Measuring Individual Differences in Generic Beliefs in Conspiracy Theories Across Cultures: Conspiracy Mentality Questionnaire

Bruder, Martin; Haffke, Peter; Neave, Nick; Nouripanah, Nina; Imhoff, Roland

doi:10.3389/fpsyg.2013.00225

ORIGINAL RESEARCH article

Front. Psychol., 30 April 2013

Sec. Personality and Social Psychology

Volume 4 - 2013 | https://doi.org/10.3389/fpsyg.2013.00225

This article is part of the Research Topic The Psychology of Conspiracy Theories View all 12 articles

Measuring individual differences in generic beliefs in conspiracy theories across cultures: Conspiracy Mentality Questionnaire

Martin Bruder¹

Peter Haffke¹*

Nick Neave²

Nina Nouripanah³

Roland Imhoff⁴

¹Department of Psychology, Zukunftskolleg, University of Konstanz, Konstanz, Germany
²Department of Psychology, Northumbria University, Newcastle, UK
³Department of Psychology, City University London, London, UK
⁴Department of Psychology, University of Cologne, Cologne, Germany

Conspiracy theories are ubiquitous when it comes to explaining political events and societal phenomena. Individuals differ not only in the degree to which they believe in specific conspiracy theories, but also in their general susceptibility to explanations based on such theories, that is, their conspiracy mentality. We present the Conspiracy Mentality Questionnaire (CMQ), an instrument designed to efficiently assess differences in the generic tendency to engage in conspiracist ideation within and across cultures. The CMQ is available in English, German, and Turkish. In four studies, we examined the CMQ’s factorial structure, reliability, measurement equivalence across cultures, and its convergent, discriminant, and predictive validity. Analyses based on a cross-cultural sample (Study 1a; N = 7,766) supported the conceptualization of conspiracy mentality as a one-dimensional construct across the three language versions of the CMQ that is stable across time (Study 1b; N = 141). Multi-group confirmatory factor analysis demonstrated cross-cultural measurement equivalence of the CMQ items. The instrument could therefore be used to examine differences in conspiracy mentality between European, North American, and Middle Eastern cultures. In Studies 2–4 (total N = 476), we report (re-)analyses of three datasets demonstrating the validity of the CMQ in student and working population samples in the UK and Germany. First, attesting to its convergent validity, the CMQ was highly correlated with another measure of generic conspiracy belief. Second, the CMQ showed patterns of meaningful associations with personality measures (e.g., Big Five dimensions, schizotypy), other generalized political attitudes (e.g., social dominance orientation and right-wing authoritarianism), and further individual differences (e.g., paranormal belief, lack of socio-political control). Finally, the CMQ predicted beliefs in specific conspiracy theories over and above other individual difference measures.

“Other centuries have only dabbled in conspiracy like amateurs. It is our (the Twentieth) century which has established conspiracy as a system of thought and a method of action” (Moscovici, 1987, p. 153).

Introduction

Belief in conspiracy theories continues to thrive in the twenty-first century. In Western cultures, recent popular conspiracy theories have revolved – among other themes – around the perpetrators (and possible knowing bystanders) of the 9/11 attacks on the World Trade Center in New York (Stempel et al., 2007; Swami et al., 2010), the deaths of Princess Diana (Douglas and Sutton, 2008), and Osama bin Laden (Wood et al., 2012), and the scientific evidence for climate change (Leiserowitz, 2006). These new conspiracy theories take their place next to “classics” such as alleged plots concerning the assassination of John F. Kennedy (McHoskey, 1995), the cover-up of alien contact (Harrison and Thomas, 1997), or the origins of diseases such as HIV (Ross et al., 2006).

