1. Introduction
Perceived consensus, which refers to the level of agreement among individuals within a social group regarding a particular issue, plays a crucial role in shaping and modifying individuals’ attitudes and beliefs toward scientific subjects (van der Linden, Reference van der Linden2021; van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019). Consequently, the perceived agreement among experts on specific topics can have significant societal implications, such as influencing individuals’ beliefs about fundamental scientific principles and garnering public support for related policies (Ding et al., Reference Ding, Maibach, Zhao, Roser-Renouf and Leiserowitz2011). Extensive research has been conducted to comprehend the impact of perceived consensus on attitudes toward pressing matters like climate change (van der Linden et al., Reference van der Linden, Leiserowitz and Maibach2016), genetically modified foods (Dixon, Reference Dixon2016), and vaccination (van der Linden et al., Reference van der Linden, Clarke and Maibach2015a). Moreover, studies have explored how perceived consensus can be utilized to counteract the influence of misinformation (Constantino et al., Reference Constantino, Sparkman, Kraft-Todd, Bicchieri, Centola, Shell Duncan and Weber2022; Maertens et al., Reference Maertens, Anseel and van der Linden2020; van der Linden, Reference van der Linden2021; Williams and Bond, Reference Williams and Bond2020). Considering its effect on the perception of scientific issues, it is reasonable to assume that perceived consensus would also shape beliefs regarding more specific conspiracy theories. In this study, we aim to investigate this matter by examining how perceived consensus among both experts and laypeople predicts beliefs in localized and specific conspiracy theories within Turkey, a non-WEIRD (WEIRD stands for Western, educated, industrialized, rich, and democratic; Henrich et al., Reference Henrich, Heine and Norenzayan2010) country. Previous research attempting to quantify WEIRDness showed Turkey to be a non-WEIRD country (Klein et al., Reference Klein, Vianello, Hasselman, Adams, Adams, Alper and Sowden2018) and other research found Turkish culture to be quite distant from US American culture, which was considered to be the baseline for a WEIRD culture (Muthukrishna et al., Reference Muthukrishna, Bell, Henrich, Curtin, Gedranovich, McInerney and Thue2020). In addition to the general sample characteristics, the conspiracy theories we selected for this study (see below for further details) are non-WEIRD as they are intrinsically anti-Western and anti-European in their content. Furthermore, we assess the accuracy of conspiracy believers’ estimations by comparing them with the actual level of consensus.
The most prominent theoretical framework regarding the consequences of perceived consensus is the Gateway Belief Model (GBM; van der Linden, Reference van der Linden2021; van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019). According to the GBM, perceived scientific consensus predicts attitudes toward scientific matters, which subsequently influence support for relevant public actions. Extensive empirical evidence supporting this model has been obtained from various large-scale samples (Hornsey et al., Reference Hornsey, Harris, Bain and Fielding2016; van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019) and longitudinal studies (Kerr and Wilson, Reference Kerr and Wilson2018; Maertens et al., Reference Maertens, Anseel and van der Linden2020). One of the extensively investigated issues within this context is attitudes toward climate change. In addition to experimental investigations on the consensus effect concerning attitudes toward climate change (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019), researchers have also examined the most effective ways to communicate climate change consensus (e.g., presenting plain facts, pie charts; van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2014), the persistence of the consensus effect (Maertens et al., Reference Maertens, Anseel and van der Linden2020), and potential psychological reactance as a by-product (van der Linden et al., Reference van der Linden, Leiserowitz and Maibach2018).
Drawing upon a high-powered sample from the United States (N = 6,301), van der Linden et al. (Reference van der Linden, Maibach and Leiserowitz2019) conducted an experiment where participants in the treatment condition were exposed to a consensus message (e.g., ‘97% of climate scientists have concluded that human-caused global warming is occurring’), while the control group engaged in a short neutral word sorting task. Consistent with previous findings (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b), they discovered that individuals exposed to the consensus message reported higher levels of perceived scientific agreement. This, in turn, led to stronger beliefs in global warming, beliefs in human causation, concerns about global warming, and subsequently greater support for public action regarding climate change (van der Linden et al., Reference van der Linden, Maibach and Leiserowitz2019).
