People actively consume and elaborate on new information. In news and social media, people may ask questions or provide evidence leading to false inferences. Although the false information is not directly stated, people may nonetheless construct and adopt false beliefs. We present two experiments in which people generated information in response to implications. In our first experiment, participants were asked to read and generate both true and false trivia answers. They then made truth judgments for items that were new to them, previously read, and previously generated. Participants rated their self-generated answers and answers they read as more true than the new statements. In our second experiment, we presented two related statements that could lead to a false pragmatic inference. When presented with the inference, people rated that statement as more true than unrelated statements. People come to believe implied false information even when it is not directly stated.

Sleep health is increasingly recognized as a critical aspect of overall wellness, as it is believed to influence numerous domains of functioning, including physical and emotional health and cognitive performance. Determinants of sleep health include a broad range of physiological, socio-cultural, and psycho-behavioral factors; here, we focus on sleep-related beliefs. In Study 1, we found associations between false beliefs (myths or misconceptions) about sleep and sleep-related behaviors. Those with stronger false beliefs have less healthy sleep profiles in terms of sleep duration, timing, and hygiene. In Study 2, we investigated the extent to which educational infographics may be effective in reducing false beliefs about sleep. We opted to use infographics because they are easily accessible and have been shown to be effective in promoting other health behaviors (e.g., nutrition). We found that one-shot exposure to sleep infographics reduced endorsement of sleep myths and increased plans to improve sleep habits.

Research highlights consistent lexical differences between conspiracy-related and non-conspiracy texts, even when matched by topic. We explored the origins of conspiratorial language using a semantic memory model, where interconnected concepts in long-term memory activate one another. Convergence toward conspiratorial language may stem from a cognitive bias where conspiracy-related concepts (e.g., deception, illuminati) are persistently activated, regardless of context. To test this, we conducted two preregistered studies (N ≈ 350 each) using a Word Association Task with cues that were either conspiracy-related (e.g., elite, evil) or unrelated (e.g., tango, oven). Word embeddings, trained on conspiracy and mainstream corpora, assessed response similarity: responses were classified as more conspiratorial if their cosine similarity was higher in the conspiracy embedding than in the mainstream embedding. We found that participants’ conspiracy mentality predicted greater convergence toward conspiratorial language, but only for conspiracy-related cues, indicating the role of contextual cues in activating conspiratorial semantic networks.

Repetition increases belief in statements. Typical studies of this illusory truth effect focus on verbatim repetition of claims without a specific source. However, claims in real life are often repeated with paraphrased wordings from various sources. Prominent cognitive accounts suggest this variation should not matter: repeated exposure to an idea should increase belief in it by making it feel more familiar, regardless of variation in its wordings or source. However, theories of social influence suggest that repetition should increase belief most when it reflects a consensus among many others. In three experiments (N = 718 US-based MTurk/Connect workers), we showed participants statements three times in a manner that reflects a social consensus (different wording and/or sources) or not (same wording and/or sources). In line with current, cognitive accounts of the illusory truth effect, these repeated exposures increased belief in statements to a comparable degree despite variations in wording or source.