Behind the Data: How We Made Debt By Degrees

To make the Debt by Degrees database, ProPublica analyzed previously unreleased data from the U.S. Department of Education and the White House. The data includes new information on student debt levels by income levels, by how well students are able to repay their loans, and by how much students earn, on average, years after graduation. The data comes from three main sources: an annual survey collected by the Department of Education (IPEDS), the national student loan database (NSLDS), and administrative earnings data from tax records.

All data has limitations. A full accounting of the caveats of these data sets is described in a new report that accompanies the data, “Using Federal Data to Measure and Improve the Performance of U.S. Institutions of Higher Education”.

Based on this report, here are some things you should know about the data in our interactive database:

• The data is only representative of students who receive federal aid from Title IV programs. This includes students who take out subsidized or regulated loans, such as Stafford and Perkins loans, as well as Pell Grants. Therefore, the data excludes several types of debt that students take on. For example, Parent PLUS loans and private bank loans are not part of this data set. However, according to the government report, only a small percentage of students’ parents take out PLUS loans (about 4.5 percent), with a median loan size of about $10,000. Few students take out private bank loans to pay for school (only 6 percent), and the loans tend to be small -- about $4,000. The data could slightly understate the full amount of debt, but it is likely not off by much.
• The graduation rates included in the data only count students who are attending college for the first time and are in school full time. It does not include students who have attended college before or those who have transferred into the school. It also excludes part-time students. Data for most four-year schools, as well as short certificate programs, will not be greatly affected by this caveat. But at community colleges, where many students attend part time and others may transfer to a four-year college before graduation, this figure is likely to be affected.
• The cost of tuition is broken down by how much each income group pays for college, based on the amount of awards, loans and institutional aid students get on average. But for public schools, these figures only represent in-state tuition. For out-of-state or international students, tuition is likely to be higher.
• The percent of Pell grant recipients at a school is widely used as a proxy to indicate how many low-income students are enrolled in a school. But international and undocumented students are not eligible for Pell grants and are not included in the Title IV program data.
• The figures related to earnings only include students who received federal loans or aid. Schools with a larger number of Title IV students may have more accurate earnings information than schools with smaller numbers of Title IV students. It is worth noting, however, that more than 70 percent of students receive either a Pell grant or a loan under a Title IV program, so this data is representative of a large cross section of the student population.
• The data on median earnings and the percent of students who make less than $25,000 after graduation includes employed people earning wages as well as unemployed people and voluntarily non-working people. This is not, therefore, an unemployment measure; it could include graduates who are unemployed or doing volunteer work, stay-at-home parents, and individuals who choose not to work, as well as people who work and make less than $25,000.
• The earnings data also includes students who attended a school but did not graduate. Their earnings might skew lower than students with degrees.
• In schools with many students who receive Pell grants, but not federal loans (a common occurrence at some community colleges), graduation rates may be distorted and appear lower than they actually are. Though this primarily affects two-year associate programs, it can also affect four-year schools with “no loan” financial aid policies.
• The cohort default rate has long been used as an indicator of how well students are able to pay off their loans after graduation. But schools can manipulate this rate by pushing their students into deferment or forbearance. The repayment rate, a new measure featured in Debt By Degrees, is a better indicator of how many students are struggling to pay off their student debt, as it includes default, deferment, forbearance, as well as students who are unable to start paying off the interest on their debt.
• For Debt by Degrees, we included only schools in the 50 states and Washington D.C. Additionally, we excluded schools without an undergraduate population and institutions of religious study, such as seminaries, bible colleges, and rabbinical schools.