Andrew Follett writes at National Review Online about a key problem plaguing science.

Americans don’t trust science the way they used to — because science is becoming increasingly untrustworthy.

In the 1970s, when Gallup asked Americans how much confidence they had in science, political-party members varied little in their responses: Republicans’ attitudes were about the same as Democrats’. But Gallup’s more recent polling shows that trust in science has deteriorated among Americans in recent decades, especially among Republicans, whose confidence was nearly 30 points lower in 2021 than in 1975. Independents have lost considerable faith in science as well, as have Americans as a whole. The only group whose trust in science has increased is Democrats.

Until around 2018, polling by FiveThirtyEight had found no significant partisan divide among Americans who said they had “a great deal” of confidence in the scientific community. But the trust gap grew more pronounced by 2021, when 65 percent of Democrats in this group had “a great deal” of confidence in scientists, compared with 32 percent of Republicans.

Democrats like to claim they’re the “party of science,” with science presumed to be fact-based and impartial, but this partisan trust gap is evidence that science has become politicized. The politicization is a result of two phenomena: a broad replication crisis in the social sciences, and its intermixing with agenda-pushing by militant ideologues in all the sciences. Republicans and independents are right to be skeptical when scientists make broad claims for their research but can’t get consistent results, or when scientific findings are twisted or misrepresented to support the researchers’ ideological beliefs.

By the current norm of social sciences, to be considered significant, a finding must have a 95 percent chance of not being random. That might sound robust, but it isn’t, as there’s enormous pressure to play fast and loose with the scientific method to get the positive results that academics want.