Other recent blockbuster ideas in psychology are also steeped in this ideology. Take mind-set interventions, which are designed to shift people’s mind-sets from “fixed” (“I failed the test — I’m just stupid”) to “growth” (“I’ll do better next time if I work harder”). “For 30 years, my research has shown that the view you adopt for yourself profoundly affects the way you lead your life,” said Carol Dweck, the originator of that idea and a professor at Stanford, in her book “Mindset: The New Psychology of Success” (ideas she echoed in a TED Talk that has been viewed more than 12.5 million times).
Perhaps better known is the implicit association test, which claims to be able to measure one’s level of implicit (or unconscious) bias via a brief computer test that involves comparing reaction time to different stimuli. Since the creators of the test also claimed that implicit bias can go a long way toward explaining persistent racially discriminatory gaps in the United States, the computer test is now one of the most common features of contemporary diversity trainings.
Since these claims were first made, though, a full-blown replication crisis has hit psychology, meaning that when researchers attempt to redo previous studies, they often find either a much less impressive result or none whatsoever. It turns out that the standard statistical methods long employed by psychologists (and other scientists) can easily produce false positive results. About 50 percent of published results from experimental psychology have failed to replicate, and the subfield of social psychology — the home base of most social priming, implicit bias and stereotype-threat research — tends to fare even worse.
Studies purporting to offer simple remedies to serious problems have been hit particularly hard. Mind-set interventions appear to be nowhere near as powerful as Professor Dweck initially advertised: A major, well-constructed 2019 study in Nature found some effect, but only a relatively small one and only for weaker students. (In a phone conversation, she pointed out that the Nature study was centered on a fairly minimalist mind-set intervention designed to be easily scalable and referenced larger effects found in earlier studies premised on more costly, time-consuming multisession interventions involving highly trained personnel.)
As for that fascinating social-priming magic embraced by Dr. Bargh, like people walking slower after seeing words with geriatric associations? “I don’t know a replicable finding,” said Brian Nosek, a psychologist and leading replication advocate, in 2019. “It’s not that there isn’t one, but I can’t name it.” The few social-priming effects that have survived this scrutiny tend to be small, inconsistent and not necessarily relevant outside of lab settings.
(In a series of emails, Dr. Bargh argued, as he has elsewhere, that his field’s replication tribulations have been overstated and pointed to some of the positive results.)
The implicit association test has experienced similar travails. It is still often a part of everyday diversity-training settings, but its creators long ago acknowledged that it is too noisy a test to be used to identify people as likely to engage in racist acts (which constitutes significant backtracking from their original claim).