Why It’s So Hard to Tell What Low-Income Students Pay at Elite Colleges

A while back, we were curious about how affordable various elite colleges were for low-income students. The federal government had recently started collecting college-level data on net prices—what students pay after grants and scholarships—by income group. That seemed like a good place to start.

We knew that the data had a big limitation: Only students who receive federal financial aid are included. But like others who’ve used the data, we figured it would still be pretty reliable when it came to the lowest-income students, since most of them receive federal Pell Grants.

So we looked for colleges that appeared to be similar in most ways—the size and academic profile of their student bodies, the size of their endowments, etc.—yet that left low-income students with vastly different prices. One such pair: Pomona and Swarthmore Colleges.

How Two Colleges Compare

We looked at the average net prices students at each college paid in 2011-12, the most recent data available at the time. At Pomona, students in the lowest income band, whose families make up to $30,000, were paying an average of $2,751, the data show. At Swarthmore, students in the same income band were paying an average of $10,793. Similar colleges, very different prices.

But then we talked to the representatives of the two colleges. Pomona and Swarthmore have similar financial-aid policies. Contrary to the reported data, it actually sounded like Swarthmore might be a bit more generous.

What was going on?

We noticed one clue: a note that Swarthmore included with its responses to the government’s data-collection survey. It suggested that the numbers Swarthmore reported didn’t really show how affordable it was. That’s why its net price for low-income students looked so high.

Many wealthy private colleges use two financial-aid systems. Like all other colleges that award federal financial aid, they have to use the government’s formula to give students its money.

But the colleges also examine families’ finances more closely (looking, for instance, at the income of a noncustodial parent) before awarding their own institutional aid. Those two systems each calculate an income for families. To categorize students’ income in its data reporting, Swarthmore used the federal financial-aid definition of income. So the data don’t match Swarthmore’s own sense of what families can pay. In other words, in the data it gave the government, Swarthmore included in its low-income category students it doesn’t think are poor.

Pomona, on the other hand, used its own income calculations to categorize students in the data it reported.

This realization eventually led to an article we published last week. That article has prompted discussion and raised more questions. Here are answers to some that you may have:

How do you know that colleges are categorizing students using two different measures of income?

We asked them. The Chronicle sent a quick, anonymous email survey to the financial-aid directors at U.S. News & World Report’s top 25 liberal-arts colleges and its top 25 national universities (minus the service academies). We asked which measure of income they used to categorize students. Of the 27 administrators who responded, 18 said they categorized students using the federal financial-aid definition of income. The other nine used their own calculations of income.

Does categorizing students one way or the other really change the average prices that much?

Some Chronicle staffers wondered the same thing when we told them about Pomona and Swarthmore. Those two colleges had concerns about sharing more data with us, so we weren’t able to pin down just how much of the difference in their net price for low-income students was caused by the different incomes they used to categorize them.

But the University of Notre Dame let us see how its net prices would look categorizing students using one income calculation versus the other. And the difference was striking. Using the federal financial-aid income definition to categorize students (as Notre Dame actually does), the average price paid by members of the lowest-income group was $11,626. Using the college’s own calculation of income, it dropped to $4,472. That is what we illustrated in last week’s article.

Why aren’t colleges all reporting their data the same way?

Average net price by income has been collected and reported for only a few years. Some college officials read the government’s directions as telling them to categorize their students according to the federal income definition even if they don’t think that makes sense. (Hence the note Swarthmore included with its numbers.) Other college officials take the instructions to mean that they should categorize students using the incomes their colleges use to determine institutional financial aid. Are some colleges trying to game the system? Maybe. But there seems to be a genuine difference of interpretation.

So which way is right?

How the colleges should be categorizing students is a question for the federal government. When we asked this week, a spokesman did not give an immediate answer beyond pointing to the instructions on the form the colleges fill out—the same instructions that led to different interpretations by different colleges.

For now, colleges reporting each way have a point:

The argument for using the federal-aid definition: The majority of colleges use only this measure of income anyhow. If the colleges that look at income both ways reported using the federal one, then the definition would be consistent across colleges. Besides, the data are collected in a government survey about federal student aid. Like it or not, the government must want its definition of income used here.

The argument for using the institutional definition: This is the income the wealthy, elite colleges are actually using to award their own aid. If one purpose of publishing the data is to let low-income students know which colleges have generous aid, it only makes sense to use numbers that show what the colleges are actually doing. Otherwise the numbers will be misleading and discouraging, especially for the poorest students.

The Chronicle does not have a horse in this race. We’d just like to know what the reported numbers actually represent.

Can we still use these data to compare colleges?

As the examples of Notre Dame and of Swarthmore vs. Pomona illustrate, the way colleges categorize students’ family income makes a big difference. So we’re not comfortable comparing colleges on their average net price by income unless we know that the colleges we’re comparing have categorized income the same way.