Research indicates that despite at least 97% of peer-reviewed literature providing evidence supporting the human-caused nature of climate change (see Lynas et al., Reference Lynas, Houlton and Perry2021; Maibach et al., Reference Maibach, Myers and Leiserowitz2014), the general public remains unaware of the high scientific consensus on this topic (Leiserowitz et al., Reference Leiserowitz, Maibach, Roser-Renouf, Feinberg and Rosenthal2014). This underscores the real-life significance of perceived consensus. Consistent with the GBM, underestimation of scientific consensus can lead to reduced support for policies aimed at addressing human-caused climate change (Ding et al., Reference Ding, Maibach, Zhao, Roser-Renouf and Leiserowitz2011) and hinder engagement in prosocial behaviors such as signing petitions or volunteering for charitable causes (van der Linden, Reference van der Linden2015). Consequently, this body of research has inspired the development of interventions that leverage social norms and consensus to positively influence individuals’ attitudes and behaviors concerning climate change (Constantino et al., Reference Constantino, Sparkman, Kraft-Todd, Bicchieri, Centola, Shell Duncan and Weber2022).
Another area of investigation within the realm of perceived consensus revolves around attitudes toward health-related subjects, particularly vaccination. Research indicates that individuals who perceive a lack of scientific consensus regarding the link between vaccination and autism are more inclined to believe in the misconception of an autism–vaccine connection (Dixon and Clarke, Reference Dixon and Clarke2013; van der Linden et al., Reference van der Linden, Clarke and Maibach2015a). Moreover, in the context of the COVID-19 pandemic, Kerr and van der Linden (Reference Kerr and van der Linden2022) found both correlational and causal evidence in two large samples (N = 7,206 and N = 1,856, respectively) demonstrating that perceived scientific consensus on the severity of COVID-19 predicted attitudes toward the threat, concerns about the virus, and subsequent support for mitigation policies. Collectively, these studies highlight the positive impact of emphasizing scientific consensus among experts, leading to more favorable attitudes toward vaccines (Dixon and Clarke, Reference Dixon and Clarke2013; Kerr and van der Linden, Reference Kerr and van der Linden2022; van der Linden et al., Reference van der Linden, Clarke and Maibach2015a).
Notably, similar to scientific beliefs, political beliefs are also influenced by perceived consensus. Weinschenk et al. (Reference Weinschenk, Panagopoulos and van der Linden2021), utilizing data from a national survey conducted after the 2020 US presidential election, found that individuals expect the majority of the general public to share their views on election norms.
Given the significance of perceived consensus in both scientific and political domains, it becomes evident that it should also be relevant to specific conspiracy beliefs. Albarracín Reference Albarracín2022 proposes that normative plausibility, defined as the extent to which individuals perceive others to believe that a conspiracy allegation is likely to be true, is a crucial factor in perceiving conspiracy theories as plausible. Consequently, it is assumed that people are susceptible to perceiving a general consensus regarding the credibility of conspiracy theories. However, research exploring the relationship between perceived consensus and conspiracy beliefs remains limited. Only a few studies exist, such as Cookson et al. (Reference Cookson, Jolley, Dempsey and Povey2021a), who demonstrated a strong association between individuals’ personal anti-vaccine conspiracy beliefs and their perception of in-group members’ anti-vaccine beliefs. Additionally, Cookson et al. (Reference Cookson, Jolley, Dempsey and Povey2021b) showed that providing normative feedback regarding the actual beliefs of the in-group concerning vaccines led individuals to update their perceived beliefs, subsequently influencing their intentions toward vaccination.
Due to the scarcity of studies examining the impact of perceived consensus on conspiracy beliefs, the present study aims to address this gap by investigating specific conspiracy theories. We aim to analyze how perceived consensus among both experts in the field and laypeople predict belief in conspiracy theories. This will be achieved by examining specific conspiracy theories within the understudied cultural context of Turkey, testing individuals’ perception of agreement against actual agreement, and utilizing a comprehensive set of correlates.
2. Overview of the current study
The current study aims to make several contributions to the existing literature on conspiracy theories. First, it seeks to address a notable gap by examining the association of perceived consensus with specific conspiracy theories. Specific conspiracy theories (i.e., conspiracy theories with very specific content; e.g., 9/11 conspiracy theories) differ from a general conspiracy mindset (i.e., a generalized tendency to believe that most world affairs are shaped by ongoing conspiracies; Imhoff et al., Reference Imhoff, Zimmer, Klein, António, Babinska, Bangerter and Van Prooijen2022) and provide more concrete examples of irrational beliefs observed in real-life situations (Sutton and Douglas, Reference Sutton and Douglas2020).
