A new study suggests that when scientific journals publicly back political candidates, public trust in science falls, especially among moderate and conservative Americans. The authors report declines in views of science’s impartiality and societal value, though people’s willingness to follow practical scientific advice appears largely unchanged. The work is published in Group Processes & Intergroup Relations.
Researchers framed the question through the lens of motivated reasoning—the tendency to process information in ways that reinforce existing beliefs and political identities. They theorised that endorsements from esteemed journals could trigger this mechanism, prompting some readers to question science’s objectivity if endorsements clash with their political views.
The research measured trust as four distinct dimensions: perceived ability (scientists’ competence), integrity (impartiality), benevolence (public-serving motive), and behavioural trust (whether people would heed recommendations).
Experiment one, conducted shortly before the 2020 presidential election, recruited 3,081 participants online who were randomly assigned to read either a neutral journal background piece or an excerpt from a real editorial endorsing Joe Biden or criticising Donald Trump.
Following the reading, participants answered questions assessing trust in science and intentions to comply with health guidelines. Endorsements led to a measurable fall in general trust and a drop in perceived integrity, with the impact disproportionately affecting moderates and conservatives.
Among conservatives, the treatment also increased belief in conspiracy theories about science and the origins of the coronavirus, indicating a reduced sense of benevolence. The researchers found no statistically significant effect on behavioural trust.
To test robustness, a second experiment conducted days before the November 2020 election involved 1,056 participants matched to national demographic quotas. The design mirrored the first study, using neutral background text versus political endorsements.
Its results closely replicated the initial findings: exposure to political messaging reduced general trust in science and heightened belief that science is biased, again driven by moderates and conservatives. This time, participants were less likely to rely on scientific sources for coronavirus updates, suggesting a perceived drop in scientific ability or competence.
A final experiment, four years later around the 2024 election, involved 2,144 participants reading material from Scientific American, Nature or The Economist. The Economist was included for its broad coverage. Endorsements endorsed Kamala Harris or criticised Donald Trump. The coronavirus material was removed and new measures assessed overall objectivity and societal usefulness of science.
Results in 2024 aligned with the previous cycles: political endorsements reduced general trust and increased the view that personal and political values drive scientific research. Participants also reported lower confidence that science is objective and useful for society, with the greatest declines among moderates and conservatives.
Limitations of the work include its use of real editorials, meaning it did not test reactions among liberal readers to endorsements of conservative candidates. The researchers note that whether liberals would react in the same way remains unknown.
Authors emphasise that while explicit endorsements carry reputational costs, silence from science on policy matters could also be damaging. They suggest future research to explore how scientific organisations can counter misinformation and advocate for evidence-based policies without appearing overtly partisan, to safeguard legitimacy across the political spectrum.
The study, titled Presidential candidate endorsements by scientific journals decrease trust in science especially for moderate and conservative Americans, is by Stylianos Syropoulos, Kyle Fiore Law, Crystal Li, Bernhard Leidner and Kevin L. Young.
