How to Choose a College Major (Using Data, Not Vibes)

What the adults who already chose would tell you

Across 3,217 career assessments, the most common career problem is not low pay or bad managers. It is “I’m just not happy with what I’m doing,” at 27.3%, ahead of money at 16.6%. Meanwhile 34.4% name earning more as their top five-year goal.

Those two numbers together are the lesson. People chase money because it is the easiest goal to name, then discover the thing that actually ruins their week is the work itself. Finance illustrates it precisely: one of the best-paid fields, and the unhappiest in our data at 64.6%, with only 3.1% saying they love their job.

Choosing a major purely by starting salary optimizes for the variable that matters second.

Start with the thinking, not the job title

Job titles change; the cognitive work is stable. Ask which of these you can do for hours without it feeling like effort:

• Untangling a problem where the answer is not known in advance
• Making something concrete and seeing it work
• Persuading, explaining, or teaching another person
• Organizing chaos into a system that runs
• Finding the pattern hiding in a mass of information
• Caring for people directly, in person

Most people can name their answer honestly in about a minute, and most ignore it because it does not map to a prestigious major. In our data, 50.6% say solving complex problems energizes them while 30.1% enjoy working with data, so these preferences genuinely differ between people rather than being universal.

Test before you commit

You are predicting your own reaction to work you have never done. That prediction is unreliable, so buy information cheaply before spending years.

1. Read the actual curriculum, every required course for all four years, not the brochure. If a third of it looks tedious, that is your future workload.
2. Talk to someone five years in, not a professor. Ask what an ordinary Tuesday looks like. Ordinary, not exciting.
3. Do a small real version before declaring: a summer job, a project, a volunteer role, an online course. Two weeks of contact beats two months of imagining.
4. Notice what you do voluntarily. The subject you read about with nobody assigning it is data you already have.

Weigh automation resistance, because your career is forty years long

You will work longer under artificial intelligence than any generation before you, which makes this a bigger factor for you than for anyone advising you.

The work that holds up combines judgment under ambiguity, direct human contact, physical dexterity in unpredictable settings, and responsibility for decisions someone must own. The work most exposed is structured, repetitive and rule-based, however credentialed it looks today. Every career recommendation we produce carries an AI Stability Score for this reason; the 2,590 recommendations we generated for members averaged 7.4 out of 10. Our breakdown of which jobs survive the AI revolution goes deeper.

What if you genuinely do not know?

Then you are in the majority, and the honest move is to buy time rather than fake certainty. Start with broad foundational courses, delay declaring as long as your institution allows, and use the first year as structured exploration. Choosing wrongly and correcting costs a semester; choosing wrongly and persisting out of stubbornness costs a decade, which is what we see in the adults who come to us at 35 to 44, the unhappiest group we measure at 50.2%.

A structured assessment converts vague preference into a concrete shortlist you can go test. Ours is built for exactly this decision, and a version aimed at students is a good starting point. Then read what to optimize for in your first job, because the major matters less than most people think once that first role begins.

Last updated: July 2026

Written by the DreamJobMatcher Research Team · About the team