There is increasing evidence that there are stable individual differences in people’s tendency to believe in such conspiracy theories; if a person believes in one conspiracy theory, he or she will also be more likely to believe in other conspiracy theories (Swami et al., 2010). In fact, this tendency even extends to beliefs in mutually contradictory conspiracy theories, and to beliefs in fully fictitious conspiracy theories. Thus, those who believe that Princess Diana faked her own death are also more likely to believe that she was murdered (Wood et al., 2012); those who believe in “real-world conspiracy theories” (i.e., that John F. Kennedy fell victim to an organized conspiracy) are more likely to believe that there was a conspiracy behind the success of the Red Bull energy drink – a conspiracy theory that was purposely developed for a social psychology study (Swami et al., 2011).

This has led some researchers to propose that the endorsement of specific conspiracy theories depends to a large extent on individual differences in the general tendency to adopt such beliefs, that is, a general conspiracy mentality (Imhoff and Bruder, in press). This term was originally phrased by Moscovici (1987) who understood the notion of conspiracy as implying “that members of a confession, party, or ethnicity […] are united by an indissoluble secret bond. The object of such an alliance is to foment upheaval in society, pervert societal values, aggravate crises, promote defeat, and so on.” (p. 154). As such, a conspiracy mentality then describes the general propensity to subscribe to theories blaming a conspiracy of ill-intending individuals or groups for important societal phenomena or, in more abstract terms, the tendency to subscribe to “general conspiracist beliefs” (Swami et al., 2010). Usually, such theories contradict common explanations and allege that these events are caused by secret plots by groups of powerful individuals. Individual differences in conspiracy mentality have important consequences as they predict prejudice against powerful societal groups (Imhoff and Bruder, in press). Consequences of this may either be intentions to engage in political action designed to undermine the perceived conspiracy (Imhoff and Bruder, in press) or – if the conspiracy is perceived to be overpowering – political disengagement (Butler et al., 1995). Further, conspiracy beliefs are powerful predictors of critical health behaviors such as adherence to medication regimens (Bogart et al., 2010) and vaccination uptake (Kata, 2010).

There have been a number of initial efforts to measure individual differences in conspiracy mentality (sometimes called “conspiracist ideation”; Swami et al., 2011), the most prominent of which is Swami and colleagues’ Belief in Conspiracy Theories Inventory (BCTI; Swami et al., 2010, 2011). This consists of 15 items measuring beliefs in specific conspiracy theories (e.g., “A powerful and secretive group, known as the New World Order, are planning to eventually rule the world through an autonomous world government, which would replace sovereign governments.”) and has been used in the UK and continental Europe (Swami et al., 2011) as well as East Asia (Swami, 2012). The internal reliability of this scale has consistently been very good and it relates in meaningful ways to other individual difference variables such as: support for democratic principles, political cynicism, negative attitudes to authority, and low agreeableness (Swami et al., 2011).

However, scales measuring beliefs in specific conspiracy theories are closely bound to specific temporal and geographical contexts. In response to these limitations, it has been suggested that there is a need to assess the general tendency to believe in conspiracies in a way that is not dependent on the cultural familiarity of selected theories (Brotherton, French, and Pickering; in press) and also independent of knowledge about specific conspiracy theories which may vary between cultures. For example, it is unlikely that the BCTI item concerning the New World Order is equally familiar in all countries around the world. Hitherto, there have been two attempts to address this challenge, however, neither explicitly address the cross-cultural validity of the measurement instruments. First, Brotherton et al. (in press) have developed a 15-item Generic Conspiracist Beliefs Scale on the basis of an exploratory factor analysis of 75 items. They were able to differentiate between five major components of generic conspiracy beliefs: governmental conspiracies, extraterrestrial conspiracies, informational control conspiracies, personal well-being conspiracies, and malevolent global conspiracies. These dimensions are also reflected in the final scale. Second, Imhoff and Bruder (in press) developed a 12-item Conspiracy Mentality Scale. The items of this instrument not only avoid mentioning any specific alleged conspiracy, but also do not name any specific groups that may be responsible for a conspiracy (example item: “Most people do not recognize to what extent our life is determined by conspiracies that are concocted in secret”). Both the authors of the Generic Conspiracist Beliefs Scale and the Conspiracy Mentality Scale provide initial evidence for the convergent and discriminant validity of their instruments. However, neither scale has been validated in non-Western cultures and so far neither scale has been adopted by researchers other than the original authors.

