DATA GUIDE

How value-added is calculated

For each college with the required data, we use a model to predict the median earnings for a given student cohort 6- or 10-years after their enrollment in the college.

As explained in the report’s appendix, value-added equals the difference between the actual (observed) median earnings and the predicted earnings based on our model. This gap can be attributed to observable aspects of college quality—curriculum value, STEM orientation, graduation rates, and faculty salaries—and to unmeasured aspects of college quality which we refer to here as an "x-factor."

In this analysis, the primary measure of value-added uses the median earnings of the student cohort that enrolled in the 2001-02 academic year and is measured 10 years following enrollment. The earnings data are from the U.S. Department of Education's College Scorecard database.

We also present the change in value-added over time by calculating value-added six years following enrollment for the student cohorts that enrolled during the 1997-98 and 2005-06 academic years. We present the difference between the two as a measure of change in value-added.

Finally, for comparison with the College Scorecard data, we present value-added using median earnings from Payscale.com for undergraduate alumni who completed their degree and have at least 10 years of work experience post-graduation. See the following section for a discussion of earnings data from both sources.

Value-added indicators are assigned scores ranging from 1 to 100, with 100 indicating the highest centile, or top 1 percent of 2- or 4-year colleges. Colleges are designated as 2- or 4-year based on the percentage of undergraduate awards granted in 2002, to approximate the circumstances of the college for the 2001-2002 cohorts.

Sources of earnings data and known bias

The College Scorecard earnings data are only for federal aid recipients, about 50 percent of all undergraduate students for the 2001-2002 cohort. Aid can be in the form of a grant or loan from the federal government. On average, these students come from lower income families than their non-aided peers and have lower earnings after enrollment than their non-aided peers. When available, we provide information on the percentage of students who received loans and an estimate for the percentage represented in the Scorecard data (download our full database for more details). For some colleges this aided cohort will not be representative of the majority of their alumni. For these reasons, we also provide earnings data from Payscale, which includes data from any degree-completing alumni with at least 10 years of work experience. These data are likely biased in that Payscale users are a non-random sample of the alumni population. Both metrics should be interpreted with caution and considered only as partial pieces of information about a college and the success of its alumni at a given time.

A further limitation of the Scorecard data is that earnings are pooled across campuses and not reported for specific campuses for some colleges with multiple campuses. We have provided a value-added score for these colleges but the earnings data are likely inaccurate and will under-rate main campuses and over-rate branch campuses with lower admissions standards.

Aspects of college quality that determine value-added

Curriculum value

The labor market value of the college’s mix of majors, as determined by the earnings of all U.S. graduates with degrees in a given major.

U.S. Census Bureau American Community Survey; the data were obtained from the University of Minnesota's Integrated Public Use Microdata Series (IPUMS).

Share of graduates prepared to work in occupations requiring STEM knowledge

The percentage of graduates who complete a degree in a field of study that prepares them for an occupation demanding high levels of science, technology, engineering, or math knowledge.

Analysis of data from O*NET and the U.S. Department of Education's IPEDS database

Graduation rate

The percentage of enrolled students who graduate from the college in six years for four-year programs and four years for two-year programs.

U.S. Department of Education's IPEDS database

Average salary of instructors

The average compensation of all instructional staff at the college.

U.S. Department of Education's IPEDS database

X-factor

The unmeasurable qualities or errors in the prediction model that account for the difference between total value-added and the contributions to value-added from the above (measurable) qualities. It can be thought of as intangible school factors that affect student outcomes.

Derived from analysis (see appendix section of report)

Student characteristics that determine predicted outcomes

Average family income

U.S. Department of Education's College Scorecard

First generation share of students

U.S. Department of Education's College Scorecard

Math SAT/ACT scores of students (imputed when missing)

U.S. Department of Education's IPEDS database

Female share of students

U.S. Department of Education's IPEDS database

Asian student share of enrollment

U.S. Department of Education's IPEDS database

Hispanic student share of enrollment

U.S. Department of Education's IPEDS database

Foreign-born student share of enrollment

U.S. Department of Education's IPEDS database

Percent attending part time

U.S. Department of Education's IPEDS database

Average age of students

U.S. Department of Education's IPEDS database

College geographic and type characteristics that determine predicted outcomes

Sub-bachelor’s share of undergraduate awards granted

U.S. Department of Education's IPEDS database

Local price index 2012

The cost of living relative to other metropolitan areas.

Bureau of Economic Analysis

State location

U.S. Department of Education's IPEDS database