Graduate school earnings: worker surveys vs program data

Graduate school has more than one earnings statistic

Students weighing a master’s, doctorate, or professional degree see conflicting numbers online. Georgetown CEW reports worker-level premiums in the American Community Survey. College Scorecard reports program-level medians after completion. The Bureau of Labor Statistics reports wages for occupations that typically require graduate training. Each is valid; none is interchangeable.

What CEW shows about workers who add degrees

The Major Payoff (Figure 10) estimates about a 29% median earnings premium for graduate degree holders relative to bachelor’s-only workers, with wide variation by bachelor’s field. Dollar gains are largest in health and medical preparatory programs (about $70,000 → $152,000 at the median in CEW’s illustration). Percentage premiums look large in education and humanities because baseline bachelor’s pay is lower—even when dollar gains are modest.

CEW also reports that half of education and public service bachelor’s holders eventually earn a graduate degree—the highest share among broad fields. STEM follows at 44%. High pay fields such as petroleum engineering show small percentage bumps from graduate study because bachelor’s pay is already high.

Also important: CEW worker premiums do not subtract graduate tuition or years out of the workforce.

What Scorecard shows about program completers

Federal program data publish 1-, 4-, and 5-year post-completion earnings—not a 10-year program window. EDsmart’s May 2026 national build covers 370 master’s CIPs with data. At 4 years post-completion, national medians in the top 50 CIPs span about 2.9 to 1 ($327,492 Dentistry vs $114,810 Taxation).

Clinical and health-related CIPs dominate the top of the distribution. A high program median does not prove that every student should borrow for that degree—it describes graduates who finished that program.

Cost and policy context

Graduate borrowing rules and repayment plans have changed in recent federal cycles. CEW (2025) notes policy shifts affecting Grad PLUS and income-driven repayment—relevant when interpreting earnings without subtracting debt service.

Pair earnings with average cost of a master’s degree and the graduate subsection of The College Payoff, 2026.

Professional degrees vs master’s degrees

Law, medicine, and dentistry appear at the top of Scorecard program tables partly because licensure gates entry and programs are long. CEW worker premiums for “graduate degree” mix master’s, doctoral, and professional training. Readers should ask which credential the table actually describes.

EDsmart’s May 2026 pull includes 370 master’s CIPs, 312 doctoral CIPs, and 144 first-professional CIPs with reported earnings.

When graduate school is worth modeling separately

Students considering a master’s in education, social work, or public administration often face lower dollar premiums than nursing or computer science—but those fields may require graduate credentials for licensure or promotion. Percentage premiums can look large when bachelor’s baselines are modest even if dollar gains are smaller than in STEM.

Pair CEW worker premiums with Scorecard program medians and a program-specific net price estimate before borrowing for graduate school.

Doctoral and professional pathways

Doctoral and first-professional programs have separate Scorecard credential levels. Earnings windows still stop at five years post-completion for many cells. Medical and dental programs show very high medians partly because only completers who enter licensed practice remain in the measured cohort. Students considering research doctorates should not extrapolate from clinical program medians.

Compare worker-level premiums in CEW with program data and with discipline-specific employment outcomes before committing to six or more years of study.

Graduate school as insurance, not a guarantee

Master’s and professional degrees often lift medians in Education Pays, but tuition, years out of the workforce, and field choice matter. A teacher with a low-cost M.Ed. may see modest pay bumps; a physician’s path includes residency and licensing delays before peak earnings. Readers should compare counterfactual earnings—what you would have earned without the degree—against total cost including living expenses.

CEW graduate payoff research emphasizes that some master’s fields barely move the needle while others (business, engineering, health professions) show larger spreads. Policy debates about “free graduate school” should separate high-return credentials from those with weak labor-market absorption.

Questions to ask before enrolling

Does your employer subsidize tuition? Are licensure exams required? What is the three-year employment rate for recent graduates in your target program? Scorecard program-level data and state wage records can supplement national medians. Pair this article with EDsmart’s College Payoff 2026 graduate-school section for institution-level context.

Graduate school decision worksheet

List total cost including living expenses and forgone salary.

Find program-level median earnings 4–5 years out, not only national worker medians.

Check employment in your target metro, not only national pools.

How this article fits The College Payoff, 2026

This piece is one of twenty companion articles to The College Payoff, 2026. The parent report synthesizes College Board Education Pays, Georgetown CEW, Census, BLS, College Scorecard, Federal Reserve SCF wealth tables, and EDsmart institution-level findings. Each breakout answers one question journalists and families ask when those sources are quoted side by side.

When federal sources update on different calendars, start from the parent report’s definitions section, then return here for the topic-specific FAQ. Interactive tables and state pages on EDsmart Data—tuition trends, debt-to-earnings by major, attainment by state, STEM outcomes—let you filter national benchmarks beyond what a single government table provides.

If a number here differs from a chart in the parent report, use the parent report and the underlying federal publication as the authority. Breakouts explain how to read those sources; they are not a separate dataset.