Measurement confounding affects the extent to which verbal IQ explains social gradients in mortality

Background IQ is thought to explain social gradientsin mortality. IQ scores are based roughly equally onVerbal IQ (VIQ) and Performance IQ tests. VIQ tests,however, are suspected to confound true verbal abilitywith socioeconomic status (SES), raising the possibilitythat associations between SES and IQ scores might beoverestimated. We examined, first, whether two of themost common types of VIQ tests exhibited differentialitem functioning (DIF) favouring persons of higher SESand/or majority race/ethnicity. Second, we assessed whatimpact, if any, this had on estimates of the extent towhich VIQ explains social gradients in mortality.Methods Data from the General Social Survey-NationalDeath Index cohort, a US population representativedataset, was used. Item response theory models queriedsocial-factor DIF on the Thorndike Verbal IntelligenceScale and Wechsler Adult Intelligence Scales, RevisedSimilarities test. Cox models examined mortalityassociations among SES and VIQ scores corrected anduncorrected for DIF.Results When uncorrected for DIF, VIQ was correlatedwith income, education, occupational prestige and race,with correlation coefficients ranging between |0.12| and|0.43|. After correcting for DIF, correlations ranged from|0.06| to |0.16|. Uncorrected VIQ scores explained11-40% of the Relative Index of Inequalities in mortalityfor social factors, while DIF-corrected scores explained2-29%.Conclusions Two of the common forms of VIQ testsappear confound verbal intelligence with SES. Sincethese tests appear in most IQ batteries, circumspectionmay be warranted in estimating the amount of socialinequalities in mortality attributable to IQ.