SAT non-g residuals predict course specific GPAs: Support for investment theory

This research examined relations of SAT residuals, obtained after removing g, with first-year course specific college GPAs (grade point averages) from the SAT Validity Study (N = 160,670). g was based on SAT scores and GPAs in two areas: science, technology, engineering, and math (STEM), and the humanities (e.g., English, art, history, foreign languages). The results revealed a domain specific pattern of effects: SAT math residuals correlated positively with math and STEM GPAs and negatively with verbal and humanities GPAs, whereas SAT verbal (reading and writing) residuals showed the opposite pattern of effects. The pattern of effects replicated with manifest GPAs and latent GPAs, with and without missing data, and with different STEM and humanities GPAs. The largest effects were obtained for models with either the SAT reading or writing subtest and with latent GPAs (absolute β ≈ .17). The difference between residual effects for math and verbal GPAs was medium in size (mean difference ≈ .27). The results supported investment theories, which predict that investment in a specific ability (math) boosts similar abilities but retards competing abilities (verbal). The findings are the first to confirm the predictions of investment theory with course specific GPAs.