Second, the study investigates the generalizability of the effect of perceived consensus to non-WEIRD contexts by utilizing specific conspiracy theories from Turkey, an understudied cultural setting (see Henrich et al., Reference Henrich, Heine and Norenzayan2010). There is a scarcity of research on the influence of perceived consensus on scientific or conspiracy beliefs in non-WEIRD populations (van der Linden, Reference van der Linden2021). While a few studies have demonstrated a consistent relationship between scientific consensus and scientific beliefs in non-WEIRD samples, such as the case of a Japanese sample (Kobayashi, Reference Kobayashi2018), further studies are needed to validate the GBM in non-WEIRD contexts. Past research has showed that trust in science and scientists might be dependent on cultural factors (Borgonovi and Pokropek, Reference Borgonovi and Pokropek2020), perception of scientists (Kossowska et al., Reference Kossowska, Szwed and Czarnek2021; Nera et al., Reference Nera, Mora, Klein, Roblain, Van Oost, Terache and Klein2022), and the level of country-level corruption (Alper et al., Reference Alper, Yelbuz, Akkurt and Yilmaz2022), which all suggest the potential use of replicating this line of research on underrepresented samples. For this study, two sets of conspiracy beliefs specific to Turkey have been identified. One set pertains to the Lausanne Treaty, which laid the foundation for the Republic of Turkey. A long-standing conspiracy theory suggests the existence of secret articles within the treaty that restrict Turkey from exercising full sovereignty over its territory, with these articles purportedly expiring in 2023, coinciding with the treaty’s 100th anniversary (Basegmez, Reference Basegmez2023). These claims were invented to attack the secular founding fathers of the Republic of Turkey, especially Ismet Inonu, who was the head of the Turkish delegation at Lausanne, to insinuate that these secular founding fathers did not get the best deal possible and secretly collaborated with the foreign and non-Muslim powers (Basegmez, Reference Basegmez2023). As such, these conspiracy beliefs are more likely to be endorsed by conservative people in Turkey (KONDA, 2018). Any suggestion of a secret article is misinformation and these allegations are definitively debunked by historians and official institutions, despite Turkey has been ruled by a conservative, religious government for the last 20 years (‘No classified article in Treaty of Lausanne: CİMER’, 2022). In short, the first set of conspiracy theories are regarding the unfounded claims about the Lausanne Treaty, which are relatively more popular in conservative circles.
The second set of specific conspiracy theories relate to alleged rich oil and mineral deposits in Turkey, with claims that ‘foreign powers’ clandestinely pressure Turkish authorities not to extract these resources (Ozavci, Reference Ozavci2023). In addition to general claims about unused mining reserves, we also included an item on ‘boron’ mines, which are believed to be worth of billions of dollars among conspiracy circles (‘Hangi komplo teorilerine inanıyoruz?’, 2020). We also included an item pertaining to the belief in the ‘contorium’ conspiracy, which revolves around fictitious reserves of a non-existent element, purportedly valued at 17 trillion dollars. This particular conspiracy theory originated as a joke by a blogger, and a national poll conducted in 2018 revealed that 34% of the population subscribed to this belief (KONDA, 2018; Malumatfurus, 2016). In short, the second set of conspiracy theories are regarding the unfounded claims about mining reserves in Turkey, which cannot be extracted because of clandestine pressures by foreign (Western, to be more specific) powers.
Third, this study employs the most comprehensive set of correlates to test how well they predict the conspiracy beliefs in question as well as the perceived consensus regarding them. The correlates encompass sociodemographic factors (age, sex, education, and perceived socioeconomic status), differences in worldview (ideology, religiosity, right-wing authoritarianism, and social dominance orientation), cognitive sophistication (cognitive reflection, preferences for intuitive and effortful thinking, and science literacy), and personality traits (narcissism, Machiavellianism, and psychopathy). Existing literature shows that the dark triad personality traits (Hughes and Machan, Reference Hughes and Machan2021; March and Springer, Reference March and Springer2019; Uscinski et al., Reference Uscinski, Enders, Diekman, Funchion, Klofstad, Kuebler and Wuchty2022) and indications of low socioeconomic status, such as lower income (Constantinou et al., Reference Constantinou, Kagialis and Karekla2021; Freeman and Bentall, Reference Freeman and Bentall2017) and lower education level (Douglas et al., Reference Douglas, Sutton, Callan, Dawtry and Harvey2016; van Prooijen, Reference van Prooijen2017), are associated with belief in conspiracy theories. Beliefs and biases such as ideology (Imhoff et al., Reference Imhoff, Zimmer, Klein, António, Babinska, Bangerter and Van Prooijen2022; Sutton and Douglas, Reference Sutton and Douglas2020; Uscinski et al., Reference Uscinski, Enders, Diekman, Funchion, Klofstad, Kuebler and