R. Richardson Pettit concludes from his empirical study of the information content of dividends in the Journal of Finance that participants make considerable use of the information implicit in dividend payments.' That conclusion is exactly the opposite of the conclusion in the Watts article in the Journal of Business on the information content of dividends.2 In another article in this issue of the Journal of Business Pettit attempts to reconcile the two findings.3 In that second article, Pettit makes the point that the two different conclusions depend on the specification of the earnings and dividend information variables in the original Pettit and Watts studies. Pettit argues that the Watts dividend information variable is misspecified, but apparently that his earnings variable is not. Consequently, his conclusion remains that dividends do have information content. The results presented in Pettit's second article do not indicate whether the dividend information content variable used by Watts is or is not misspecified. However, it is apparent from Pettit's original article and from his reconciliation article that the Pettit earnings variable is misspecified. In fact, Pettit's earnings misspecification is such that we cannot conclude from Pettit's results, using his methodology, that dividends do have information content. Pettit's results (in the second article) with a reformulation that used my methodology on regular dividends suggest that dividends could perhaps convey information to market participants. However, those results are not clear cut. Let us look at these points in more detail.