Psychology in 2026 has no sovereign question. It has a fight over what counts as a finding, and underneath that fight, a change in what it costs to produce one. The cost change explains most of the status movement, so I would start there.
The reform movement won. Preregistration, multi-site collaboration, adequate power, open data, severe testing. The victory raised the price of a high-prestige result. A clever experiment on ninety undergraduates used to buy a career. Now the price of admission is a consortium, a registry entry, a population cohort, or a randomized trial inside a real institution. Prestige in psychology has migrated from ingenuity to capital. That migration happened just as the American state began withdrawing the capital so fewer places can afford to do work the field respects, and the places that can will increasingly sit outside the United States.
The live questions.
First, whether psychology’s objects exist. When a researcher says depression, intelligence, grit, or self-control, what is he claiming is there? Denny Borsboom and the Amsterdam methods group own this question. Eiko Fried carries it into clinical research, where the network view treats symptoms as producing and sustaining one another. Nobody has answered it. The question has enormous downstream reach, because every other question in the discipline assumes it has been settled.
Second, whether psychology deserves belief. Simine Vazire made that question an editorial program by running Psychological Science, which she has edited since the start of 2024 on a four-year term. Brian Nosek supplies the infrastructure rather than the challenge. Vazire holds the more powerful position, because an editor decides what a finding has to look like before it can be a finding.
Third, whether artificial intelligence explains cognition or imitates its outputs. Iris van Rooij and Olivia Guest at Radboud published the critical case in 2026, arguing that psychology’s weak theory and AI‘s habit of confusing engineered artifacts for minds combine into a perfect storm. Guest puts the practical objection well: the point of science is to produce knowledge, and science cannot be automated. The constructive answer comes from Stanford, where the argument runs that a model trained on human-scale input and subjected to real experiments can serve as a testable theory. Van Rooij is rising because AI made her old demand for formal explanation urgent.
Fourth, what digital life does to young people. This is where public status and disciplinary status split most visibly. Jonathan Haidt (b. 1963) and Jean Twenge (b. 1971) hold the public and legislative ground. Amy Orben and Candice Odgers hold the methodological ground. The interesting development is that Orben stopped arguing and started spending. She and Dan Lewer
are running the IRL Trial in Bradford, roughly four thousand students aged twelve to fifteen, half of them randomly assigned to daily social media limits and a nine at night to seven in the morning curfew, with a comparison group under no restriction. Orben calls it the first large-scale experimental reduction of social media use in healthy minors anywhere. Read her own words carefully and the standard framing collapses: she says the evidence that social media harms individual children, including severe harm, is irrefutable. The dispute concerns population-level causal size, not whether harm occurs. Her group also expects results on the Australian under-sixteen ban within about a year. Within eighteen months this question gets an answer, and whoever is right collects a great deal of status.
Fifth, what genetics predicts and what prediction licenses. Kathryn Paige Harden at Texas asks the question in its morally loaded form. Original Sin: On the Genetics of Vice, the Problem of Blame, and the Future of Forgiveness came out from Random House in March 2026. Scientists called it magnificent. The Los Angeles Review of Books treated it as scientific Calvinism and a rehash of eugenics.
Sixth, morality and collective behavior under technological and institutional pressure. Crockett at Princeton took a 2026 Troland Research Award from the National Academy of Sciences for a mechanistic theory of moral cognition, with the citation pointing to how technologies such as social media exploit moral learning and erode community trust. Betsy Levy Paluck offers the answer template: pick a process, intervene in a school or a post-conflict society, and measure behavior.
Seventh, whether the discipline describes humanity or affluent institutional life. Joseph Henrich still owns the question. The answer is arriving through international networks, which weakens any single department’s claim to the topic.
Eighth, whether competing theories can be made to risk defeat. The consciousness field ran an adversarial collaboration on integrated information theory and global workspace theory, preregistered the contrasting predictions, separated theory owners from data collection, and embarrassed both camps. The method won, not either theory. Other subfields are copying it.
On the asker and answerer question, I would put it more sharply than a division of labor. Asking is portable and cheap. Answering is expensive and local. Metascience gave a generation of psychologists a route to high status that required no laboratory, no grant, no participants, and no wait. That is the cheapest prestige available in the discipline, which explains its growth better than its intellectual merit does, and its intellectual merit is real. The corresponding risk is a field full of people who audit and few who build. Vazire, van Rooij, Borsboom, and Guest ask. Nosek, Paluck, Orben, Frank, Goodman, Craske, and Jessica Schleider build.
The questions have become consortium-shaped. Princeton has the fastest rise, holding Crockett, Tom Griffiths, Paluck, Shafir, and Cohen at the intersection of computation, morality, and field experimentation. Stanford has the broadest claim and absorbs new questions rather than chasing them. Amsterdam owns the foundational question about what psychological objects are. Cambridge’s MRC Cognition and Brain Sciences Unit owns adolescent digital mental health, and Orben directs the group there. Leiden, Radboud, and Northwestern are rising in narrower lanes. Two of my four leaders sit outside the United States, and I expect that share to grow.
Moving down. The losses attach to styles. Small samples. Counterintuitive single effects. Theory written after the results. Universal claims from American sophomores. Generic screen-time correlations. Candidate genes. Diagnostic categories treated as natural objects. Prediction models announced before prospective validation. The celebrity social psychologist who owns one memorable finding has lost the most.
The methodological closure is mostly deserved. It also has victims. Preregistration culture penalizes description, exploration, case study, and any population too small or too rare to power a trial. Some true things about unusual people are now hard to publish.
The editorial closure operates through harm standards. Nature Human Behaviour reserves the right to require changes or decline work when anticipated harm to groups outweighs benefit, and presents this as research ethics. Editors who hold that power can use it against accurate findings, and the policy does not have to be abused for authors to anticipate it and self-censor.
The professional closure runs through organizational position-taking. When a body like the American Psychological Association adopts a firm stance on a contested clinical question, youth gender medicine being the current example, it converts an empirical dispute into a loyalty test. Nobody has to be punished for the incentive to work.
The fiscal closure now outweighs the other three combined. The fiscal 2027 budget request released April 3, 2026 proposes dissolving the National Science Foundation’s Social, Behavioral and Economic Sciences directorate and zeroing its funding, with continuing grants aligned to administration priorities transferred elsewhere. That directorate funds roughly sixty-three percent of American academic research in the social and psychological sciences. NSF leadership told staff the same day that it would begin dissolving the directorate before the end of the fiscal year, telling rotators to prepare to return to their universities. The agency has awarded about six hundred thirteen grants in fiscal 2026, roughly a fifth of the 2021 through 2024 pace, and has no confirmed director. The entire National Science Board was fired on April 25, 2026. Congress restored much of what the White House tried to cut a year earlier, so the outcome remains open, and the directorate page was still live in early July. The internal pressures shift what gets published. This one shifts what gets done.
I would add a fifth force nobody has priced yet. If large models can produce plausible synthetic participants and plausible literature reviews, the cheap end of psychology loses its reason to collect human data at all. Van Rooij and Guest argue that on epistemic grounds. The near-term effect will be economic, and it will arrive fastest in exactly the underfunded departments that can no longer afford real samples.
Editors, prize committees, and infrastructure owners still control status flows. The trade publisher has become an independent route around all three. Harden and Haidt both took it. Both gained enormous public standing and paid an internal price. That route will get more crowded as the grants dry up.
