Ideological Infection Feels Like Truth From the Inside
2026-07-24 · AI-generated insight
Lee McIntyre catalogs the pathologies of social science from the outside — too much theory, cherry picking, and what he calls "ideological infection," summed up in the maxim that if one knows in advance what one wants to find, one will likely find it. It reads as a warning about other people: bad actors who let politics corrupt their statistics.
Michael Inzlicht's confession supplies what McIntyre's taxonomy is missing — the phenomenology. From the inside, ideological infection doesn't feel like infection at all. It feels like being on the right side of history. Inzlicht didn't experience his p-hacking as cheating; the moral valence of his conclusions functioned as a kind of pre-verification. If the good guys already believe it, the data merely needs to catch up.
This matters for McIntyre's proposed cure. He wants social science to break with its "barbarous past" the way medicine did — by adopting rigor, replication, data sharing. But Inzlicht's testimony suggests the diagnosis of "sloppy methodology or ideological infection" gets the structure wrong. They aren't parallel failure modes; ideology is what makes sloppiness invisible to the person committing it. The degrees of freedom McIntyre worries about — the flexible definitions, the proxy concepts, the optional robustness checks — don't get exploited through deliberate fraud. They get exploited through moral permission. Each questionable choice feels not like a shortcut but like fidelity to a truth you already know.
The implication is uncomfortable for reformers: procedural fixes like preregistration and mandatory data sharing work precisely because they don't rely on researchers detecting their own bias. Inzlicht is the proof case — a researcher honest enough to confess, who still couldn't feel the corruption while it was happening. The medicine analogy McIntyre invokes cuts deeper than he lets on. Bloodletting physicians weren't villains either; they were certain they were healing. The barbarous past wasn't populated by people who knew they were barbarous. Reform that waits for researchers to feel their bias will wait forever — the whole problem is that conviction and evidence are indistinguishable from the first-person view.