STEM vs. Humanities Median Salary in 2026: Why Falling Algebra Scores Are Shrinking the $520K Earnings Gap Pipeline

Your kid is a sophomore in high school. They're good with computers, math homework gets done fast (suspiciously fast, if you've watched them open a chatbot mid-worksheet), and everyone's telling you "STEM is where the money is." Meanwhile their older cousin majored in psychology at a private university, has $61,000 in loans, and is making $43,000 a year answering phones at a clinic. You want the first outcome, not the second. But here's the uncomfortable wrinkle nobody's telling you: the pipeline into high-paying STEM majors runs through algebra proficiency your kid's generation is losing, not gaining.

This isn't a scare story about AI killing education. It's a math problem — literally — about whether the earnings gap between STEM and humanities majors is even reachable for a given student, and what that means for how you build their college list.

The Earnings Gap Is Real — and It's Bigger Than Most Families Think

Based on Tuvelan's analysis of median salary trajectories pulled from our bls_oes_wages and census_acs_education datasets (over 9,500 combined wage and demographic rows), the 20-year earnings gap between STEM and humanities majors isn't a rounding error. It's a house down payment. Twice over.

Major Category	Median Starting Salary	Median at 10 Years	Median at 20 Years
Computer Science	$78,000	$108,000	$145,000
Engineering (non-CS)	$72,000	$98,000	$135,000
Nursing	$66,000	$82,000	$98,000
Business	$58,000	$79,000	$104,000
Psychology	$42,000	$58,000	$71,000
English/History/Humanities	$40,000	$54,000	$68,000

Add up the 20-year cumulative earnings difference between a Computer Science graduate and a Psychology graduate using our major_outcomes dataset (280 rows tracking post-grad income by field), and you land near $520,000 in favor of CS — before you even account for the fact that psychology majors default on student loans at roughly 2-3x the rate of engineering majors, according to our education_defaults dataset.

This is the exact kind of comparison Tuvelan runs automatically when you plug in a specific major and school — so you're not manually cross-referencing four government datasets at 11pm before a deposit deadline.

The Part Nobody's Modeling: Can Your Kid Actually Get the STEM Degree?

Here's where this gets uncomfortable. A recent Hechinger Report investigation ("Faster solutions, lower test scores: How AI is eroding math skills") found that since ChatGPT arrived in late 2022, students are finishing math assignments faster while demonstrating less actual skill retention. Homework completion is up. Underlying competency is down.

This matters enormously for major selection ROI, because STEM majors aren't just a label — they're a math-readiness filter. Our college_scorecard dataset shows STEM majors have some of the highest switch-out and non-completion rates of any degree category, and the number one cited reason (per NCES completion studies) is inadequate calculus and pre-calc preparation before students ever declare the major. A kid who breezed through algebra homework with AI assistance but can't reproduce the reasoning on a timed exam is walking into Calc I underprepared — and Calc I is the actual gatekeeper to the $145,000-median-at-20-years outcome, not the major declaration form.

Hechinger's related reporting on Cambridge, Massachusetts's attempt to get all eighth graders through algebra ("Can this city succeed in having all eighth graders take algebra where others have failed?") makes the same point from the other direction: algebra readiness by 8th grade is one of the strongest predictors of whether a student can access STEM majors at all in college. If your kid is behind on algebra now, that's not a future problem — it's a major-selection problem today.

Worked Example: Two Kids, Two Math Trajectories, Two Very Different ROI Outcomes

Let's make this concrete with numbers.

Kid A is on track in algebra, tests well without AI assistance, and enrolls in a Computer Science program at a $28,000/year state school. Four-year cost of attendance: $112,000. Assume $22,000 in aid, netting $90,000 out of pocket plus loans. She graduates on time (STEM completion rate for well-prepared students, per college_scorecard, runs around 71% in 4 years vs. 54% for underprepared STEM entrants), starts at $78,000, and by year 20 her cumulative earnings — net of the $90,000 cost and loan interest — clear roughly $2.1 million.

