Are Headstart gains on the g factor? A meta-analysis

Headstart studies of compensatory education tend to show impressive gains on IQ scores for children from low-quality environments. However, are these gains on the g factor of intelligence? We report a meta-analysis of the correlation between Headstart gains on the subtests of IQ batteries and the g loadings of these same subtests (K = 8 studies, total N = 602). A meta-analytic sample-weighed correlation of − .51 was found, which became − .80 after corrections for measurement error. We conclude that the pattern in Headstart gains on subtests of an IQ battery is highly similar to the pattern in test–retest gains and is hollow with respect to g. So, Headstart leads to gains in IQ scores, but not to gains in g. We discuss this finding in relation to the Flynn effect, training effects, and heritability.