Recently, several news outlets wrote alarming articles about how selective public universities have been reducing the share of low-income students they enroll. The articles cited new research showing that two-thirds of selective public universities enrolled a smaller share of these students in 2010 than in 1999. Reporters leapt at the chance to connect this big finding to equally big trends in higher education: declining state funding; increasing merit aid for wealthy students; greater out-of-state recruiting; even rising economic inequality and feelings of being “left behind” that have shaped American politics since last year’s presidential election.”

However, the media have grossly overinterpreted the new finding, which comes from a report by the think tank New America. Simply counting the number of universities that have seen declines in the share of low-income students they enroll exaggerates the trend. Using the same data, time period and set of schools as the New America report, we factor in how many students attend each of those universities and the magnitude of the decline at each one. The results show that there has been hardly any change in the share of low-income students at selective public universities.

The data for the New America report, and the analysis to follow, comes from Raj Chetty and his team at the Equality of Opportunity project. The New America report defines low-income as the bottom two quintiles of the income distribution, and it defines a “selective” public university as institutions that are not open enrollment (borrowing the category from Chetty, et al.). After removing some schools for which data are missing, the study covers 381 institutions.

As our analysis shows, different income groups’ access to selective public colleges has been remarkably stable across the years covered in the study. Between 1999 and 2010, the share of students at these colleges from the bottom 40 percent of the income distribution declined by less than a percentage point, from 20.6 percent to 20.1 percent. The share of students from the top income quintile did increase, from 37.0 percent to 38.4 percent, while the share of students from the middle class declined the most, from 17.5 percent to 16.8 percent.

Indeed, those trends are all moving in the direction that the press coverage of the New America report would have us believe — more wealthy students, fewer poor students. The issue is magnitude. These are arguably very small changes, and are difficult to square with headlines and articles that imply the changes are large, alarming and proof that universities are responding to state budget cuts by increasingly “shutting out poor students.

That is not to say the finding from the New America report is wrong. It is perfectly accurate. But the chart we have shown is also accurate, and it is important to understand why both findings can be correct. for they imply different interpretations — and most importantly — different policy responses.

While it is true that two-thirds of selective public colleges saw a reduction in their shares of low-income students, the drop was small at many schools. For example, West Virginia State University, which saw a drop in low-income enrollment of just 0.05 percent, is treated the same as schools with large reductions, such as the University of Montana Western, which saw a drop of 18.6 percent. They all count in the “two-thirds” figure, despite the fact that most of the reductions in low-income enrollment are closer to West Virginia State’s than University of Montana Western’s.

Second, universities where the distribution of students from different income groups has remained fairly constant over time tend to enroll more students. Universities that saw less than a 2-percentage-point change in low-income students’ enrollment share are, on average, 66 percent larger than colleges that saw a reduction in low-income share of more than 5 percentage points. Yet small, outlier colleges are given the same weight in the New America report as large colleges with relatively stable enrollment patterns, which also happen to enroll most students attending selective public colleges.

Put another way, the data include many large universities with positive and negative changes in the share of low-income students they enrolled, but these changes are generally small. By contrast, smaller institutions with large drops in low-income students’ shares of enrollment punch above their weight when one looks only at universities. These outlier schools give the impression that access to selective public colleges for low-income students has deteriorated substantially, when in reality the situation has changed very little.

The ramifications for policymakers are important. Those who internalize the media’s interpretation of the two-thirds statistic risk overreacting. Policymakers are bound to reach for broad, cross-cutting solutions if they are to counteract trends they believe are happening at two-thirds of universities, when solutions actually need to be narrowly tailored for a subset of universities. Of course, some policymakers might not be persuaded to do anything after seeing how little things have changed.

Jason Delisle is a resident fellow at the American Enterprise Institute. He formerly worked at New America, but had no involvement with the report on low-income students.

Preston Cooper is a research analyst at the American Enterprise Institute.