Wuchty2022), religiousness, (Alper et al., Reference Alper, Bayrak and Yilmaz2020; Frenken et al., Reference Frenken, Bilewicz and Imhoff2023), right-wing authoritarianism (Bruder et al., Reference Bruder, Haffke, Neave, Nouripanah and Imhoff2013; Grzesiak-Feldman and Irzycka, Reference Grzesiak-Feldman and Irzycka2009; Prichard and Christman, Reference Prichard and Christman2020; Swami, Reference Swami2012), and social dominance orientation (Bruder et al., Reference Bruder, Haffke, Neave, Nouripanah and Imhoff2013; Swami, Reference Swami2012) are some of the important factors that are associated with conspiracy beliefs. Also, cognitive dispositions such as cognitive reflection (Alper et al., Reference Alper, Bayrak and Yilmaz2020; Binnendyk and Pennycook, Reference Binnendyk and Pennycook2022; Kantorowicz-Reznichenko et al., Reference Kantorowicz-Reznichenko, Folmer and Kantorowicz2022; Swami et al., Reference Swami, Voracek, Stieger, Tran and Furnham2014; Yelbuz et al., Reference Yelbuz, Madan and Alper2022) and scientific reasoning (Čavojová and Ersoy, Reference Čavojová and Ersoy2020) are found to be negatively related with conspiracy beliefs. By accounting for variations in individuals’ worldview, cognitive abilities, personality traits, and sociodemographic characteristics, the current study aims to investigate the predictors of specific conspiracy beliefs and the corresponding perceived consensus. Furthermore, these analyses will provide valuable insights into the shared predictors underlying both conspiracy beliefs and perceived consensus, thus laying the foundation for future research endeavors aimed at understanding the underlying mechanisms driving their association.
Fourth, unlike previous studies, this research examines not only the perceived consensus among experts but also among laypeople (i.e., other participants in the same study). It is expected that, similar to the impact of perceived agreement among experts, perceived agreement among laypeople would also predict higher beliefs in the specific conspiracy theory under investigation.
Fifth, similar to studies conducted on climate change (e.g., van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019), the study takes into account the actual consensus among experts by consulting individuals with relevant expertise. Additionally, the consensus among laypeople is assessed by calculating the number of participants who agree with the conspiracy theories. These two distinct measures allow for a comparison of the strength of the effect of perceived consensus among experts versus laypeople.
Lastly, to examine the potential priming effect, the study randomizes the order of self-reported beliefs and expectations about others’ agreement with the conspiracy theories. This experimental design enables investigation into whether considering others’ perspectives has an immediate causal influence on belief in conspiracy theories.
3. Method
3.1. Sample
The sample was collected as part of an ongoing longitudinal study. Participants completed the measures of cognitive sophistication and personality (see below) in the first wave, while they completed other measures in the second wave, which took place five months after the first wave. After removing duplicate entries and participants who did not complete all materials, the resulting sample consisted of 729 participants (M age = 31.64, SD = 9.68; 497 females, 228 males, 4 other). The dataset is available online at https://osf.io/mfuyz?view_only=00f5e092d9c0440b8a074568fc6dc36d.Footnote 1
To ascertain the actual levels of expert consensus, we sought the input of several professionals in the respective fields. Specifically, we obtained self-reported data from a group of professors specializing in the history of the Turkish Republic (N = 16), who provided their agreement levels regarding the conspiracy theories related to the Lausanne Treaty. Additionally, we collected responses from engineers specializing in mining, geology, or geophysics (N = 216) to gauge their agreement levels concerning the conspiracy theories surrounding unused mining reserves.
3.2. Materials
3.2.1. Conspiracy theories on the Lausanne Treaty
Participants stated how much they agree (1 = strongly disagree, 5 = strongly agree) with two items (‘There are secret articles in the Treaty of Lausanne’. and ‘Some clauses in the Treaty of Lausanne will expire in 2023’.). They also estimated the percentage of experts of the history of Turkish Republic who would ‘agree’ or ‘strongly agree’ with these items on a scale ranging from 0% to 100%. Using a similar scale, they also estimated the percentage of participants of the current study who would ‘agree’ or ‘strongly agree’ with these items.
3.2.2. Conspiracy theories on the mining reserves in Turkey
The same response format as in the conspiracy theories on the Lausanne Treaty was used. The only difference was the items. There were three items regarding the conspiracy theories on the mining reserves in Turkey (‘Turkey cannot utilize most of its boron mines because foreign countries do not give permission.’, ‘Turkey has rich underground energy resources, but they cannot be extracted due to international agreements.’, ‘There is a “Contorium” reserve worth 23 trillion dollars in the Bosphorus, but foreign powers do not allow it to be extracted’.).
