Why were the returns of stocks with low book-to-market ratios and high market capitalizations lower, on average, than the returns of stocks with high book-to-market ratios and low market capitalizations? In this paper we pit the characteristics hypothesis against the affect hypothesis. The characteristics hypothesis says that some characteristics, such as low book-to-market ratio and high market capitalization, are associated with high future stock returns in typical investors’ minds. The affect hypothesis says that the names of some companies elicit positive affect which is associated with high future stock returns in typical investors’ minds. We find, through experiments, that the evidence is more consistent with the affect hypothesis than with the characteristics hypothesis.