On the Usefulness of the Conspiracy Mentality Concept

Conspiracy Mentality and Belief in Specific Conspiracy Theories – A Chicken and Egg Problem?

Nera now criticizes this view as intuitively suggesting a unidirectional causal relation from conspiracy mentality as a general propensity to the endorsement of specific conspiracy theories. I beg to differ here – claiming a relatively stable propensity does not imply a unidirectional causal claim. Indeed, we can think of the relation between conspiracy mentality and specific conspiracy belief in both an inductive and a deductive way. The inductive perspective would imply that some people find very specific conspiracy theories highly plausible (for whatever reason) and use these experiences of plausible conspiracy theories to induce a general rule that many (or most) events or phenomena can be explained by plots hatched in secret. Such a perspective allows for the interesting research question of the characteristics and situational context of such gateway conspiracy theories that initiate this process. From a deductive perspective, people have a certain belief about how the world operates in principle (e.g., most things are determined by secret plans of a few powerful agents) and use this belief as a schematic foil against which world events are interpreted. As we know from a plethora of research, humans have a strong tendency and a rich arsenal of instruments to engage in belief-consistent information processing (Oeberst & Imhoff, 2023). This perspective is well in line with the findings that conspiracy mentality is associated with the adoption of extremely novel conspiracy theories (emerging quickly after almost any event of at least intermediate relevance) but also completely fictitious ones (Imhoff & Lamberty, 2017; Meuer et al., 2021). Importantly, however, these two perspectives are not mutually exclusive, and both models of causation can be true: conspiracy mentality as an over-generalization of specific beliefs and as an interpretative foil to make sense of (new) aspects of the world. Equally importantly, one of Nera’s examples from genetics seems to imply that causation is restricted to deterministic causation, as he refutes the idea that a genetic predisposition that increases the risk for a condition given the presence of other circumstances has a causal impact. This is an overly narrow understanding of causation. Increasing the likelihood of developing a disease might not necessarily (deterministically) lead to developing this disease, but it is still a causal impact in the probabilistic sense, making it more likely to occur.

The Ontology of Conspiracy Mentality

Despite the possibility of such a peaceful coexistence of the potential causal models (including additional ones where third variables like hyperactive agency detection – partially – cause both), there seems to be some uneasiness regarding the question of how to think about conspiracy mentality. Do we follow a reflective model whereby the endorsement of each specific conspiracy belief merely reflects the latent variable “conspiracy mentality” as frequently assumed for personality traits? Or do we construe the relation in the sense of a formative model whereby conspiracy mentality is the tendency to show this behavior (without any latent disposition)? This question also refers to causality, but not the one discussed above, where A (adopting a specific conspiracy theory) leads to B (harboring a generalized conspiracy worldview) over time (or vice versa). Instead, it asks whether variations on the endorsement of specific conspiracy theories are caused by an entity that exists independent of this measurement (the realist position behind the reflective model) or whether conspiracy mentality is just the summary of our measurement that does not exist in the real world (the constructivist position better alignable with a formative model; Borsboom et al., 2003).

Here, I will defend the former position that – at least at the hypothetical level – people differ on their conspiracy mentality. The alternative position would render any generalization beyond a specific scale used problematic: If conspiracy mentality is just a summative characterization of the way participants responded to specific items, any different scale (or different subset of items) will form different conspiracy mentalities. This does not seem to be the way the field treats belief in conspiracy theories. Instead, most researchers seem to employ scales tapping into the endorsement of several specific conspiracy theories (that may change from study to study) or generic statements assumed to tap into the worldview behind it interchangeably (for a critique of treating the two approaches as identical, see Imhoff et al., 2022).

Generic Statements of Specific Theories?

Starting from the hypotheses that people do differ in the extent to which they endorse the wide-spread existence of conspiracies, the question is how to best tap into these individual differences. Several authors have proposed to ask for agreement with several specific conspiracy theories and take the average agreement as the best indicator of the latent variable. As I will argue below, this approach incorporates the issue we have labeled contamination elsewhere (Imhoff et al., 2022). Psychometrically speaking, it introduces systematic construct-irrelevant variance if the chosen items are not perfectly balanced in terms of their truth value and their ideological leaning.

