Yes — field of study matters far more than school type, and the two interact

Summary

Yes, and the effect of what you study is far larger than the effect of where you study it. Nationally, prime-age (25-54) bachelor's degree holders earn a median 70% more than workers with only a high school diploma and face much lower unemployment (2.9% vs 6.2%), per Georgetown University's Center on Education and the Workforce (CEW). But within the bachelor's degree alone, median earnings five years after completion range from about $27,000 (business operations support) to $138,000 (operations research) across fields tracked in the Dept. of Education College Scorecard — a more than 5x spread, among programs large enough to report earnings (see exclusion note below). Institution type (public, private nonprofit, private for-profit) shifts earnings within a given major by a much smaller margin, roughly 10-25%, and mostly reflects a private-nonprofit premium concentrated in already-high-paying fields (e.g. Computer Science: $104K public vs $128K private nonprofit) rather than a general "prestige" effect. Because private institutions also charge roughly double the net price of public ones, the same low-earning major can have a positive return on investment (ROI) at a public school and a negative one at a private school, per Georgetown's ROI modeling.

1. The overall bachelor's-vs-high-school premium

Georgetown CEW's Major Payoff report (Oct 2025) finds prime-age (25-54) workers with a bachelor's degree earn 70% more at the median than workers with only a high school diploma, and have a much lower unemployment rate (2.9% vs. 6.2%). As a simple cross-check using this corpus's ACS data (2023, median earnings by educational attainment, averaged across the 50 states + DC — an unweighted state-average proxy, not a population-weighted national figure), the bachelor's-vs-high-school gap is $64,199 vs $38,995, a 65% premium — directionally consistent with Georgetown's more precise, individual-level 70% figure. A graduate/professional degree adds further, to an average $82,319 in the same ACS proxy.

2. Field of study drives most of the spread

Using Dept. of Education College Scorecard data (this corpus's edu.field_of_study_roi, bachelor's-level programs with 500+ completions, 2022 or later vintage), median earnings 5 years after program completion:

• Highest-paying fields: Operations Research ($138,172), Computer Engineering ($111,687), Computer Science ($111,294), Electrical/Electronics Engineering ($101,615), Petroleum Engineering ($102,241).
• Lowest-paying fields: Business Operations Support ($27,237), Drama/Theatre Arts ($34,676), Culinary Arts ($37,065), Music ($40,923), Fine/Studio Arts ($41,440).

Exclusion applied here: this ranking is built from edu.field_of_study_roi filtered to median_earnings_5yr IS NOT NULL. Of 340,012 total bachelor's-level program-year rows for 2022+, 220,340 (65%) have no median_earnings_5yr value and were dropped before ranking — the College Scorecard suppresses earnings for cohorts too small to protect student privacy, so small programs are systematically missing rather than randomly missing. The rankings above should be read as "among programs large enough to report earnings" (all filtered here to 500+ total completions in the surviving rows), not as covering every bachelor's program; a small, low-enrollment program's true earnings could differ from what any of these figures show, in either direction.

This matches Georgetown's own Major Payoff finding: median earnings for prime-age STEM majors average $98,000 vs. $58,000-69,000 for education/public-service and arts/humanities majors, and even within STEM the range runs from $64,000 (miscellaneous agriculture) to $146,000 (petroleum engineering) — field of study alone spans a wider range than the college-vs-no-college gap itself.

3. Institution type matters much less — and mostly within a field, not across the whole population

Pooled across all bachelor's fields in this corpus (joining edu.field_of_study_roi f to edu.college_scorecard cs on institution and year), average 5-year median earnings by ownership type are close: Public $62,092, Private nonprofit $64,502, Private for-profit $62,024 — a 4% spread, an order of magnitude smaller than the field-of-study spread above. Exclusion applied here: this join query used the same filter, f.median_earnings_5yr IS NOT NULL, on the identical 340,012-row 2022+ bachelor's population, so the identical 220,340 rows (65%) were excluded here as well, for the same College Scorecard privacy-suppression reason described above — these three ownership averages are computed only over programs large enough to report earnings, and smaller institutions (disproportionately smaller for-profits and small private colleges) are under-represented in all three group averages. This lines up with Georgetown's own conclusion that "major choice explains more lifetime earnings variance than school prestige."

Institution type does matter more within a specific major, and the direction differs by field:

• Computer Science, bachelor's: Public $103,767 vs. Private nonprofit $127,504 (private nonprofit CS programs, concentrated among selective research universities, show a real earnings edge); Private for-profit $75,760 (n=32 program-years — small sample, directional only).
• Fine/Studio Arts, bachelor's: Public $40,820 vs. Private nonprofit $43,581 (a much smaller gap than in CS); Private for-profit $28,203 (n=10 — very small sample).

