/Es as inputs in models of stock valuation and portfolio formulation, our findings should be of interest generally to investors and financial analysts.