This analysis of gender performance differences in a first-level New Zealand university course in computer science is predicated on the model of academic ability proposed by Charles Spearman, the eminent educational psychologist and statistician. The regression model is based on the theoretical constructs of general academic ability and specific academic ability. An attraction of this approach is that controls are effected for between-gender differences in these facets of academic ability. The analysis, which incorporates corrections to the standard errors to control for undesirable properties in the residuals, reveals a large, and unexpected, gender difference.