We believe that these activities attest to the scientific relevance of developing valid, reliable, and efficient instruments to measure generic conspiracy beliefs. In this context, we will propose the Conspiracy Mentality Questionnaire (CMQ), a short (5-item) measure of generic conspiracy beliefs that we administered in a large (N = 7,766) international study spanning North American (US), Western European (UK and Ireland; Germany), and Middle Eastern (Turkey) cultures (see Study 1a). The cross-cultural dimension has so far been largely absent from research on conspiracy theories (for a notable exception examining the UK and Austria, see Swami et al., 2011). Developing that dimension would be highly desirable because it is likely that individual characteristics and cultural factors interact when it comes to the belief in specific conspiracy theories, and its consequences on attitudes and behaviors. In particular, a number of studies have shown that subcultures within national groups are differentially prone to belief in conspiracy theories. For example, African American and Latino communities in the US are particularly likely to endorse conspiracy theories claiming that HIV was spread to extinguish specific ethnic groups (Ross et al., 2006). Connecting these findings with research looking at individual differences in the propensity to believe in conspiracy theories is a promising endeavor – even more so when extending the perspective to different cultures around the globe. At the global level, conspiracy theories have been identified as a driving factor in the discourses of conflict in the Middle East (Pipes, 1998). Again, identifying how such broader societal phenomena relate to, and interact with, individual characteristics constitutes a worthwhile future research agenda. Our scale is designed to facilitate such future efforts.

Overview and Hypotheses

In Study 1a, we explore the factorial structure, assess the internal consistency, and test the measurement equivalence of the CMQ across its three language versions (English, German, Turkish). Study 1b tests the temporal stability of the CMQ across a 2-week interval. In Study 1a and three subsequent studies, we then examine the validity of the scale in predicting beliefs in 33 specific conspiracy theories. Further, in Studies 2–4, we test the convergent and discriminant validity of the construct of conspiracy mentality compared to other generalized political attitudes [right-wing authoritarianism (RWA), social dominance orientation (SDO)], personality measures (Big Five, schizotypy, paranoid ideation), and further individual differences measures (e.g., domains of paranormal belief, power-, and control-related self-perceptions, anomia, death anxiety).

On the basis of the pertinent literature we expect that conspiracy mentality will evince reliable correlations with many of these constructs. In particular, conspiracy mentality should be positively related to instruments gauging perceptions of low socio-political power (Abalakina-Paap et al., 1999) and anomia (Goertzel, 1994), paranoia and schizotypy (Grzesiak-Feldman and Ejsmont, 2008; Holm, 2009), paranormal belief (Ramsay, 2006), RWA (Adorno et al., 1950), and the personality dimensions of agreeableness (negatively) and openness to experience (positively) (Swami et al., 2010, 2011). At the same time, we expect that none of the correlations will reach a level that would cast doubt on the viability of conspiracy mentality as an independent construct. In addition, we hypothesize that conspiracy mentality will be the strongest and most consistent predictor of specific beliefs in conspiracy theories, even when controlling for other individual difference measures.

Study 1A

Study 1a is the largest data collection to date measuring both beliefs in specific conspiracy theories and the general tendency to believe in conspiracy theories across four (groups of) countries: the US, UK, and Ireland, Germany, and Turkey. It allowed us (a) to examine the internal reliability and item loadings of the CMQ, (b) to test measurement equivalence of the instrument across three different language versions, (c) to investigate mean differences in conspiracy mentality across cultures as well as sex differences, and (d) to use the CMQ to predict the belief in specific conspiracy theories.

Methods

Participants

Participants initially were recruited by solicitation e-mails to participants in prior online studies conducted by the first author and by posting the study on sites listing online studies (e.g., Reips and Lengler, 2005; Kathryn Gardner’s onlinepsychresearch.co.uk and John H. Krantz’ Psychological Research on the Net). The study proved highly popular and the link spread further through social networking sites, online groups and forums, and links in relevant articles of mainstream media outlets.

