High-scoring, low-income students no more likely to complete college than low-scoring, rich students

In the New York Times on March 7, Charles Murray offers some solutions to the class divide, then dismisses them nearly as quickly as he mentions them on the grounds that they wouldn’t actually work or aren’t necessary. Whether his facts on the class divide are accurate is not the subject of this post, but rather a closer look at a couple of his “solutions.”

Murray makes some decent points about the problems with unpaid internships and the benefits they may afford only those who come from families wealthy enough to allow such experiences. Aside from offering children of well-off parents the ability to pad their resume with unpaid internships , my colleague Ross Eisenbrey argues further that illegal unpaid internships are a scourge on the labor market. Murray rightly states that, “Internships that pay the minimum wage are still much more feasible for affluent students than for students paying their own way through college.”

The part of his article that I take issue with are his arguments about access to higher education. Murray suggests replacing ethnic affirmative action with socioeconomic affirmative action (an argument for another day), then later dismisses it as unnecessary, because “a high proportion of academically gifted children from the working class already get scholarships to good schools.”  Let’s take a look at the evidence.

The relevant issue is the quality of education accessible to children from families in different positions in the income scale. The figure below compares the family income of children in the entering classes at top-tier universities.  Nearly three-quarters of those in the top-tier universities come from families with the highest incomes, while 3 percent and 6 percent of the entering class come from the lowest and second lowest income groups, respectively – or, the bottom 50 percent of families.

Still, Murray might argue that those findings represent meritocracy at work, as those from high-income families have, perhaps through their privileged positions, acquired the intellectual tools to succeed at top schools. The second figure belies this argument. This figure shows that even after controlling for academic ability, higher income children are still more likely to complete college. Each set of bars shows the probability of completing college for children based on income and their math test scores in eighth grade. For example, the first set of bars (for the students with the lowest test scores) shows that 3 percent of students with both low scores and low incomes completed college, while 30 percent of low-scoring children from high-income families managed to complete college.

The fact that college completion is higher for each successive income group among similar scoring students is evidence against a completely meritocratic system. The pattern implies that at every level of test scores, higher income led to higher completion rates. The key comparison in this figure is the fact that high-scoring students from low-income families complete college at nearly the same rate as low-scoring, high-income students (29 percent vs. 30 percent). In other words, high-scoring, low-income children are no more likely to complete college than low-scoring, rich children.

In no way do these data suggest that a high proportion of children (gifted or not) from low-income families achieve placement or completion at universities (and definitely not top schools).

  • Benjamin Hodes

    While I find your arguments impressive, doesn’t the 2nd figure provide insight into a different question- that of college completion rates rather than access to top schools. Your points are well taken but I don’t think they quite refute Murray’s contention.

    Benjamin Hodes
    Pittsburgh, PA

    • Dave

      Was this some sort of standardized math test score? Or was the just a percentage of how the students did within their own income groups?

  • Andrew Campbell

    Any possible explanations for this? Is it due to financial resources. Low income have to drop out more but high income don’t?

  • Alan

    What happens if you look only at two-parent households?

  • John Secular Smith

    This appears to be the source of the data.


    I’m having trouble finding where you actually got that table.  Any page numbers or references you can send our way?

    • Kringsat

       Table 21, p. 50. “Percentage distribution of 1988 eighth-graders’ educational attainment by 2000, by eighth-grade mathematics achievement and selected student characteristics: 2000”

  • Inmyhat

    These data represent college entrance and not the high school they attended. Maybe the real issue is public education reform as private school proves to be working with respect to completing college.

  • Quartile

    Why is this graph avoiding the dollar value of what is considered a ‘top’ income quartile? Because this might be at a top tier private school as opposed to a top tier public school? It does make a huge difference.