Are the effects of lead exposure linked to the g factor? A meta-analysis

Does lead reduce IQ at the level of g, test specificities, or both? A bare-bones psychometric meta-analysis utilizing the Method of Correlated Vectors was performed on a sample of 16 studies for which subtest-level data could be obtained satisfying stringent inclusion rules. The aggregate correlation across samples between subtest-level estimates of both g loading (g) and the deleterious impact of lead exposure (d) was 0.10 (K = 16, total N = 1935, 80% CI after correction for sampling error = 0.10 to 0.10). So, lead exposure is associated with a slightly positive vector correlation, which is consistent with the results of other studies examining the effects of other neurotoxins on IQ using MCV; this outcome is consistent with two scenarios. The first is that lead exposure may have effects on both g and test specificities owing to systemic effects on many different brain regions. The second is that two antagonistic factors are at work. It might be that the 'control' and exposure groups used in these kinds of studies are confounded with pre-existing differences in g - lower g being a risk factor for poorer life outcomes (including lower socioeconomic status and concomitantly heightened risk of lead exposure), whereas lead has it's primary effect on the test specificities, with both effects opposing one another, as reflected in the small magnitude vector correlation value. Strategies for distinguishing between these scenarios are discussed.