The Peter Orszag graphs of enrollment rates and family income that you posted are not particularly informative, as they omit a lot of variables correlated with income that are known to have large impacts on achievement. In short, the administration is likely ready to mislead us again.

One of the biggest unmentioned confounders is marriage. Married couples are more likely to have higher incomes than single, divorced, or cohabitating couples. See this paper for a representative overview of the evidence. The income differences are substantial even controlling for education, work experience, and other observable indicators of earnings capacity. It is possible that the income differences shown are really proxies for family structure.

Growing up in a stable, married, two-biological-parent family also appears to give children an advantage in educational attainment. See this paper for a representative result using the National Education Longitudinal Study.

Finally, getting into college depends on test scores and grades from high school. Marriage is also a confounding variable for high-school performance. Kids from intact families are less likely to have problems, with the attendant lower grades, in high school.

The income measure itself can be a problem, as government income measures typically include only cash income, ignoring “in-kind” assistance such as Medicaid, housing subsidies, and so on.

One also wonders why they chose math scores — and what math scores they used. SAT? ACT? A couple of minutes with Google and voilà, a similar graph shows up at the notorious epi.org site. Follow the reference given and you end up here, with NAEP 8th-grade math scores.

Who knows a) what math scores they are using (and why they don’t also show verbal scores), b) how selection bias affects the scores they say they have, and c) how whatever scores they use correlate with the tests used for college entry.

In short, without controlling for the social structures that contribute both to income and educational success, and knowing what the heck they are measuring, those graphs tell us very little about what the influences of income are and whether or not college entry is “fair.”

The fact that they apparently are going to be rolled out as evidence that we need higher-education reform controlled by government suggests that a more detailed analysis of the issue should be begun right now.

Linda Gorman

Director, Health Care Policy Center

Independence Institute

Golden, Colo.