Note the cross-cutting effect: a Computer Science graduate from a public university ($103,767) out-earns a Fine Arts graduate from any type of institution, including private nonprofit ($43,581), by more than 2x. Major dominates institution type.

4. Why institution type still matters for ROI, not just raw earnings

Cost differs sharply by institution type even where earnings differ little. In this corpus, average net price per year (2022+, College Scorecard) is $10,816 for public in-state students vs. $22,921 for private nonprofit and $21,499 for private for-profit — roughly double at private institutions. Georgetown's ROI methodology (Ranking 4,600 Colleges by ROI, Feb 2025, updated Sept 2026) combines earnings and net price directly and finds exactly this interaction: a low-earning major like fine arts still produces a positive 40-year ROI at a public university, but the same major at a private institution charging roughly $55,000/year can show a 40-year ROI below 100% (i.e., a plausible net loss relative to forgone earnings and cost). Georgetown's correlation analysis across its 4,600-institution dataset also finds that graduation rate correlates moderately with ROI at bachelor's-granting institutions (r ≈ 0.42-0.49 for 10-year ROI, public and private nonprofit alike, though not for-profits), while Pell Grant recipient share correlates negatively with ROI (r ≈ -0.33 to -0.35 at 10 years) — institutions serving more low-income students tend to show lower measured ROI, likely reflecting completion risk and other confounds rather than the major/instruction itself.

Data and methodology notes

Sources used: Dept. of Education College Scorecard, program-level (edu.field_of_study_roi, sourced via college_scorecard_programs) and institution-level (edu.college_scorecard) data, queried directly from this corpus, 2022+ vintage, bachelor's-level programs with 500+ total completions to avoid small-sample noise (single-field breakdowns by ownership, e.g. for-profit CS/Fine Arts, retained smaller n's and are flagged directional only). Census ACS median-earnings-by-education (state-level, 2023) used as a simple cross-check, not a population-weighted national estimate. Georgetown CEW's own reports ("The Major Payoff," Oct 2025; "Ranking 4,600 Colleges by ROI," Feb 2025/Sept 2026) supply the individual-level, nationally representative premium (70%) and the ROI/correlation findings this corpus's raw earnings tables cannot compute on their own (ROI needs a discounted cash-flow model over enrollment cost + multi-decade earnings, which Georgetown builds and this corpus's tables do not attempt to replicate).

A separate, institution-level exclusion also underlies the average-net-price comparison in Section 4: of 31,365 edu.college_scorecard institution-year rows for 2022+, 8,865 (28%) lack mean_earnings_10yr and are excluded from any institution-level earnings statistic drawn from that table, again due to College Scorecard's privacy suppression for small cohorts — disproportionately affecting smaller for-profit and specialized institutions.

Limitation: "sector" (public/private/for-profit) is not populated directly on field_of_study_roi; the institution-type breakdown here was built by joining that view back to college_scorecard.ownership on institution ID and year.

Every query behind this report

14 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT cip_title, credential_level, credential_title, COUNT(*) n, AVG(median_earnings_5yr) avg_earn_5yr, SUM(ipeds_awards) total_awards
FROM edu.field_of_study_roi
WHERE credential_level = 3 AND median_earnings_5yr IS NOT NULL AND "year" >= 2022
GROUP BY cip_title, credential_level, credential_title
HAVING SUM(ipeds_awards) > 500
ORDER BY avg_earn_5yr DESC
LIMIT 15

SELECT cip_title, credential_level, credential_title, COUNT(*) n, AVG(median_earnings_5yr) avg_earn_5yr, SUM(ipeds_awards) total_awards
FROM edu.field_of_study_roi
WHERE credential_level = 3 AND median_earnings_5yr IS NOT NULL AND "year" >= 2022
GROUP BY cip_title, credential_level, credential_title
HAVING SUM(ipeds_awards) > 500
ORDER BY avg_earn_5yr ASC
LIMIT 15

SELECT sector, COUNT(*) n, AVG(median_earnings_5yr) avg_earn_5yr, SUM(ipeds_awards) total_awards
FROM edu.field_of_study_roi
WHERE credential_level = 3 AND median_earnings_5yr IS NOT NULL AND "year" >= 2022
GROUP BY sector
ORDER BY sector

SELECT ownership,
  CASE ownership WHEN 1 THEN 'Public' WHEN 2 THEN 'Private nonprofit' WHEN 3 THEN 'Private for-profit' END AS own_label,
  COUNT(*) n,
  AVG(mean_earnings_10yr) avg_mean_earn_10yr,
  AVG(median_earnings_8yr) avg_median_earn_8yr,
  AVG(median_debt_completers) avg_debt
FROM edu.college_scorecard
WHERE "year" >= 2022 AND mean_earnings_10yr IS NOT NULL
GROUP BY ownership
ORDER BY ownership