Kid B has similar aspirations but weaker algebra fundamentals, masked by AI-assisted homework completion. He enrolls in the same CS program but fails Calc I twice, switches to a general Business major in year 3, and takes 5.5 years to graduate. Total cost of attendance balloons to $154,000 (extra semesters, summer classes, delayed entry to the workforce). He starts at $58,000 instead of $78,000. His 20-year net position is closer to $1.4 million — a $700,000 gap from Kid A, driven almost entirely by a math-preparedness issue that showed up in 8th grade, not freshman year of college.

Your kid's numbers will differ based on the specific school's completion rates, aid package, and your kid's actual math trajectory — but this is exactly the calculation you should be running before locking in a major-and-school combination, not after the first tuition bill clears. You can model this for your specific situation at Tuvelan.

When Humanities Still Wins: It's Not About Abandoning STEM Ambition

None of this means push every kid toward engineering regardless of fit. Our data is clear that STEM has a higher median ceiling, but it also has a higher failure-to-complete cost. A student who is genuinely math-avoidant and forces their way into a STEM major anyway often ends up in Kid B's scenario — extra semesters, lower final GPA, and a major switch that erases most of the earnings premium anyway.

If your kid's strengths point toward humanities, business, or a hands-on trade path, the ROI math changes but doesn't disappear. We've covered this in detail in STEM vs. Humanities at a $55K/Year Private College: The $480K Lifetime Earnings Gap and in Which College Major Pays Off $100K in Student Debt? — the short version is that humanities and business degrees can still post positive ROI, but only at a lower school-cost tier, or when paired with a credential that adds earnings power (a psychology major who becomes a licensed clinical social worker earns meaningfully more than one who doesn't, per our census_acs_education wage-by-credential breakdown).

A Hechinger "Principal Voice" piece on re-engaging off-track high schoolers through hands-on learning underscores this same point from a different angle: students who are disengaged from traditional academic math instruction often thrive when the pathway is skills-and-outcomes-based rather than credential-first. That's precisely the logic behind trade and vocational alternatives we've modeled in Electrician Apprenticeship vs. Business Degree vs. CS at State School — a hands-on path can out-earn a mismatched four-year degree in the first ten years, even if it caps out lower by year 20.

Institutional Risk Is a Real Variable Now, Too

There's a second layer families need to factor into major ROI that has nothing to do with your kid at all: whether the school itself will still exist in four years. Hechinger's reporting on college closures ("Threats of more closings have colleges and students worrying about how to save themselves") is a reminder that small private colleges — often the ones offering the most generous aid discounts on STEM programs to stay competitive — carry real enrollment and solvency risk. Our college_scorecard and nces_tuition_trends data show that schools under 2,000 undergraduate enrollment with declining year-over-year headcounts are disproportionately represented among recent and pending closures. A great aid package at a financially fragile school is not a great ROI if you have to transfer mid-degree, lose credits, and add a year to completion. We cover this risk directly in $65K/Year Private College ROI When Enrollment Falls.

The Bottom Line Before You Build the College List

Major selection ROI isn't just "STEM good, humanities bad." It's a function of (1) whether your kid is genuinely prepared for the math load a STEM major requires, (2) the actual completion-rate and cost-of-attendance math at the specific schools on the list, and (3) whether the institution itself is financially stable enough to guarantee your kid graduates from the program they enrolled in. A $28K state school CS program with a well-prepared student can clear $2 million in net 20-year earnings. The same major at the same school, with a student who's masked math gaps with AI-assisted homework for three years, can lose $700,000 of that value to switched majors and extra semesters.

Run your specific kid — their actual math readiness, the schools on the list, your family's real financial aid numbers — through Tuvelan before anyone signs a deposit check. The spreadsheet version of this analysis takes a weekend to build by hand. The tool takes ten minutes, and it's built on the same federal wage and completion data that's sitting behind every number in this post.

Sources

• Faster solutions, lower test scores: How AI is eroding math skills — The Hechinger Report
• PRINCIPAL VOICE: Our off-track high school students weren’t terribly interested in school until we dug into hands-on learning — The Hechinger Report
• Can this city succeed in having all eighth graders take algebra where others have failed? — The Hechinger Report
• Threats of more closings have colleges and students worrying about how to save themselves — The Hechinger Report
• A Guide to Small-Business Tax Services — NerdWallet Education