Is college still worth the cost as tuition continues rising and underemployment rates remain high among graduates? In a report released late last year, Gallup and Purdue University relayed the answers of college graduates themselves—a nationally representative sample of 30,000 Americans with Internet access and at least a bachelor’s degree.

Overall, half of graduates “strongly agree” that their college education was worth the cost, and another 27% agree less enthusiastically. Several factors were linked to grads’ positive appraisal:

  • Alumni of institutions more highly ranked by U.S. News & World Report were more likely to strongly agree that college was worth the cost, though the rankings “account for about one-third of the variation” in responses.
  • Those who attended public (in- or out-of-state), private nonprofit, and research universities saw their college educations similarly, while graduates of private for-profit colleges were more negative. Only 26% of these Americans—who are disproportionately likely to be racial/ethnic minorities with a heavy student debt load—strongly agreed college was worth it, and 28% disagreed or strongly disagreed.
  • People who graduated college in the last ten years are significantly less likely than others to consider their education worth the cost (38% strongly agreed; 31% of the underemployed). But recent grads earning at least $60,000 a year are about as positive as the overall sample of college graduates.
    In a multivariate analysis accounting for employment status, student loan debt, and personality traits, several factors that are harder to measure, and harder for institutions to foster, also emerge as important in alumni’s assessment of their education. The top three were having at least one professor who cared about the respondent as a person, having a mentor who encouraged the respondent to pursue his or her goals, and having a professor who made the respondent excited about learning.
  • Finally, and predictably, the more student loan debt a graduate has, the less likely that grad is to strongly agree that college was worth the cost for them.

And carrying a significant debt load is not uncommon. Per the report:

Among those recent graduates who received their degrees between 2006 and 2015, 63% say they took out student loans for their undergraduate education, with the median reported amount at $30,000.

Overall, 35% of recent graduates took out loans totaling more than $25,000, which is the level at which debt burden appears to have a more serious impact on graduates’ lives.

Non-trivial minorities of these Americans say their debt has caused them to delay getting married (19%), having kids (26%), moving out of their parents’ home (27%), buying a home (43%), and other life milestones.

A new analysis by Brad Hershbein at the Brookings Institution calls attention to another element affecting whether college pays off for individual Americans: the income of their parents. Data from the Panel Study of Income Dynamics show, in Hershbein’s words, that:

College graduates from families with an income below 185 percent of the federal poverty level (the eligibility threshold for the federal assisted lunch program) earn 91 percent more over their careers than high school graduates from the same income group. By comparison, college graduates from families with incomes above 185 percent of the FPL earned 162 percent more over their careers (between the ages of 25 and 62) than those with just a high school diploma.

It’s easier to digest in the form of this striking chart he constructed:

Hershbein brookings college chart

Since we know richer students tend to have higher standardized test scores in high school and attend more selective colleges, it’s not entirely surprising that they would reap larger gains from their college degrees. Still, the extent of the gap startled me. Hershbein and his colleagues are currently examining possible explanations for it. In the meantime, their findings look like one more piece of evidence that college is not the great equalizer many of us hoped it would be.