Rubin, Piché, Michlin, and Johnson (1984) recently presented data allegedly demonstrating a substantial relationship between social-cognitive ability and narrative writing skill. Certain theoretical and statistical considerations led us to suspect that the claimed relationship was not actually present in the data reported by Rubin et al. Consequently, two empirical studies were conducted to test for the hypothesized relationship between social-cognitive ability and narrative writing skill, one study reanalyzing data reported by Rubin et al. and the second analyzing original data. The results of the two studies indicate no relationship between social-cognitive ability and rated quality of narrative essays. These findings are discussed in terms of a theoretical model of the relationships among cognitive abilities, discourse aims, and discourse models.