A total of 7,766 volunteers responded to 38 items measuring participants’ belief in conspiracy theories. 1.2% (n = 93) of the submissions were excluded due to (a) repeated submissions from one computer, (b) identical responses to all items, (c) self-reported poor command of the questionnaire language, or (d) missing responses to more than one questionnaire item. Further, (e) we only included participants with valid data on the five items measuring participants’ conspiracy mentality (see below). This resulted in a final sample of N = 7,673 participants (4,919 men, 2,694 women, 60 unreported) between the age of 18 and 67 years (M = 29.1, SD = 10.2). Participants’ were resident in Germany (n = 5,018), the US (n = 1,126), Turkey (n = 981), or the UK and Ireland (n = 548).

Materials and procedure

When arriving on the study website, participants could choose between one of three language versions (English, German, or Turkish). They then received a very broad and non-judgmental definition of the term conspiracy theory:

“A conspiracy theory is a theory that provides an alternative explanation to the established understanding of a historical or current event. Often, it is claimed that this event is the result of conscious manipulations by individuals or secretive powers. Due to our incomplete knowledge about the world, it can usually not ultimately be decided which explanatory model is true – the established understanding of an event or the respective conspiracy theory.”

We further informed participants that we were interested in their personal beliefs in such theories. Participants then provided demographic information and proceeded to the main questionnaire page starting with the following instructions: “For each of the statements below, please use the respective rating scale to indicate how likely it is in your opinion that the statement is true. Remember that there are no “objectively” right or wrong answers and that we are interested in your personal opinion.”

The 38 items administered in this study [dubbed the Conspiracy Theory Questionnaire (CTQ) by Darwin et al., 2011] contained 33 items measuring the belief in specific conspiracy theories and five items assessing participants’ general tendency to believe in conspiracies or their conspiracy mentality. We refer to these latter five items as the CMQ (see Appendix for complete item wording).

The items included in the study were either adapted from existing ad hoc scales assessing beliefs in conspiracy theories (e.g., Goertzel, 1994; Abalakina-Paap et al., 1999; Sjörberg, 2002) or developed drawing on the content of websites and online forums dedicated to conspiracy theories. Items assessing the belief in specific conspiracy theories covered, for example, conspiracies concerning assassinations (e.g., “Princess Diana’s death was an assassination rather than an accident.”), alien landings (e.g., “There are specialized government services who attempt to harass UFO witnesses into silence.”), technological developments (e.g., “A cure for most forms of cancer has already been found, but medical circles prefer to continue to extract research funding from governments and keep their findings secret.”), and secret activities of powerful organizations (e.g., “The Vatican Bank in Rome has close links to the Mafia.”). The five items assessing participants’ conspiracy mentality consisted of general statements capturing a conspirational view of world events.

Participants indicated on 11-point scales how likely they thought each item to be true from 0 (0% – certainly not) to 10 (100% – certain). Each scale point was additionally labeled with increasing probabilities in steps of 10 percentage points.

The original English items were translated into German and Turkish in a procedure similar to that proposed by Brislin (1970): Each item was first translated by a native speaker of the respective language and then translated back into English by a different person competent in both languages. Disagreements between the original and the backtranslated version and their implications for the translated version were resolved by discussion. In total, n = 5,026 participants completed the German version, n = 1,640 the English version, and n = 1,007 the Turkish version of the questionnaire.

After completing the questionnaire, participants had the opportunity to compare their own score with that of a normative sample, were debriefed, and thanked for their support.

Statistical analyses

Rather than being taken for granted, tests should be employed to determine whether a different language version of one questionnaire measures the same construct in an equivalent manner across different cultures (see Geisinger, 1994; van de Vijver and Hambleton, 1996; Hambleton, 2005). As a first step, we explored the factorial structure of the five items across the full sample and within each language version using principal component analyses.

