How to read earnings by major

Who is actually in the number

The federal field-of-study figures follow graduates of one program at one college who received federal financial aid, and report what they earned a set number of years after finishing. Tax records, not a survey, so nobody is guessing at their own salary.

That also means the figure covers only aided graduates who are working and filing. Students who went straight to graduate school, left the workforce, or never took federal aid are not in it.

The three ways it gets misread

The first is treating a national figure for a field as a prediction for one college. Computer science graduates of two colleges two hundred miles apart can differ by a factor that swamps the difference between fields.

The second is comparing fields without holding the college constant. Much of what looks like a gap between majors is a gap between the colleges that happen to offer them, and between the students those colleges admit.

The third is reading a small cohort as a stable number. Where a program graduates a handful of aided students a year, one investment banker moves the median. Very small cohorts are suppressed entirely for privacy, so a blank is not a zero.

What to do with it instead

Use it to rule things out, not to rank things in. A field that pays badly everywhere is telling you something real. A field that pays well at one college and poorly at another is telling you about those two colleges.

Fields by school shows which colleges report awards in each field and where earnings are published, and any college’s own page carries its outcomes beside its cost.