Leaving Academia: The Pros and Cons, Weighed Against the Evidence

A social sciences lecturer in her third year at a UK university did the arithmetic last summer. Counting only the hours inside her contracted week, she had spent twenty on research in a year. The rest went on teaching, marking, committees and the work of colleagues whose posts had been cut, with the writing pushed into evenings so her output kept pace with promotion criteria nobody had adjusted. Her salary was around £40,000. She took the sum to an academic advice forum and asked whether it was worth staying.

That isn't a story about weakness. It's a story about a labour market, and the labour market has numbers. The National Science Foundation's Survey of Doctorate Recipients counted 908,700 people with science, engineering or health doctorates in employment in the United States in 2023. Of those, 40.5% worked in educational institutions and 50.9% in business or industry. In that single year, industry employed more of America's science doctorates than education did.

Which makes the usual framing odd. Leaving academia gets discussed as a loss or a failure, when in most fields it is the statistically ordinary thing to do. The question worth asking isn't whether people go. It's what they gain, what it costs them, and how little either side knows before the decision gets made.

The odds are published, and nobody reads them out

Start with the ceiling. A Royal Society report cited by the editors of Nature Microbiology put the share of PhD graduates who go on to secure a university professorship at 0.45%. Field and country move that figure, and permanent research posts sit outside it, but no reasonable reading makes a chair the expected outcome of a doctorate.

Destination data shows where the rest go. In the 2023 Survey of Earned Doctorates, engineering recorded the lowest share of new doctorate holders with non-postdoctoral academic jobs, at 10%, against 80% heading into industry. Biomedical sciences sent 61.6% of graduates with jobs lined up into industry, down from 66.5% the year before, and 22.1% into academic posts. Of that academic group, roughly three-quarters were in non-tenure-track roles, up from 66% in 2021.

Ambition has been adjusting for years. In Nature's 2022 global survey of graduate students, 48% said they would ultimately prefer to work in academia, down from 56% in 2019, while 28% named industry. The preference is shifting faster than the training built around it.

What leavers report, once they're out

The best evidence from the other side of the move is a survey of 856 former research staff from universities and institutes across Europe, run for Vitae by the Careers Research and Advisory Centre. These weren't people who couldn't cut it: 76% had published as principal author and 26% had won competitive grant funding. Eight in ten had wanted an academic career when they began.

Their reasons, in order of agreement: better long-term prospects (76%), more job security (75%), no longer being prepared to work on short and fixed-term contracts (69%), better work-life balance (56%), a better working environment (51%), better salary potential (49%). Only about half said they couldn't secure a position in their field. The dominant motive wasn't rejection. It was a preference for a contract that renewed.

What they found is the part usually left out of the warnings. Nature's editorial on the findings recorded that more than four in five were satisfied in their new jobs, and that many stayed close to research through publishing, outreach or administration. Four-fifths said they still drew on the capabilities they had built as research staff. Two-thirds named communication as the competency their current job demanded most, ahead of the subject knowledge they had spent a decade acquiring. The skills travelled. The job title didn't have to.

Photo by Brett Jordan on Unsplash

The bill: identity, autonomy, and a pension that doesn't travel

None of that makes the move cheap. The Vitae study found the hardest part was the transition rather than the destination, with respondents describing culture shock — less autonomy over their own time, and a pace set by other people's deadlines — along with a loss of social identity, sharpest among those who called themselves "still an academic researcher at heart". Women were roughly twice as likely as men to report difficulty with the loss of status and self-image, and twice as likely to say they had struggled to decide whether to leave at all.

There's a complicating finding in a 2025 paper in Studies in Higher Education, drawn from 780 people who had left or were considering leaving UK universities. Workload was the most cited drawback, named by 53.4%, with management concerns at 45.6%. The benefits they named were students (29.9%), colleagues (29.2%) and flexibility (28.8%). As the authors argue, the two lists aren't opposites: the autonomy academics prize is the same permission that lets workload expand without limit. The thing people miss and the thing that drove them out can be one thing. That's why the exit often feels like ambivalence rather than escape.

Money arrives later than people expect, and usually as a comparison error. The NIH's 2026 stipend notice sets postdoctoral pay on its training and fellowship awards at $63,480 in the first year, rising to $77,076 at seven or more years of experience, and many institutions use those levels as a floor. A first industry salary tends to get measured against that floor rather than against a faculty salary the researcher never held. Pensions are the quieter cost: benefits accrued in a university scheme don't automatically follow you into another sector, and portability is one of the specific problems the OECD has asked governments to fix.

