Response Bias

What it is

Any yes-or-no judgement mixes two things. Discrimination is how well you separate signal from noise: the gap between how you treat the real thing and the fake. Bias is where you set your threshold: how much evidence you need before you say yes at all.

Lower the threshold for scepticism and you will rate fake headlines worse. You will also rate true headlines worse. The average rating of the fake items falls either way, so a study that reports only that cannot say which happened.

Signal detection theory exists to separate the two, and receiver operating characteristic analysis is the standard tool for doing it.

In effect

Ariana Modirrousta-Galian and Philip Higham applied it to the best-known misinformation interventions, the browser games Bad News and Go Viral, in the Journal of Experimental Psychology: General in 2023.

The earlier evaluations had participants rate the reliability or manipulativeness of true and fake news items before and after playing, usually against a control group that played Tetris or did nothing, and compared mean ratings. Reanalysed with the two terms separated, the games did not improve discrimination. What moved was the bias.

Tina Seabrooke, with the same authors, then re-examined the Bad News game using Indian headlines in Psychonomic Bulletin & Review in 2025 and reported no evidence of improved discrimination.

What it does not say

It does not say the games do nothing. Raising scepticism is a real effect and may be wanted in some settings. It says the effect is not the one claimed.

It does not refute inoculation theory, which is sixty years old and far larger than two browser games.

And it is one research group. A reanalysis and a follow-up by overlapping authors is not independent confirmation, which is the same complaint the authors are making about the positive literature.

The transferable use is a question rather than a verdict. Any training that claims to make you better at spotting something bad can succeed by making you warier of everything, so ask whether the evaluation measured the gap or only the average.

---