One aspect that deserves particular attention is the contamination of specific conspiracy belief with (low) truth value. Items tapping into conspiracy mentality, on the other hand, are not prima facie right or wrong but reflect different worldviews. While some authors have argued that it is a downside of the conspiracy mentality concept that it does not entail the “tendency to subscribe to normatively weak beliefs” (Sutton & Douglas, 2020; p. 121), I would argue that this is a strength. Only a subset of specific conspiracy theories are normatively implausible beliefs. Although arguably these are the ones most frequently studied and most resonant of lay understanding of the term, there are a number of conspiracy theories that have evidence in their favor: that the Nazis conspired to annihilate European Jewry, that Mohammed Atta and other Al-Qaeda terrorists plotted in secret to fly passenger planes into the World Trade Center, or that a group of members of the Nixon administration secretly wiretapped and broke into the headquarters of the Democratic Party (see Imhoff & Lamberty, 2020b, for a thorough discussion). What these conspiracy theories have in common is that they are well accepted and there is good evidence for their accuracy. Nevertheless, believing in them suggests believing in secret coordination and cover-up. We would thus expect them to be associated with conspiracy mentality as well. On the flipside, the frequently reported association of conspiracy belief with low cognitive skills (Stasielowicz, 2022) or more intuitive and less analytic processing (Binnendyk & Pennycook, 2022) might just be an artifact of choosing implausible conspiracy theories as items.

As argued elsewhere (Imhoff et al., 2022), the same is true for ideological content contamination. Specific conspiracy theories are often tainted with such content contamination (e.g., expressing certain worldviews other than conspiracy mentality like rightwing, xenophobic, anti-technology, or anti-capitalist stands), and thus their correlation will be confounded by sources of variance other than conspiracy mentality. To a certain extent, this problem also arises when we tap into the general propensity to endorse conspiracy beliefs by averaging the agreement to a number of specific conspiracy theories (as done in other scales of conspiracist ideation; Brotherton et al., 2013). As long as the specific conspiracy theories are perfectly balanced with regard to sources of content contamination and truth value, these influences will get lost in aggregation. As soon as they are not, however, the composite score will have similar issues as single items tapping into specific beliefs. Thus, if we are interested in the latent variable behind endorsing conspiracy theories – and not the latent variable behind endorsing epistemically questionable or ideologically laden statements – measures of conspiracy mentality will prove more useful than the frequently employed (often mostly implausible) specific conspiracy theories.

Is Conspiracy Mentality a Useful Concept?

Ultimately, this is the core criterion of whether a concept is valuable: whether it is useful. Claiming the usefulness of the concept is not the same as making a claim about its ontology. There is an argument to be made that all psychological constructs are instrumental fictions by definition – be it intelligence, neuroticism, attention, or conspiracy mentality – which renders a debate over whether conspiracy mentality exists unfruitful (on the fiction of illusory essences, see Brick et al., 2022). More relevant than the issue of ontological existence is whether the concept provides a parsimonious way to describe regularity in human behavior. It is true that this systematicity has to go beyond empirically demonstrating an association of conspiracy mentality with specific conspiracy beliefs over and over again – and defining conspiracy mentality by this association (the Black Box argument decried by Nera). This indeed borders circularity or tautology.