For international researchers there's a harder constraint. Universities and their affiliated nonprofits can sponsor H-1B visas outside the annual cap, which means a postdoc or lecturer on a university visa has never been through the lottery. Moving to an employer subject to the cap means entering it. The workaround that exists — holding a genuine part-time appointment at a cap-exempt institution while working concurrently elsewhere — depends on the university post continuing, which turns the academic job into a dependency rather than a stepping stone. Some researchers qualify instead for self-petitioned green cards as people of extraordinary ability or under a national interest waiver, but those routes turn on documented impact, not on the degree.

The commitment sits in the strategy; the budget sits elsewhere

Universities have got better at saying the right things about career development. Practice lags. In Nature's 2025 survey of 3,785 PhD students across 107 countries, 28% agreed their supervisor had useful advice for careers outside academia, 42% that their supervisor made time for frank career conversations, and 52% that they were encouraged to attend careers training. The people best placed to advise on the exit are, in most departments, the people least equipped to.

Where institutions have acted, the interventions are concrete and modest in cost. The Biotechnology and Biological Sciences Research Council requires a three-month internship within its Doctoral Training Partnership programmes, which puts a student inside a patent office, a journal or a medical writing team before they have to guess what those jobs are like. Prosper, launched in 2019 by three English universities with Research England funding, built its model around employers and principal investigators rather than careers advisers alone. The OECD's 2023 policy paper on doctoral career pathways recommended that funders and institutions track where researchers actually go, publish it, and stop treating non-academic experience as a mark against candidates in academic recruitment.

Now look at the year those recommendations landed in. The University and College Union counted cuts equivalent to more than 15,000 jobs announced at UK universities in the twelve months to October 2025, up from roughly 5,000 the previous March. The Office for Students reported that 43% of universities were forecasting a deficit for 2024-25. The sector's pay offer for 2025-26 was 1.4%. A careers service can explain the outside world beautifully; it can't make the inside one secure.

Testing the move before committing to it

Most of the costs above can be sampled at small scale before anyone resigns. The clearest pattern among people who made the transition well is that they built evidence outside the seminar room first — a placement, a committee with external partners, freelance editing, a data project for a nonprofit — and used it to test whether the work suited them, not merely whether the salary did.

Some labs now make that easier by design. Writing in The Transmitter, one neuroscientist describes staffing part of her group with former doctoral students and postdocs who hold industry jobs and contribute a fraction of their time as project scientists or statistics consultants, co-authoring papers as they go. Shared posts split between two research groups are another version of the same idea. These arrangements cost a principal investigator less than a full position and keep a route back open.

Two practical notes for anyone doing this while on a visa. Ask prospective employers whether the organisation is subject to the annual H-1B cap, rather than whether it sponsors visas; university HR and corporate HR answer those two questions differently, and the first one is what decides your year. Then read the terms of any part-time academic appointment before relying on it, because the concurrent-work arrangement holds only for as long as the exempt post does.

On the application side, the translation problem is real but mechanical. Christopher Caterine, a classics PhD who left for consulting, went through more than 150 informational interviews during his own move and wrote the method up as a book. Find people doing the job, ask what the week actually looks like, and let them tell you which of your skills they would pay for. An academic CV is a record of production. The document that gets you hired is an argument about a problem you can solve.

Photo by Zanyar Ibrahim on Unsplash

What could be counted next

The odd thing about a system that produces so many leavers is how little it knows about them. Vitae's most recent analysis of UK Graduate Outcomes data found 41% of 6,020 working doctorate holders in higher education, down from 49% in an earlier cohort, with 14% in research roles outside higher education against 10.2% previously. That's a destination map, and it comes from graduate survey data rather than from any institution's own record of the people who left its payroll last term.

A university that wanted to understand its own exit could measure four things without new legislation. Where its departing fixed-term researchers went, at twelve months and at three years. Whether those who left said careers advice had been useful, asked at the point of leaving rather than the point of arrival. How many fixed-term contracts were converted to open ones. And what share of doctoral supervisors had been trained to talk about work outside the institution, which is a staffing decision and therefore a budget decision.

Those four numbers would settle an argument currently running on anecdote. They would also tell a lecturer with twenty hours of research time in her week something no forum can: whether her department is an anomaly or the pattern, and what the people who left before her actually found.