INTRODUCTION

A central question in the psychology of conspiracy beliefs is why some people endorse them while others do not. Many conspiracy claims—for example, ‘global warming is a hoax’, or ‘the harmfulness of vaccines is being covered up’—run counter to evidence, and widespread belief in unsubstantiated conspiracies can have serious negative consequences for individuals and society (see Jolley et al., 2022). A widely supported finding in the literature is that conspiracy beliefs are associated with deficits in cognitive sophistication, including critical thinking, cognitive flexibility, scientific reasoning and scientific literacy (Binnendyk & Pennycook, 2022; Fasce & Picó, 2019; Georgiou et al., 2021; Gjoneska, 2021; Hornsey et al., 2023; Lantian et al., 2021; Pytlik et al., 2020; Yelbuz et al., 2022). This supports the idea that lower cognitive sophistication leads people to endorse conspiracy beliefs, and carries the implicit assumption that higher cognitive sophistication leads people to reject them.

One major challenge, however, is that definitions of ‘conspiracy theory’ are often imprecise and inconsistent across studies. Neutral or non‐normative definitions characterize conspiracy theories as explanations where the ultimate causes of important social and political events and circumstances are attributed to two or more powerful actors operating in secret (e.g. Douglas et al., 2019; Keeley, 1999). Such definitions are agnostic about whether a conspiracy theory is warranted or unwarranted, and therefore includes claims backed by little evidence (e.g. the Apollo moon landings were faked) as well as conspiracies that genuinely happened. The Watergate scandal, MK Ultra program and Iran–Contra affair are some examples of verified conspiracies (i.e. where protagonists have confessed, and the conspiracy is now part of the official historical record). Tobacco companies and the fossil fuel industry have also intentionally suppressed research conflicting with their interests (McGreal, 2021). Conspiracy thinking may even be adaptive, helping people detect dangerous outgroup coalitions (van Prooijen & van Vugt, 2018) or make sense of real scandals and abuses of power (Butter & Knight, 2020). Indeed, belief in conspiracies is higher in more corrupt countries, where conspiracies seem more plausible (Alper, 2023; Hornsey & Pearson, 2022).

In public discourse, however, conspiracy theories are typically perceived and stigmatized as implausible or bizarre claims that run counter to evidence (Douglas et al., 2021; Imhoff & Bertlich, 2024; Nera et al., 2020, 2022; Uscinski & Enders, 2023). Many academic definitions also build in epistemic value judgements. For instance, conspiracy theories have been defined as implausible (Brotherton, 2013), epistemically unwarranted (Lobato et al., 2014), lacking sufficient proof (Wagner‐Egger et al., 2019, 2023) and epistemically risky or ‘inherently prone to being false’ (Douglas & Sutton, 2023 p. 284). Psychological research has focused almost exclusively on studying belief in conspiracy theories as normatively defined (Imhoff & Bertlich, 2024), but this has several practical and epistemic problems (see Imhoff & Lamberty, 2020; Nera & Schöpfer, 2023). For one, if conspiracy theories are defined as unwarranted, endorsing them becomes functionally equivalent to endorsing claims that lack evidence. However, conspiracy theories defined in this way make up only a subset of all conspiracy claims, and a narrow focus risks conflating unwarranted conspiracy belief with conspiracy belief more broadly (e.g. Binnendyk & Pennycook, 2022; Brashier, 2023; Lantian et al., 2021).

Because prior work focuses almost exclusively on measuring belief in unwarranted conspiracy claims, the negative association between cognitive sophistication and conspiracy belief becomes somewhat circular, revealing only that less cognitively sophisticated people are more likely to believe implausible claims (Imhoff & Bertlich, 2024; Stanovich & Toplak, 2025). The implicit assumption is that rejecting conspiracies must therefore signal cognitive sophistication. If only belief in unwarranted conspiracy theories is assessed, however, it is unclear whether people disbelieve conspiracies because of the evidential basis of the claims or a heuristic tendency towards outright rejecting cospiracies. If cognitive sophistication plays a role, it should be reflected not in blanket rejection but in the capacity to discriminate well‐supported claims from those that lack evidence (see Stanovich & Toplak, 2025).

In a recent study, Frenken et al. (2024) presented participants with plausible and implausible conspiracy theories about a fictional country. Less reflective people were more likely to endorse implausible theories but endorsed plausible ones at rates comparable to their more reflective counterparts. In a belief‐updating task, those high in conspiracy mentality also continued to find conspiracy theories plausible despite strong contradictory evidence. Similarly, Hattersley et al. (2022) show that flawed reasoning is related to belief in implausible conspiracy theories but unrelated to plausible ones.

Examining plausible and implausible claims separately, however, provides an incomplete picture. This is because belief in plausible and implausible claims is often correlated (Bensley et al., 2020; Frenken et al., 2024; Moulding et al., 2016; Pennycook et al., 2025). Moreover, those higher in conspiracy mentality are more likely to endorse both (Frenken et al., 2024), reflecting general response tendencies rather than careful evaluation of each claim. Recent work has nevertheless begun to disentangle response tendencies from discernment. Tagand et al. (2026), for example, demonstrate that those low in conspiracy mentality more accurately distinguish warranted from unwarranted conspiracy claims, but also tend to reject epistemically warranted claims (see also Stanovich & Toplak, 2025).

In the present studies, we extend this work by examining how well people distinguish between verified and unverified conspiracies, and what this reveals about the relationship between conspiracy beliefs and cognitive sophistication. We define verified conspiracies as conspiracy claims supported by documented evidence and now form part of the official public record, whereas unverified conspiracies are conspiracy claims that lack compelling, supporting evidence. We define cognitive sophistication broadly as the capacity for complex thought, encompassing a constellation of cognitive abilities and thinking dispositions (see Toplak et al., 2014). Here we operationalize cognitive sophistication as general knowledge (breadth of factual knowledge about the world), cognitive reflection (the tendency to override incorrect intuitive responses through deliberation) and scientific reasoning (the ability to think critically about scientific evidence).

Our aim is to examine the relationship between cognitive sophistication and conspiracy belief in a way that disentangles the ‘accuracy’ (discrimination) of the beliefs from response bias (the tendency to endorse or reject conspiracy claims). We do this by examining the correlates and predictors of conspiracy discrimination and whether discrimination is a stronger predictor of cognitive sophistication than response bias. We then build on this in Studies 2 and 3 by assessing whether indiscriminate rejection of conspiracy claims, like indiscriminate belief, is associated with lower cognitive sophistication.

Transparency and openness

Study 1 is an exploratory re‐analysis of unpublished data. Its analyses were not preregistered. Methods, hypotheses and planned analyses for Study 2 were preregistered prior to data collection (see https://osf.io/j9y4e/wiki?wiki=w3bgq) but the primary focus of Study 2 deviated from these original plans (details below). Materials, data, and analytic scripts for Studies 1 and 2 are available at: https://osf.io/pnavw/. Methods, hypotheses and planned analyses for Study 3 were preregistered prior to data collection (see https://osf.io/hdex4/wiki?wiki=6b82x), and all materials, data and analytic scripts can be found at: https://osf.io/xsjr2.

For each study, we report how we determined sample sizes, data exclusions and materials used. Data were analysed and visualized using R and R Studio. Each study received ethical approval from the UNSW Human Research Advisory Panel C. All participants provided informed consent. The authors declare no conflicts of interest.