The College Scorecard is an effort
to increase transparency in post-secondary education, measuring how
well individual schools perform at preparing students to be successful.
The data is provided to help students and families compare college costs and
outcomes, arming them to make informed decisions about the future and goals.
The data in this project is not actually surveyed by the project itself, but is instead drawn from a number of different data collections from other federal entities. The Scorecard combines the data from the following federal collections:
Students are grouped into cohorts based on their entry year into higher education using the date they are first shown to be receiving financial aid as well as their self-reported grade level on their FAFSA documentation. Students' earnings and debt metrics in the Scorecard are then based on their performance within these cohorts.
Because the Scorecard is simply a conglomeration of a few different datasets, the contents frequently overlap with our imported primary sources. At Ididio, our main use of the Scorecard is to evaluate outcome measures for students using Title IV funding, and these measures are originally from a combination of NSLDS and IRS data which we only access through the Scorecard. The NSLDS data contains information about where economically a student and their family lies prior to their education, while the IRS tax records show subsequent financial information for graduates. Earnings information is not available on a program level, so there is no differentiation between, for example, a student in a medical program and a student in a business program. We primarily value the data for its window into student debt repayment.
College Scorecard data are downloaded from the Department of Education. We last retrieved this data on May 11, 2021.
In November 2019, College Scorecard released program-level loan and earnings data at the CIP-4 (see CIP Codes) level for most institutions. Ididio shares this data in two ways. First, we aggregate the program-level data to create a robust picture of debt and starting salaries for each program to help people broadly which choosing which degree or certificate to pursue. Second, we share the data directly in our program listings for each institution. In both cases, but especially in the latter case, there are issues with the way this data is collected that warrant a cautious interpretation of the values shared:
The loan and earnings data are created using student lists from the National Student Loan Data System (NSDL). That means that any earnings from students who did not use federal loans could be omitted from the reporting. The exception is that if a student borrowed for a previous degree, then that student is in the NSDL system, and the earnings are calculated.
There are unreleased suppression rules that cause many institutions' data from being fully shared. In particular, there are many cases in which we have some loan or earnings data, but no information to allow us to determine the percentage of program graduates for whom the earnings or loan data applies.
While the institution-level earnings data from College Scorecard are based on all students with federal loans who attended, regardless of completion status, program-level earnings data are necessarily only based only on students who actually finished a degree and who are working. Therefore, a program can look appealing when in reality a large percentage are unable to complete a credential or obtain a job.
The data is reported at the CIP-4 level which may combine several institution degree fields.
The Department of Education's now-defunct [Gainful Employment(https://studentaid.gov/data-center/school/ge) standards are difficult to locate directly on the Federal Student Aid site because they are no longer calculated. They define discretionary income to be earnings minus 150% of the poverty guideline, which we determined to be $12,060 in 2017 (the second earnings year collected by College Scorecard) based on the Department of Health and Human Services's threshold estimate for the country as a whole. The guidelines suggest the loan payment should be less than 8% of annual earnings, and less than 20% of discretionary income. When payment amounts are greater than 12% of annual earnings or 30% of discretionary income, those are outright failing, with debt-to-earnings ratios in between these thresholds flagged as concerning. Using the data from College Scorecard, we are able to recreate the gainful employment labels for each program-school combination. As the Institute for College Access and Success notes, our calculations cannot reproduce the Gainful Employment standards are they were intended because the original calculations looked at the debt for all students including non-borrowers, while College Scorecard figures only calculate the debt load for borrowers. Therefore, the debt load will be higher in these calculations, causing more colleges to fail the standard. The original calculation was for federal debt only, and the new calculation includes private loan debt. Additionally, the original calculation was able to compare to the higher of a median or mean salary of graduates, while the median salary is the only value provided in College Scorecard.
While our calculations are not based on the precise data originally gathered for gainful employment standards, we hope they fairly capture the spirit and intent of those calculations. We only flag schools that fail both standards, which should ameliorate the more stringent results from the differing loan calculations and single salary comparison.