We also find evidence that the illusory truth effect is stronger for implausible headlines. Rather than being a bias that prevents the processing of other info, repetition appears to form a prior (towards truth) that is integrated with what people know about the item. 11/

Second, intuitive participants (measured by CRT) show a much stronger illusory truth effect in the high-quality condition. However, they show almost minimal effects of rep in low-quality. Deliberative participants show small illusory truth in both. 10/

Is the effect of repetition moderated by source credibility? YES!

The effect of repetition is about ¼ the size in the low-quality condition. This moderation occurs for true and false items. 6/

To test this prediction, ppl see 3 True & 3 False headlines repeated in 3 exposure phases + 1 judgment phase. Ppl randomized to a high-quality condition see many novel headlines that are largely true. Those in low-quality see mostly false. Feed quality -> source credibility. 5/

The model unifies 4 findings in the lit: a) baseline illusory truth, b) each additional repetition has a smaller effect on beliefs, c) repetition effects are larger for implausible items, d) novel items are believed less when repetition is common. More explanation in fig 4/

Despite what it feels like, ppl consume mostly true info from credible sources (fig from @jennyallen.bsky.social ). Sources are usually good but sometimes err – friends lie, credible news sources retract.

In a formal model, we investigate what this implies for repeated info. 2/

New WP!
The illusory truth effect (repetition -> belief) is core to psych of beliefs, & thought to be a deep bias impacting misinfo, persuasion & advertising

Why would cognition include such a flaw? We argue it is a rational adaptation to high-quality info environments 🧵1/

These patterns of belief updating are consistent with a model of hierarchical Bayesian inference. When presented with consensus, people update their beliefs about climate change and scientists in a principled manner.

We test 2 interventions in conjunction with consensus information. One of these interventions, which focused on the long-established history of climate science, proved particularly effective at boosting climate belief above and beyond scientific consensus.

However, we find that people attribute consensus to the skill or bias of scientists in similar ways.

All parties think that scientists are more skilled as consensus increases. But, perceived bias also increases.

These attributions are predictive of the effect on climate belief.

We vary the level of scientific consensus in a large, nationally-representative sample to explore how individuals attribute consensus to climate change and the attributes of scientists.

First, we find that higher consensus increases belief in climate change across parties!

Beliefs about climate scientists can meaningfully change the impact of consensus messaging. At the same time, consensus, in turn, shapes beliefs about scientists. This is suggestive of the occurrence of hierarchical inference, as predicted by influential theory.

One of the most common climate communications is the consensus message, like this great example from Cook and Lewandowsky (2016), where people are told that 97% of climate scientists agree that human-caused climate change is happening.

Very happy to share a new working paper with some great coauthors!

97% of climate scientists agree that human-caused climate change is occurring. But how do people interpret this consensus? Is it a testament to expertise or a signal of bias?

psyarxiv.com/ezua5/