Second, we tested for measurement equivalence (ME; also measurement invariance or metric invariance) across the three language versions. The focus here was on the invariant operation of the items and, in particular, on equivalent factor loadings across groups (Byrne, 2008, 2010). Using SPSS AMOS (Version 20), we conducted a multi-group confirmatory factor analysis and compared a configural model without constraints with a second model, in which factor loadings were constrained to be equal across groups. The second model is computed by freely estimating parameters for the first group and then constraining factor loadings for the other groups to be equal to those of the first one. Measurement equivalence can therefore be assumed if, through comparison, the models are deemed to be equally fitting of the data as this implies that the additional constraints did not substantially reduce model quality. Statistical testing for measurement equivalence is sometimes based on the comparison of each model’s χ² values; however, this is a highly conservative method and is often seen as “impractical and unrealistic”(Byrne, 2009). When, as is the case in our analyses, the sample size is large, this often leads to statistically significant χ² differences, even in case of negligible differences between the two models. We thus follow recommendations in the relevant literature and also compare changes in other fit indices between the unconstrained and the constrained model as a second source of information to evaluate measurement equivalence. In particular, we examined the comparative fit index (CFI), the Tucker–Lewis-Index (TLI), and the root mean square error of approximation (RMSEA). According to Hu and Bentler (1999), a cut-off value of 0.95 for CFI and TLI, and a value of less than 0.06 for RMSEA indicate a good model fit. For the RMSEA, values of up to 0.08 are still considered to indicate reasonable model fit (Browne and Cudeck, 1992) and values between 0.08 and 0.10 are considered to reveal mediocre fit (MacCallum et al., 1996). PCLOSE provides a one-sided test of the null hypothesis that RMSEA equals 0.05. PCLOSE-values larger than 0.05 indicate that the null hypothesis is not rejected – the model is then described as a “close fitting” model (Kenny, 2012). For changes in CFI and TLI values (ΔCFI, ΔTLI) in the context of examining measurement or structural equivalence (SE), Cheung and Rensvold (2002) proposed that only differences larger than 0.01 should lead one to reject the assumption of measurement equivalence. For ΔRMSEA, Chen (2007) recommended that only a difference of > 0.015 should lead to a rejection of measurement equivalence.

Testing the cross-cultural measurement properties of a psychometric scale can be extended to tests of SE. Here, the focus is on the unobserved (or latent) variables. Tests of SE can be used to check whether the factorial structure of a measurement instrument is the same across groups (Byrne, 2009).

Results

Descriptive statistics

Item statistics (mean values, standard deviations, and item discrimination coefficients) for each of the five CMQ items for each of the language versions are presented in Table 1. Each of the three language versions had good (German: α = 0.84; English: α = 0.84) or acceptable (Turkish: α = 0.72) internal consistency and very good to medium item discrimination statistics (r_itc > 0.37).

TABLE 1

Table 1. Conspiracy Mentality Questionnaire item statistics for each of the three language versions.

Factorial structure and model fit

We first conducted an exploratory factor analysis across all participants to examine the factorial structure of the CMQ. The screen plot criterion suggested a one-factor solution explaining 60.6% of the variance. All factor loadings were larger than 0.71. Repeating this analysis for each language version separately also suggested a one-factor solution for each language version (see Figure 1). However, the proportion of explained variance was lower for the Turkish-speaking sample (50.3%) than for the German-speaking (61.0%) and English-speaking (61.2%) samples. Except for one item, all items strongly loaded on the extracted component (individual factor loadings were larger than 0.69 across all language versions). The one exception was the item “I think government agencies closely monitor all citizens” in the Turkish CMQ version with a still acceptable loading of 0.54.

FIGURE 1

Figure 1. Screen plot for exploratory factor analysis for each of the three language versions (Study 1a).