While I agree with the notion that such a Black Box definition is not useful, I disagree that this is the state of current understanding of what conspiracy mentality is. We can – and should – characterize what individual differences in conspiracy mentality entail in other ways than making the connection to the endorsement of specific conspiracy theories and quite some research has done exactly that. Here, I would suggest to further differentiate what might be seen as constituents of a conspiracy mentality from its associates. Constituents (e.g., anti-elitism, distrust of authorities) seem to be an integral part of conspiracy mentality, as they are proximally implied by the very wording or items tapping into conspiracy mentality (e.g., “Those at the top do whatever they want”). Associates are less trivially related and provide the opportunity for open empirical questions about the relation to conspiracy mentality. Such more distal associates have been established by showing reliable correlations of conspiracy mentality with low general trust (Thielmann & Hilbig, 2023), also behaviorally (Meuer & Imhoff, 2021), a higher perceptual threshold to detect trustworthiness (Frenken & Imhoff, 2023), a bias to suspect negative intention and secret coordination, even behind mundane everyday events (Frenken & Imhoff, 2022), low epistemic trust in high-power sources (Imhoff et al., 2018), or feelings of isolation when refraining from internet use (Jetten et al., 2023). The finding that conspiracy mentality is more pronounced in cultural contexts that are plagued by corruption (Alper & Imhoff, 2023) may be an indicator that part of this construct taps into the sensitivity for valid cues to untransparent and illegitimate actions.

Future Directions

It cannot be stressed enough, however, that more research is needed and that elucidating the antecedents of so remarkably stable individual differences (e.g., Imhoff & Bruder, 2014; H. Wang & Van Prooijen, 2023) in conspiracy mentality will prove a worthwhile endeavor. Why people end up in a radicalized worldview characterized by conspiracy beliefs and ready to use violent means to pursue one’s goals (Imhoff et al., 2021) is a timely and socially relevant question. From a basic research perspective, however, there is no reason to restrict the inquiry to such rabbit hole dynamics. People fall on all sorts of places on a continuum ranging from absolute disagreement to absolute agreement with conspiracy mentality items, and while it is understandable and tempting to zoom in on why some people end on one of the ends of that spectrum, that will not be the whole story. A better understanding of any antecedent of virtually any position or dynamic within that spectrum will provide us with a better and more complete understanding of the phenomenon. In light of the high intermediate stability of conspiracy mentality scales (H. Wang & Van Prooijen, 2023), it seems worthwhile to direct our attention to sources of differences in earlier biographical phases, such as adolescence (Bertlich et al., 2023; Jolley et al., 2021).

Although these questions are still to be answered, it seems fair to attest that psychological research has begun to explore the nature of conspiracy mentality beyond circular trivialities. A second point of disagreement is the question on whether we should treat conspiracy mentality as a unidimensional or multifaceted construct. Nera (this issue) emphasized the conceptual heterogeneity of the items comprising conspiracy mentality scales and suggests that it might be better to construe it as multidimensional. In my perspective, the heterogeneity of the items is a feature, not a bug. Psychological constructs should span a certain breadth to incorporate different aspects. The fact that these heterogeneous items rank respondents in a similar order (e.g., in Latent Profile Analyses; Frenken & Imhoff, 2021) is much more telling than if less diverse items clustered together. To be clear, whether a concept is unidimensional or multidimensional should not be relegated to psychometric analyses but requires a conceptual decision. Clearly, there is no right or wrong here: By zooming in and trying to separate different aspects, one can come to a scale that will tap into distinct facets. The question is, is it useful?

Nera suggests at least implicitly that it would be useful to take a more fine-grained approach and tear distinct facets of conspiracy mentality apart. From my perspective, it is not self-evident that a more and more detailed granularity of concepts will prove useful. Let us entertain an analogy to classical personality constructs here: Logically, the propensity to speak with friends on the phone, to feel comfortable around strangers, and to go out frequently are independent. The fact that they do nevertheless cluster together is a strength (not a bug) of the extraversion concept. It even allows the prediction of extraverted behavior that is not part of any of the items. The same argument could be made for the case of conspiracy mentality. If distrusting the elites, suspecting agency where there is none and seeing the general population as gullible clusters together (although these could be logically orthogonal), that is a strength. If further dissecting the concept into subfacets, however, will help make substantially more accurate predictions or provide better explanations, it will prove useful. If the gain is marginal or entirely absent, parsimony should prevail.