TL;DR
This paper investigates whether conspiracy beliefs form a coherent "belief system" by analyzing the structure and organization of 20 different conspiracy theories. The study reveals that these beliefs are primarily organized along two independent dimensions: one related to partisan/ideological identities and the other to anti-social orientations. This finding has significant implications for how we study and understand conspiracy beliefs, suggesting that the specific theories examined heavily influence research conclusions.
What Problem Does This Paper Address?
While conspiracy theories are often referred to as a "belief system," empirical research has largely focused on the correlations between individual conspiracy beliefs rather than their underlying structure. This leaves a gap in understanding how these beliefs are organized in people's minds. Without this understanding, it's difficult to accurately decipher which theories "go together," predict who might believe in specific theories, or understand the dispositions associated with them. The paper aims to address this by exploring the substantive dimensions and subgroupings that organize conspiracy beliefs, moving beyond simple correlation analysis.
Methods
The researchers employed Multidimensional Scaling (MDS), a technique designed to represent dissimilarities between objects as distances in a low-dimensional space. Specifically, they used nonmetric MDS because the survey data used ordinal scales. This method allows for the visualization of how different conspiracy beliefs relate to each other. To further interpret the structure, they also used hierarchical cluster analysis (HCA) to group similar conspiracy beliefs and embedded external variables (like political identity and personality traits) into the MDS configuration to identify the substantive dimensions.
Datasets
The study utilized data from two national surveys conducted in March 2020 and October 2020.
- March 2020 Survey: Included 2,023 respondents and data on beliefs in 20 conspiracy theories.
- October 2020 Survey: Included a sample representative of the U.S. population and data on 16 conspiracy theories, with 7 of these being new compared to the first survey.
These surveys are noted as being larger in scope than many previous studies on conspiracy beliefs.
Key Insights and Ideas
- Low Dimensionality is Key: Conspiracy beliefs aren't a chaotic mess; they can be explained by a surprisingly simple, low-dimensional structure, consistent with the idea of a belief system.
- Two Independent Pillars: The primary organizing principles for conspiracy beliefs are not what you might initially expect. They are largely driven by either:
- Political Identity: Where you stand on the left-right spectrum.
-
Anti-Social Orientations: Traits like narcissism, psychopathy, and a willingness to accept political violence. Crucially, these two dimensions appear to be largely uncorrelated.
-
Content Matters: Conspiracy theories naturally cluster based on their subject matter. For instance, theories about partisan actors tend to group together, as do those related to science and medicine.
- "Conspiracy Thinking" as a Shorthand: The general predisposition to believe in conspiracy theories ("conspiracy thinking") acts as a strong, consistent correlate across all types of conspiracy beliefs, similar to how liberal-conservative self-identification can predict specific issue stances.
- Beware of Generalizations: The specific conspiracy theories included in a study significantly impact the conclusions drawn about their correlates. What might be true for one set of theories may not hold for another.
Key Results
- A two-dimensional solution was found to adequately explain the variability in conspiracy beliefs across both surveys, with a low STRESS coefficient (.074 in the first survey, .045 in the replication).
- Hierarchical cluster analysis identified 6 distinct clusters of conspiracy theories, some of which were strongly associated with partisan identities (e.g., theories maligning Democrats or Republicans) and others with anti-social traits.
- External variable regressions revealed that partisan and ideological identities formed one dimension, while psychopathy, Machiavellianism, narcissism, and acceptance of violence formed a second, orthogonal dimension.
- The study found that inferences about the correlates of conspiracy beliefs are highly contingent on the specific theories examined. For example, trust in government showed different relationships with different clusters of conspiracy beliefs.
Limitations and Future Work
- Correlational Nature: The study acknowledges that its analyses are correlational, meaning causal claims cannot be made.
- Limited Number of Theories: While more theories were included than in many previous studies, the authors note that more data would provide a clearer picture of stable groupings.
- Sociodemographic Controls: The analyses controlled for basic sociodemographic factors but omitted more specific substantive controls that could influence findings.
- Future Research Directions:
- Researchers should be more circumspect in their labeling of conspiracy theories, using "partisan," "ideological," or "political" labels when appropriate.
- Future studies should poll on diverse types of conspiracy theories and generate distinct scales to test hypotheses across these types.
- Investigating the stability and durability of these belief systems over time is crucial.
- Applying this framework to non-U.S. samples is needed to assess cross-cultural generalizability.
- Further work could combine social and epistemological factors with political and psychological ones to understand attraction to conspiracy theories.
Summary updated