[OSF] In a longitudinal sample of 2,593 individuals from Minnesota, we investigated whether individuals with IQs ≤ 90 who completed college experienced the same social and economic benefits higher-IQ college graduates did. Although most individuals with IQs ≤ 90 did not have a college degree, the rate at which they completed college had increased ~6× in men and 10× in women relative to rates in the previous generation.
The magnitude of the college effect on occupational status, income, financial independence, and law abidingness was independent of IQ level, a finding replicated using the nationally representative NLSY97 sample. Additional analyses suggested the association of college with occupational status was consistent with a causal effect and that the educational success of individuals with low-average IQs may depend in part on non-ability factors, family socioeconomic status and genetic endowment.
We discuss our finding in the context of the recent expansion in college attainment as well as the dearth of research on individuals with low-average IQs.
[Keywords: low-average IQ, returns to college, non-ability contributors to educational attainment, general cognitive ability]
…The final stage in our analysis involved using the cotwin control (CTC) method to assess whether the college effects identified at the first stage were consistent with a causal effect (McGue et al 201014ya). As is true with any method for analyzing observational data, CTC analysis does not purport unequivocally to establish causality. Rather, it seeks to determine whether an association is consistent with causality using a test that is more stringent than the standard approach of statistically correcting for measured confounders. The CTC method derives its rationale from the counterfactual model of causality. By comparing outcomes in twins discordant on exposure (here completion of college), CTC analysis in effect uses one twin (ie. the college-completing twin) as an approximation to the counterfactual for the cotwin (ie. the non-college-completing twin.) The power of the CTC approach is a consequence of the matching of the twins. In our application of the CTC method, we investigated only monozygotic (MZ) twin pairs discordant for college completion, which controls for confounding due to genetic and rearing environmental factors because MZ twins are perfectly matched on these factors even when they are discordant for college completion. The use of the CTC method to assess the returns to education has a long history in economics and psychology (1994; Stanek et al 201113ya). In implementing the CTC analysis here, we followed the procedures described by et al 2019 for covariate adjustment.
[That’s quite a bit of confounding, and a far smaller causal effect…]
…Results of the CTC analyses are summarized in Figure 4 & Table S9 (SOM). College-completing MZ twins had higher means on all 4 outcomes than their non-college completing cotwins, although this difference was generally modest in magnitude (d < 0.25) and non-statistically-significant except for occupational status (χ2 = 33.6, p < 0.001, d = 0.54, 95% CI = 0.36–0.72). The college-completing twin also scored on average higher on IQ (mean difference of 1.8 IQ points, 95% CI = 0.4–3.2) and the Personality composite (d = 0.49, 95% CI = 0.32–0.67), although adjusting within-pair differences on social outcomes for these potential confounders had minimal effect on estimates (Table S9).
…Consistent with earlier research, we found that completing a college degree was associated with all 4 social outcomes, with the magnitude of the college effect being large for occupational status, moderate for income and financial independence, and modest, but still statistically-significant, for legal problems. Importantly, the magnitude of the college effect on these outcomes did not vary statistically-significantly by GCA level. College was neither the great equalizer (ie. it did not reduce GCA differences, 2011) nor a producer of a Matthew effect (ie. it did not expand differences, et al 2015). Those with low levels of GCA appeared to benefit from college to the same degree as those high in ability. Importantly, we were able to replicate this key finding from the Minnesota sample in the independent NLSY97 sample.
…Limitations: …Second, the social and economic outcomes we investigated might be considered a low bar for assessing the benefits of higher education. We believe it likely that our results would have been different had our focus been on the extremes of intellectual achievement (eg. patents, scientific publications) shown in previous research to be associated with very high GCA ( et al 2008). Our results may also look quite different in 10 years, when the sample reaches their prime career years [measurement error].
[A more qualitative comparison would have made this much more interesting: how did they get degrees? Which degrees? What do their colleagues think? What do they think? etc.]