-- same field (Computer Science, bachelor's) across sector types
SELECT sector, COUNT(*) n, AVG(median_earnings_5yr) avg_earn, SUM(ipeds_awards) awards
FROM edu.field_of_study_roi
WHERE credential_level = 3 AND cip_title = 'Computer Science.' AND median_earnings_5yr IS NOT NULL AND "year" >= 2022
GROUP BY sector ORDER BY sector

SELECT sector, COUNT(*) n, AVG(median_earnings_5yr) avg_earn, SUM(ipeds_awards) awards
FROM edu.field_of_study_roi
WHERE credential_level = 3 AND cip_title = 'Fine and Studio Arts.' AND median_earnings_5yr IS NOT NULL AND "year" >= 2022
GROUP BY sector ORDER BY sector

SELECT cs.ownership,
  CASE cs.ownership WHEN 1 THEN 'Public' WHEN 2 THEN 'Private nonprofit' WHEN 3 THEN 'Private for-profit' END AS own_label,
  COUNT(*) n, AVG(f.median_earnings_5yr) avg_earn, SUM(f.ipeds_awards) awards
FROM edu.field_of_study_roi f
JOIN edu.college_scorecard cs ON f.unit_id = cs.id AND f.year = cs.year
WHERE f.credential_level = 3 AND f.cip_title = 'Computer Science.' AND f.median_earnings_5yr IS NOT NULL AND f.year >= 2022
GROUP BY cs.ownership ORDER BY cs.ownership

SELECT cs.ownership,
  CASE cs.ownership WHEN 1 THEN 'Public' WHEN 2 THEN 'Private nonprofit' WHEN 3 THEN 'Private for-profit' END AS own_label,
  COUNT(*) n, AVG(f.median_earnings_5yr) avg_earn, SUM(f.ipeds_awards) awards
FROM edu.field_of_study_roi f
JOIN edu.college_scorecard cs ON f.unit_id = cs.id AND f.year = cs.year
WHERE f.credential_level = 3 AND f.cip_title = 'Fine and Studio Arts.' AND f.median_earnings_5yr IS NOT NULL AND f.year >= 2022
GROUP BY cs.ownership ORDER BY cs.ownership

SELECT cs.ownership,
  CASE cs.ownership WHEN 1 THEN 'Public' WHEN 2 THEN 'Private nonprofit' WHEN 3 THEN 'Private for-profit' END AS own_label,
  COUNT(*) n, AVG(f.median_earnings_5yr) avg_earn, SUM(f.ipeds_awards) awards
FROM edu.field_of_study_roi f
JOIN edu.college_scorecard cs ON f.unit_id = cs.id AND f.year = cs.year
WHERE f.credential_level = 3 AND f.median_earnings_5yr IS NOT NULL AND f.year >= 2022
GROUP BY cs.ownership ORDER BY cs.ownership

SELECT cs.ownership, AVG(cs.tuition_out_state) tuition_oos, AVG(cs.avg_net_price_private) net_priv, AVG(cs.avg_net_price_public) net_pub
FROM edu.college_scorecard cs WHERE cs.year>=2022 GROUP BY cs.ownership ORDER BY cs.ownership

SELECT "year",
  SUM(median_earnings_hs_graduate) / COUNT(*) AS avg_hs,
  SUM(median_earnings_bachelors) / COUNT(*) AS avg_bach,
  SUM(median_earnings_graduate_degree)/COUNT(*) AS avg_grad,
  COUNT(*) n_states
FROM census.acs_earnings_by_education
WHERE "year" = '2023'
GROUP BY "year"

SELECT
  COUNT(*) total_rows,
  SUM(CASE WHEN median_earnings_5yr IS NULL THEN 1 ELSE 0 END) null_5yr
FROM edu.field_of_study_roi
WHERE credential_level = 3 AND "year" >= 2022

SELECT
  COUNT(*) total_rows,
  SUM(CASE WHEN mean_earnings_10yr IS NULL THEN 1 ELSE 0 END) null_10yr
FROM edu.college_scorecard
WHERE "year" >= 2022

SELECT COUNT(*) total, SUM(CASE WHEN f.median_earnings_5yr IS NULL THEN 1 ELSE 0 END) null_5yr
FROM edu.field_of_study_roi f
JOIN edu.college_scorecard cs ON f.unit_id = cs.id AND f.year = cs.year
WHERE f.credential_level = 3 AND f.year >= 2022