Measurement equivalence across language versions

Using multi-group confirmatory factor analysis, we proceeded to test a one-factor model across the three different language versions. In this configural model, there were no constraints imposed on the parameters in the model; however, there was the minimal requirement that the number of factors and the number of items associated with each factor are equal across groups. Thus, this model tests whether the one-factor structure observed in the exploratory analyses can be confirmed when taking all three groups into account simultaneously.

As expected for our large sample, χ² values were significant. However, examining alternative fit indices revealed very good model fit (see Table 2) with the CFI and the TLI being larger than 0.95 and the RMSEA smaller than 0.05. PCLOSE indicated a close fitting model. Thus we proceeded to tests of more restrictive models.

TABLE 2

Table 2. Tests for measurement and structural equivalence of the conspiracy mentality questionnaire across the three language versions (English, German, and Turkish).

In the measurement equivalence model, factor loadings were constraint to be equal across groups. Apart from the significant χ² value, this model showed very good model fit which was highly similar to that of the unconstrained model. Again, the CFI, TLI, and RMSEA were within a value range that is generally accepted to indicate very good fit (see Table 2) and PCLOSE remained non-significant. In addition, the changes in all of these values between the configural model and the measurement equivalence model were minimal, all within a range supporting ME (Cheung and Rensvold, 2002; Chen, 2007). Thus, the CMQ is of equivalent measurement quality across all of its three language versions.

Building on the confirmatory results concerning measurement equivalence, we proceeded to test the SE of the CMQ across the three language versions. This is done by constraining both factor loadings and factor (co) variances to be equal across groups. The SE model showed acceptable model fit with both the CFI and the RMSEA falling within the range of well-fitting models and the TLI only falling very slightly short of the value of 0.95 (Hu and Bentler, 1999). Comparing the SE model with the less restrictive ME model revealed acceptable invariance only when using differences in RMSEA to assess model equivalence. These differences were minor (ΔRMSEA = −0.012) and within the suggested range (Chen, 2007). However, both changes in the CFI and the TLI are somewhat larger than would be expected for equivalent models (ΔCFI = 0.018; ΔTLI = 0.020). Inspection of the model modification indices showed that removing the constraint on the latent factor variance for the Turkish group to be equal to that of the two other groups would have resulted in a better fitting model. Indeed, such a model was highly similar to the measurement equivalence model with ΔCFI = 0.001, ΔTLI < 0.001, and ΔRMSEA < 0.001. Thus, although the measurement properties of the individual items of the CMQ are stable across language versions, questions remain as to whether the latent construct of conspiracy mentality is as coherent in Turkey as in Western European and North American countries.

Differences in mean levels of conspiracy beliefs

Having established that our five item measure operates well in all three language versions, this allowed us to investigate between-country differences in conspiracy mentality. We used a one-way ANOVA with Tukey HSD post hoc tests to examine mean differences between countries and independent groups t-tests to test sex differences between men and women in each country (Bonferroni-corrected α = 0.0125). Previous research in different disciplines suggests that conspiracy beliefs may be particularly rife in the Near East (e.g., Zonis and Joseph, 1994; Pipes, 1998; Gray, 2010) and we thus hypothesized that our Turkish sample would show higher mean scores than the samples from Western Europe and the US. As far as sex is concerned, Darwin et al. (2011) reported no significant differences. However, our larger sample allowed us to re-examine the idea that women may be more prone to believe in conspiracy theories than men. This may be because women are structurally disadvantaged in many societies and powerless individuals and groups are more susceptible to conspiracy beliefs (Abalakina-Paap et al., 1999). Furthermore, women are more likely to believe in paranormal phenomena (Irwin, 1993), a tendency that in itself is linked to higher conspiracy beliefs (Darwin et al., 2011).