3.3. Correlates
3.3.1. Sociodemographic factors
Participants completed measures on age, sex (male, female, other), education status (1 = primary school, 7 = doctoral degree), and perceived socioeconomic status,
3.3.2. Worldview
Participants completed measures on ideology (1 = extremely left-wing, 7 = extremely right-wing) and religiosity (1 = not religious at all, 7 = very religious). They also filled out 6-item right-wing authoritarianism (Bizumic and Duckitt, Reference Bizumic and Duckitt2018; α = .703), and 4-item social dominance orientation (α = .472; Pratto et al., Reference Pratto, Çidam, Stewart, Zeineddine, Aranda, Aiello and Henkel2013) scales.
3.3.3. Cognitive sophistication
Participants completed 3-item cognitive reflection (Frederick, Reference Frederick2005; Yilmaz and Saribay, Reference Yilmaz and Saribay2016; α = .774), 6-item preference for intuitive (α = .914) and 6-item preference for effortful thinking (α = .816) (Bayrak et al., Reference Bayrak, Doğruyol, Alper and Yilmaz2023; Newton et al., Reference Newton, Feeney and Pennycook2023), and 16-item science literacy (McPhetres et al., Reference McPhetres, Rutjens, Weinstein and Brisson2019; α = .816) measures.
3.3.4. Personality
Participants completed measures of 4-item narcissism (α = .740), 4-item Machiavellianism (α = .809), and 4-item psychopathy (α = .632) (Jonason and Webster, Reference Jonason and Webster2010; Özsoy et al., Reference Ozavci2017) scales.
4. Results
For both conspiracy theories on the Lausanne Treaty (‘There are secret articles in the Treaty of Lausanne’. and ‘Some clauses in the Treaty of Lausanne will expire in 2023’.), the percentage of professors of the history of Turkish Republic (N = 16)Footnote 2 who ‘strongly agreed’ or ‘agreed’ was 0%. However, both conspiracy believers (who ‘strongly agreed’ or ‘agreed’) and non-believers (who ‘strongly disagreed’ or ‘disagreed’) perceived expert consensus to be higher than the actual rate. Conspiracy believers, compared to non-believers, overestimated the expert consensus by a larger margin (see Figure 1). The same pattern was found in participants’ estimations of laypeople consensus (average agreement among all participants of the study) (see Figure 2).
Figure 1 Conspiracy believers’ and non-believers’ estimation of expert consensus and the actual rate of expert consensus for conspiracy theories on the Lausanne Treaty.
Figure 2 Conspiracy believers’ and non-believers’ estimation of laypeople consensus and the actual rate of laypeople consensus for conspiracy theories on the Lausanne Treaty.
Beliefs in both conspiracy theories about the Lausanne Treaty positively correlated with perceived consensus among laypeople and experts in the history of Turkish Republic. Participants’ own beliefs in these conspiracy theories correlated relatively more strongly with the perceived expert consensus, compared to perceived laypeople consensus (see Table 1).
Table 1 Correlations between belief in conspiracy theories on the Lausanne Treaty and perceived consensus
Note: ‘Difference’ test refers to the z-test for the difference between two dependent correlations with one variable in common (Lee and Preacher, Reference Lee and Preacher2013).
For all three conspiracy theories on the mining reserves in Turkey, the percentage of engineers of mining, geology, and geophysics (N = 216) who ‘strongly agreed’ or ‘agreed’ ranged from 0.9 to 6.48% (see Figure 3). Similar to the conspiracy theories on the Lausanne Treaty, both conspiracy believers (who ‘strongly agreed’ or ‘agreed’) and non-believers (who ‘strongly disagreed’ or ‘disagreed’) perceived expert consensus to be higher than the actual rate. Conspiracy believers, compared to non-believers, overestimated the expert consensus by a larger margin (see Figure 3). The same pattern was found in participants’ estimations of laypeople consensus (average agreement among all participants of the study) (see Figure 4).
Figure 3 Conspiracy believers’ and non-believers’ estimation of expert consensus and the actual rate of expert consensus for conspiracy theories on the mining reserves in Turkey.
Figure 4 Conspiracy believers’ and non-believers’ estimation of laypeople consensus and the actual rate of laypeople consensus for conspiracy theories on the mining reserves in Turkey.
Beliefs in all three conspiracy theories regarding the allegedly underused mining reserves of Turkey positively correlated with perceived consensus among laypeople and experts in mining, geology, and geophysics engineering. Participants’ own beliefs in these conspiracy theories correlated relatively more strongly with the perceived expert consensus, compared to perceived laypeople consensus (see Table 2).
