Genetic laboratory technician doing embryo biopsy with long thin needle, conducting compulsory medical research. (Credit: © Viacheslav Yakobchuk - stock.adobe.com)
The Embryo Selection Study That Undercuts the ‘We’d Never Do That’ Argument
In A Nutshell
- A new study of more than 2,000 people in the U.S. and China finds that non-medical traits like low IQ and antisocial behavior shape embryo selection almost as strongly as a disease like heart disease, even though people say testing for them feels less acceptable.
- Roughly 61% of Americans said they’d be willing to test embryos for non-medical traits, far more than earlier studies found, and a majority did not call the testing morally wrong or worth banning.
- When forced to choose between two hypothetical embryos, people avoided one with a higher chance of vision impairment far less often than one with a higher chance of heart disease, low IQ, or antisocial behavior.
- A leading theory that people avoid traits they see as harder to control did not hold up. What predicted avoidance instead was how serious and life-disrupting a trait seemed.
There’s a well-documented gap in human psychology between what people say they’ll do and what they actually do. A new study published in Nature Human Behaviour finds that gap showing up in one of the most consequential decisions a person can face: choosing which embryo to implant during IVF.
Most people, when asked directly, say testing embryos for traits like low intelligence or antisocial behavior feels different from testing for a disease, and less acceptable. That kind of testing was originally developed to catch severe inherited diseases, and past research has generally found the public draws a line at using it for anything else. But when researchers gave more than 2,000 participants in the United States and China hypothetical embryo profiles and asked them to choose between two options, those same traits drove decisions just as powerfully as the risk of heart disease.
Put simply: as a group, people were more cautious about testing for non-medical traits. But a separate group handed that same information in a forced embryo choice let it shape their pick just as strongly.
Genetic Testing Was Built for Disease, Now It Screens for Much More
Preimplantation genetic testing is a procedure used during in vitro fertilization in which embryos are genetically screened before being placed in the womb. Originally designed to detect serious inherited conditions, it is now offered by private clinics to assess a range of traits, some with no connection to disease. Earlier research had consistently found that the public draws a firm line: testing for medical conditions is acceptable, testing for non-medical traits is not.
This study, which included more than 1,400 American adults and more than 600 participants from China, tested whether that stated preference holds up in a forced-choice setting. Researchers split American participants into two groups: one was asked whether they would be willing to test embryos for each of four traits (heart disease, vision impairment, low IQ, and antisocial behavior), while the other skipped the opinion questions and instead made forced decisions, choosing embryo A or embryo B based on the genetic profiles shown to them. The Chinese sample completed only the forced-choice task, letting researchers compare what people say in the abstract against what a separate group does once the information is already in hand.
Given a Real Choice, Non-Medical Traits Mattered Just as Much
Among those asked about willingness to test, two out of three approved screening for medical conditions like heart disease. For traits with nothing to do with disease, like low IQ or antisocial behavior, that number dropped, but not by much: three in five still said yes.
What surprised the team was how many American participants expressed willingness to test for non-medical traits at all. Prior studies had found that only a minority, sometimes as few as 20 to 37 percent, approved of testing for traits like intelligence or personality. The higher numbers here may reflect shifting attitudes as genetic testing becomes more familiar, or how the traits were framed: earlier research often asked about positive traits like high intelligence, while this study asked about negative ones, like low IQ, which may register more like a health concern. The paper raises this framing possibility itself, so it remains tentative.
The most telling finding came from the forced-choice task. Choosing between two hypothetical embryos, participants steered away from the one with above-average odds of low IQ or antisocial behavior at least as often as they steered away from the one with above-average odds of heart disease, sometimes more so. China told the same story, and at times avoided low IQ even more than Americans did.
Vision impairment stood apart. Participants were far more willing to pick an embryo with above-average chances of vision problems than one flagged for any of the other three traits, which formed a cluster they treated as roughly comparable in seriousness.
Severity, Not Controllability, Drove Which Traits People Avoided
Researchers had initially suspected that a concept called controllability might explain these patterns: people might avoid traits that feel impossible to manage, and feel less urgency about ones that can be addressed. Vision impairment can be corrected with glasses. Heart disease is harder to control.
That theory fell apart. When participants rated how controllable each trait felt, the ratings didn’t line up with their choices, and sometimes pointed the opposite way. What actually seemed to drive people’s picks was how serious and life-altering a trait felt, not whether it could be managed. Researchers are careful to call this a hunch rather than a proven mechanism, since it wasn’t part of their original predictions.
That gap between what people say and what a forced decision draws out of them shows up on both sides of the Pacific, even though no one in this study was making a real reproductive choice. And it raises a practical wrinkle for how genetic testing gets regulated. The U.S., the U.K., China, and much of Europe draw a hard line between medical and non-medical testing on paper. This study suggests real decisions don’t respect that line nearly as cleanly.
Disclaimer: This article summarizes findings from a peer-reviewed study and is intended for general informational purposes. It does not constitute medical, reproductive, or genetic counseling advice. Anyone considering IVF, preimplantation genetic testing, or embryo selection should consult a qualified medical or genetic counseling professional.
Paper Notes
Limitations
Several important limitations apply. Only four traits were tested, chosen to span a range of perceived impact rather than represent every possible trait, so results should not be generalized to all traits. The scenarios used standardized, simplified risk differences that do not reflect the real-world variability of predictive accuracy across different conditions. All choices were hypothetical: participants were not making actual medical decisions, and real-world choices during IVF may be shaped by emotional, relational, and institutional pressures a survey cannot capture. Including both American and Chinese participants adds cross-national breadth but does not represent the full range of global regulatory environments. The exploratory finding pointing to perceived seriousness as a driver of avoidance was not pre-registered and was not directly tested through experimental manipulation. Because participants in the forced-choice task were given genetic information regardless of whether they would have sought it, they may have treated it as relevant to the task even if they would not endorse its use in an actual clinical setting. The results show that available information can shape a choice; they do not establish that someone who says they oppose testing would necessarily act on that information if a clinic offered it anyway.
Funding and Disclosures
Funding came in part from the Wellcome Trust, the National University of Singapore, the Social Science Research Council of Singapore, and the National Institute for Health and Care Research Oxford Health Biomedical Research Centre, along with several smaller grants supporting individual co-authors. Study author Julian Savulescu discloses a role as a Bioethics Committee consultant for Bayer. The remaining co-authors report no competing interests.
Publication Details
Authors: Edmond Awad, Clara Colombatto, Joanna Demaree-Cotton, Brian D. Earp, Jim A. C. Everett, Peng Liu, G. Owen Schaefer, Ilina Singh, Dominic Wilkinson, and Julian Savulescu | Affiliations include: Uehiro Oxford Institute and Department of Psychiatry at the University of Oxford; Department of Economics at the University of Exeter; School of Psychology at the University of Waterloo; Centre for Biomedical Ethics at the National University of Singapore; School of Psychology at the University of Kent; Center for Psychological Sciences at Zhejiang University; John Radcliffe Hospital; and Murdoch Children’s Research Institute. | Journal: Nature Human Behaviour | Article type: Registered Report | DOI: https://doi.org/10.1038/s41562-026-02562-w | Stage 1 protocol accepted in principle: October 7, 2025 | Protocol available at: https://osf.io/vb9c2
