Under “Spearman’s Law of Diminishing Returns” (SLODR) higher levels of general cognitive ability are associated with a more differentiated cognitive performance across ability tests. In the present study, evidence is presented that suggests that the phenomenon is a function of two methodological variables. First, the SLODR effect is greater when general ability is estimated on the basis of subtests which yield a more positively skewed distribution of scores. Second, SLODR direction is a function of the average subtest skewness in a battery, with SLODR associated with positive skewness and reverse SLODR associated with negative skewness. Contrary to previous speculation, task difficulty appears to be a relatively less important moderator. It is argued that together, these results imply problems with the manner in which SLODR is currently tested and raises the possibility that previous confirmations of SLODR are unreliable.