Table 2 Correlations between belief in conspiracy theories on the mining reserves in Turkey and perceived consensus
Note: ‘Difference’ test refers to the z-test for the difference between two dependent correlations with one variable in common (Lee and Preacher, Reference Lee and Preacher2013).
Interestingly, irrespective of the specific content of the conspiracy belief, a positive correlation was identified between participants’ personal beliefs and their perceptions of consensus among both experts and fellow participants (see Figures 5 and 6). The magnitudes of these correlations were consistent across all comparisons, indicating that certain individuals tend to perceive an elevated consensus on various conspiracy theories, a tendency that aligns with their own convictions. The correlation between belief in a conspiracy theory and the perception of consensus seems to be independent of the content. This finding suggests that some individuals exhibit a generalized propensity to both believe in a diverse array of conspiracy theories and perceive a higher degree of consensus regarding those theories among the populace. Furthermore, the correlations between perceived consensus among experts and perceived consensus among laypeople were observed to be of similar magnitudes across all comparisons (see Figure 7).
Figure 5 The correlations between participants’ own belief and their perceptions of expert consensus on different conspiracy theories.
Figure 6 The correlations between participants’ own belief and their perceptions of laypeople consensus on different conspiracy theories.
Figure 7 The correlations between perceptions of expert and laypeople consensus on different conspiracy theories.
Next, we examined the factors associated with conspiracy beliefs. The belief in conspiracy theories related to the Lausanne Treaty was found to be higher among older individuals, females (as compared to males), those with lower education levels, and individuals with lower perceived socioeconomic status (see Figure 8). Consistent with previous research on conspiracy beliefs, Lausanne conspiracy beliefs displayed positive correlations with right-wing ideology, religiosity, right-wing authoritarianism, and a preference for intuitive thinking. Conversely, they exhibited negative correlations with reflective thinking and science literacy. Contrary to our expectations, individuals with higher beliefs in these conspiracy theories demonstrated lower levels of social dominance orientation and higher preference for effortful thinking. However, their level of belief did not show any correlation with the Dark Triad personality traits. The pattern of correlations was identical for the belief in conspiracy theories regarding mining reserves (see Figure 8).
Figure 8 The correlations between main variables of interest.
Note: ‘Experts’ refers to participant’s perception of expert consensus, while ‘laypeople’ refers to perceived laypeople consensus. ‘Self’ refers to participant’s own belief in the conspiracy theory in question. ‘SES’ stands for socioeconomic status, ‘RWA’ right-wing authoritarianism, ‘SDO’ social dominance orientation, ‘CRT’ cognitive reflection test. Participants who chose ‘other’ for their sex or education were excluded in calculating the correlations. Higher score in sex corresponds to being male, as opposed to female. A full correlation table with correlation coefficients is available in the Supplementary material.
Perceived consensus among laypeople regarding conspiracy beliefs related to the Lausanne Treaty exhibited positive correlations with female gender, higher levels of religiosity, right-wing ideology, right-wing authoritarianism, and a preference for intuitive thinking (see Figure 8). Similarly, perceived consensus among laypeople regarding conspiracy beliefs surrounding unused mining reserves showed positive correlations with female gender, higher levels of religiosity, right-wing authoritarianism, and a preference for intuitive thinking, while displaying negative correlations with science literacy and psychopathy.
Notably, perceived consensus among experts on conspiracy beliefs related to the Lausanne Treaty exhibited the exact same pattern of correlations as participants’ own beliefs in those conspiracy theories. It demonstrated correlations with all predictors in the same direction. Likewise, perceived expert consensus on conspiracy beliefs surrounding unused mining reserves displayed the same pattern of correlations as participants’ own beliefs, with the exception of a slightly stronger correlation with psychopathy (see Figure 8).
We also have checked whether the order of materials had an effect on the reported levels of belief in conspiracy theories. Completing the items on perceived expert and laypeople consensus before or after the items on the participant’s own belief did not have an effect on the participants’ own beliefs regarding Lausanne-related, t(727) = -.564, p = .573, d = -.042, or mining reserves-related conspiracy theories t(727) = -.703, p = .482, d = -.052. Results suggest that there was no priming effect and thinking about the level of agreement among other people had no effect on conspiracy beliefs.
5. Discussion
Our findings indicate that participants exhibited a tendency to overestimate the consensus among experts and laypeople regarding two sets of conspiracy theories: The Lausanne Treaty and mining reserves in Turkey. In other words, individuals believed that these theories were more widely accepted both in the public sphere and within academia than they actually were. Furthermore, this inclination to overestimate consensus was more pronounced among conspiracy theory believers as compared to non-believers in both sets of conspiracy theories.