Regarding differences in conspiracy mentality across the four countries, the ANOVA showed a significant main effect, F(3, 7669) = 120.32, p < 0.001, $η_{p}^{2}$ = 0.045. Tukey HSD follow-up comparisons revealed significant group differences between each pair of countries, ps < 0.01, except for the comparison between the US (M = 6.3; SD = 2.0) and UK/Ireland (M = 6.3; SD = 1.9), p = 1.00. Turkish participants’ (M = 7.3; SD = 2.1) conspiracy mentality was markedly higher than that of all other groups (ds > 0.58) whereas German participants scored lower than the other groups (M = 5.9; SD = 2.1). However, effect sizes of mean differences between Germany, the US, and the UK were small, |ds| < 0.20.

As far as sex differences are concerned, the only significant effect was observed for the US, t(1119) = 3.40, p = 0.001, d = 0.20, with women (M = 6.5; SD = 1.9) scoring slightly higher than men (M = 6.1; SD = 2.1).

In sum, the mean comparisons supported our contention that the susceptibility to believe in conspiracy theories is more pronounced in Near Eastern countries (i.e., Turkey) than in Western countries (Germany, UK/Ireland, US). In line with Darwin et al.’s (2011) findings, we also did not observe strong support for systematic sex differences, although the findings revealed that the hypothesis of women’s greater tendency to belief in conspiracy theories held in our US sample (despite a small effect size).

Predicting specific conspiracy beliefs

We also tested the extent to which conspiracy mentality as measured by the CMQ predicts endorsement of each of the 33 specific conspiracy theory items also included in the questionnaire. The CMQ correlated strongly with each of the specific conspiracy items, with correlations ranging from 0.37 to 0.76, each significant at a level of p < 0.001. The average correlation coefficient was $\bar{r} = 0.58$ , obtained by averaging Fisher’s z transformed correlation coefficients (StatSoft Inc., 2012).

Study 1B

Study 1b examined the 2-week test-retest reliability of the CMQ.

Method

The German version of the CMQ was administered twice at the end of an otherwise unrelated study. The time lag between the invitations to the two assessments was 15 days. All participants were members of a German university. Of 177 participants who completed the CMQ at t1, 75% or 133 participants (82 women, 50 men, 1 unidentified; M_Age = 24.27, SD_Age = 5.21) also completed assessment at t2. Dropout was independent of CMQ scores at t1, t(175) = 1.07, p = 0.28.

Results

Internal consistency of the 5-item CMQ was satisfactory both at t1 (α = 0.77) and t2 (α = 0.82). The correlation between the two assessment points was 0.84.

Summary Studies 1A and 1B

Study 1a established the factorial structure and internal consistency of a new 5-item psychometric instrument – the CMQ – designed to measure the general tendency to believe in conspiracy theories that characterizes a conspiracy mentality. It further provided support for measurement equivalence of three different language versions of the CMQ (English, German, and Turkish). As far as SE was concerned, there remained some doubts as to whether a one-factorial structure can adequately capture the construct of a “conspiracy mentality” in Turkey. Comparisons of mean levels of conspiracy mentality revealed higher levels in Turkey than in Western European and North American countries. Sex differences were small with the only significant sex effect showing higher levels of conspiracy mentality among US women than men. Study 1b demonstrated satisfactory test-retest reliability of the CMQ over a 2-week interval.

Study 2

Studies 2–4 are concerned with the convergent and discriminant validity of the CMQ with other individual differences measures. Study 2, a focused re-analysis of data presented by Darwin et al. (2011) examined the relations between conspiracy mentality as assessed by the CMQ and (a) people’s tendency to belief in paranormal events, (b) paranoid ideation, and (c) a schizotypal personality disposition. A tendency to believe in the paranormal manifests itself in the acceptance of events and processes that accepted science currently deems impossible; paranoid ideation is characterized by fears and suspicions about physical and psychological threats potentially posed by social agents; and schizotypy is a mild form of schizophrenia that involves suspicion, magical thinking, paranoid beliefs, and different forms of social anxiety. All three of these constructs have been theoretically and empirically linked to beliefs in conspiracy theories (Ramsay, 2006; Grzesiak-Feldman and Ejsmont, 2008; Holm, 2009; Darwin et al., 2011).