One possible explanation is “Spearman's hypothesis,” which
states that the magnitude of score differences across demographic groups is a direct function of how strongly
the test measures g. To test this hypothesis, Jensen (1985, 1998) developed the method of correlated vectors.