Secondly, we observed a positive correlation between belief in these conspiracy theories and perceived consensus among both laypeople and experts in the respective fields. Notably, this association demonstrated a relatively stronger link with perceived expert consensus as compared to laypeople consensus. To the best of our knowledge, this study represents the first attempt to compare perceived laypeople and expert consensus in their predictive value for conspiracy beliefs. The findings highlight the relative influence of perceived expert consensus, indicating that individuals’ perception of what individuals with relevant knowledge believe is more influential in shaping conspiracy beliefs. These results hold significance, particularly considering that the media often fail to accurately portray the level of consensus among experts, often adopting a ‘false balance’ approach by presenting opposing views on an issue. This false balance undermines the perception of consensus in expert opinion (Koehler, Reference Koehler2016) and creates the erroneous impression that experts are divided on the matter (Dixon and Clarke, Reference Dixon and Clarke2013).
Interestingly, we also found that perceived consensus on and belief in conspiracy theories correlate even when they are about different conspiracies. Thus, the results suggest that some people who endorse conspiracy theories tend to believe in a variety of different conspiracies at the same time and also have a higher perception of expert and laypeople consensus on these conspiracy theories. These results are consistent with an established finding in the conspiracy research, which suggests that all conspiracy beliefs correlate with each other (Williams et al., Reference Williams, Marques, Hill, Kerr and Ling2022). Future research should explore the role of a generalized tendency to harbor an inflated perception of consensus regarding one’s personal conspiracy beliefs. Such inquiries could elucidate the significance of a broader perception of heightened consensus, as opposed to perceived consensus on a particular issue, as underscored in existing literature (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019).
We found no priming effect when considering participants’ reported conspiracy beliefs in relation to thinking about the consensus of experts and laypeople. Thus, this study contributes to the existing research on perceived consensus in several meaningful ways. Firstly, it investigates the impact of perceived consensus on specific conspiracy theories, which provide concrete examples of the irrational beliefs observed in everyday life, rather than more general conspiracy mindsets that have been predominantly studied thus far (Sutton and Douglas, Reference Sutton and Douglas2020). Secondly, it explores whether the effect of perceived consensus, along with the theoretical framework of the GBM (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019), generalizes to a non-WEIRD context, specifically Turkey, and conspiracy beliefs specific to that context. Thirdly, the study incorporates a comprehensive set of variables to test the correlates of conspiracy beliefs and perceived consensus on these beliefs, accounting for variations in individuals’ worldview, cognitive sophistication, personality traits, and sociodemographic characteristics. Moreover, this study examines both perceived consensus among experts and laypeople, facilitating a comparison of their respective impacts on individuals’ belief in conspiracy theories. By randomizing the order of self-reported beliefs and expectations about others’ agreement with the conspiracy theories, the study also investigates whether considering others’ perspectives has an immediate causal influence on belief in conspiracy theories.
Furthermore, our findings contribute to two important theoretical aspects. Firstly, they support Albarracín’s (Reference Albarracín2022) proposed influence of normative plausibility on conspiracy beliefs, highlighting that the extent to which individuals perceive others to believe in a conspiracy plays a significant role in their perception of the conspiracy’s plausibility. Secondly, in light of the GBM (van der Linden, Reference van der Linden2021; van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019), our findings suggest that the predictions of the GBM extend to more specific conspiracy beliefs in non-WEIRD contexts. Our robust results indicate that higher perceived consensus among both experts and laypeople is associated with a greater tendency to believe in a specific conspiracy theory.
While the findings of the present study provide valuable insights into the association between perceived consensus and conspiracy beliefs, it is important to acknowledge that the correlational nature of the design prevents us from establishing causal relationships. One potential explanation for this association is the argument that a higher level of perceived consensus on conspiracy theories causes higher belief in those conspiracy theories. This aligns with the findings from the growing body of research on the GBM, which suggests that perceiving higher expert consensus on issues such as climate change can reduce belief in misinformation related to those issues (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019). However, it is also plausible that causality operates bidirectionally, as belief in conspiracy theories could influence individuals’ perceptions of consensus regarding those theories. The false consensus effect, a well-established cognitive bias, leads individuals to overestimate the prevalence of their own beliefs among others (Ross et al., Reference Ross, Greene and House1977). Even when one’s beliefs are not mainstream at the time, individuals tend to believe that their opinions align with the future consensus (Rogers et al., Reference Rogers, Moore and Norton2017). Thus, it is conceivable that higher levels of conspiracy beliefs may lead to higher perceptions of consensus on those beliefs. Another potential explanation is the presence of common predictors that may act as confounders in the relationship between conspiracy beliefs and perceived consensus. Preliminary evidence from our study (see Figure 5) suggests that certain predictors, including cognitive sophistication factors such as science literacy and cognitive reflection, as well as factors related to conservatism such as religiosity and right-wing ideology, predict both conspiracy beliefs and perceived consensus in the same direction. Notably, perceived expert consensus exhibited an identical pattern of correlations with all predictors, mirroring the associations observed with conspiracy beliefs. This points to a significant overlap in how both constructs are related to these relevant variables. It is possible that certain underlying factors influence both conspiracy beliefs and perceived consensus, resulting in strong correlations between the two. Future research should further investigate these potential explanations and ascertain which direction(s) of causality are empirically supported.
6. Limitations & future directions
In addition to its contributions, the current study suggests several potential directions for future research. While we did test for potential order effects and found that thinking about perceived consensus does not have an immediate impact on belief in specific conspiracy theories, the study did not include direct manipulations of perceived consensus among experts and laypeople. An intriguing avenue for further investigation would involve examining how individuals react to feedback on actual consensus after they have reported their perceived consensus. Prior studies have indicated that providing feedback on the actual consensus within one’s group regarding certain conspiracy theories can lead individuals to revise their beliefs in accordance with the group (Cookson et al., Reference Cookson, Jolley, Dempsey and Povey2021a, Reference Cookson, Jolley, Dempsey and Povey2021b). Thus, given our finding that individuals, particularly conspiracy theory believers, tend to overestimate the consensus among both experts and laypeople concerning specific conspiracy theories, it is important to assess whether individuals adjust their judgments after being presented with the actual consensus among groups.
Furthermore, the present study did not incorporate measurements of affective judgments, such as levels of worry or emotional responses. As affective judgments are argued to have a greater influence on driving public action compared to cognitive judgments (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019), it would be valuable for future research to examine the influence of perceived consensus on affective judgments regarding specific conspiracy theories. This would provide insights into the role of affective factors in shaping public responses and actions related to conspiracy beliefs.
Relatedly, an important consideration for future research, particularly in non-WEIRD contexts, is whether perceived consensus not only predicts belief in specific conspiracy theories but also influences behavioral intentions, as posited by the GBM (van der Linden et al., Reference van der Linden, Leiserowitz, Feinberg and Maibach2015b, Reference van der Linden, Maibach and Leiserowitz2019). The current evidence addresses the initial assumptions of the GBM, specifically the influence of perceived consensus on individuals’ beliefs, but lacks insight into the subsequent impact of perceived consensus on behavior and behavioral intentions. Therefore, future studies should aim to test the complete model by incorporating behavioral measures and examining whether all assumptions of the model conceptually replicate in non-WEIRD contexts.
Furthermore, this study focuses on assessing belief in specific conspiracy theories that are specific to the studied country. It is plausible that the influence of perceived consensus on specific conspiracy theories may vary across different countries. However, according to the principle of normative plausibility and the GBM, the underlying mechanism should remain consistent. Future research should investigate potential country-level variations or determine whether the same phenomena hold true globally.
In conclusion, utilizing a non-WEIRD sample and employing a comprehensive set of correlates, our study demonstrates that perceived consensus plays a significant role in driving belief in specific conspiracy theories. We also highlight the alarming tendency for individuals to greatly overestimate the consensus among experts and laypeople, even in conspiracy theories that experts deem groundless. Notably, conspiracy theory believers exhibit a higher degree of overestimation compared to non-believers. Consequently, these findings indicate that perceived consensus holds promise as a potential avenue for addressing and combating conspiracy beliefs.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/jdm.2023.33.
Data Availability Statement
The raw data can be accessed at https://osf.io/mfuyz.
Acknowledgments
The current study is based on the data of an ongoing longitudinal project funded by The Scientific and Technological Research Council of Turkey (TÜBİTAK; project number: 221K046; project title: ‘The Temporal Changes in Epistemically Suspect (Conspiracy, Paranormal and Pseudoscience) Beliefs: Longitudinal Analysis of the Predictors and the Moderators’). We used ChatGPT (chat.openai.com) for proofreading purposes. We asked ChatGPT to proofread and improve the language when necessary in our texts. ChatGPT’s role was to revise existing bodies of text and it did not author any part of the article by itself. We thank Onurcan Yilmaz, Serkan Dolma, Eylul Deran Atalay, Aysenur Duzgun, Nazif Karaca, Tugcenaz Elcil, and Fatih Bayrak for their work and assistance in our ongoing longitudinal research.
Competing interest
The authors declare none.
Open access
