Edited by: Igor Grossmann, University of Waterloo, Canada
Reviewed by: Paul Conway, Florida State University, USA; Ulrich Von Hecker, Cardiff University, UK
*Correspondence: Gábor Orosz
This article was submitted to Personality and Social Psychology, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Conspiracy theory (CT) beliefs can be harmful. How is it possible to reduce them effectively? Three reduction strategies were tested in an online experiment using general and well-known CT beliefs on a comprehensive randomly assigned Hungarian sample (
The level of beliefs in conspiracy narratives differs among people. Conspiracy narratives claim that powerful people or organizations cooperate in secret, to achieve sullen objectives by deceiving the public (Abalakina-Paap and Stephan,
Despite theoretical diversity (see Krekó,
According to previous studies, individual differences appear regarding beliefs in CTs. These individual differences can be categorized into personality and attitudinal dimensions. Among personality variables, belief in CTs is negatively related to self-esteem (Abalakina-Paap and Stephan,
If conspiracy theories are associated with malevolent effects, the question arises: How is it possible to change conspiracy beliefs? According to Fishbein and Ajzen (
Based on Fishbein and Ajzen's conceptualization (Fishbein and Ajzen,
(a) In order to change the link between the object and the attribute, further logical pieces of information or logical steps can be provided, thus allowing us to elaborate on the logical structure which can result in a more complex link. For example, the object of the belief can be the banks, and the attribute can be exploitative. A low level of elaboration lacks arguments as to why this link exists, whereas a high level of elaboration includes many logically compatible and underpinned arguments. We hypothesize that pointing out the logical flaws of this link, providing an elaborated explanation without logical flaws could make one deconstruct the preexisting link and construct a new one between the object and the attribute. This should lead to belief changes. In the previous example, if we provide detailed information about how much money banks get from taxpayers' money, which is four times less than what they pay to the state in many different special taxes in any given period of time, the link becomes more elaborate in terms of complexity and coherence. We assume that this can lead, to some extent, to belief change. From the perspective of functional attitude theories (Katz,
(b) The second possibility of conspiracy belief change involves increasing the distance between the self and those who hold a certain link between the object and the attribute. To achieve this, one can demonstrate that those people who hold such beliefs are characterized by negative traits or they are targeted as being ridiculous. As practically no one wants to be ridiculed by others, the ridiculing argument can be fueled by the ego-protective function (Katz,
(c) We suppose that the third form of conspiracy belief change relates to the identification with the object of the belief. Therefore, in this case, the primary goal is not to change the link between the object and the attribute, but to focus on the reduction of the distance between the self and the object of the CT. For example, CTs concerning objects (e.g., Jews) generally include negative attributes (e.g., secretly manipulative). If, as a result of a message (e.g., claiming that Christians faced similar conspiracy theories beforehand), people put ourselves in the position of Jews, they can empathize more easily with them, which can lead to more positive attitudes toward them. These empathizing arguments can be related to value expressive functions in terms of presenting an image that is in line with a positive and caring self-concept (Katz,
The goal of the present study was to experimentally examine the above-described three strategies of CT belief change in an online setting on comprehensive samples. We expected that all three strategies would effectively reduce CT beliefs.
This research employed a nationally representative probability sample of 813 Hungarians, selected randomly from an Internet-enabled panel, including 20,000 members, with the help of the Solid Data Ltd., in October–November 2014. For the preparation of the sample, a multiple-step, proportionally stratified, probabilistic sampling method was employed. Members of this panel used the Internet at least once a week. The demographic characteristics of the panel are permanently filtered. More specifically, individuals were removed from the panel if they responded too quickly (i.e., without paying attention to their response) and/or had fake (or not used) e-mail addresses. The sample was nationally representative in terms of gender, age, level of education, and location of residence. The study was conducted in accordance with the Declaration of Helsinki and with the approval of the Institutional Review Board of the Eötvös Loránd University. Before starting the questionnaire, participants received information about the study in terms of examining beliefs about how societies work. Subsequently, participants read and approved the informed consent.
Eight hundred and thirteen (813) individuals (50.8% female) participated in the present online experiment, between the ages of 18 and 75 (
In the present study, participants were recruited via Internet. After agreeing with the informed consent form, participants listened to the first audio recording (for the transcript, see Appendix 1 in Supplementary Material). This is a 4:30 min recording that presented a conspiracy super theory including the victimization of Hungary by the financial imperium, the hidden control of Jews over the world, the EU as a non-functional oppressive power, and the bankers who exploit the Hungarian financial system. The text provided vivid, but confusing details about the mechanisms that “actually” shape the fate of Hungary and the world. This super CT met the above mentioned characteristics of CTs in terms of nothing happens by chance, nothing is what it seems, everything is interconnected with everything, and the world is divided into good and evil.
Having listened to the recording, participants expressed their acceptance concerning eight questions on the four main topics (victimization of Hungary, EU, power of the Jews, bankers). Then, they were asked about their general acceptance of the listened CT. The final two questions referred to the perceived competence and intelligence of the speaker who reported the CT. After this questionnaire, participants were randomly assigned to one of the four conditions. They listened to another speech with, either rational (3:36 min), ridiculing (3:28 min), or empathetic (2:54 min) arguments against CTs. In the control condition (3:15 min), they listened to a weather forecast. The transcripts of each condition can be seen in Appendix 2 in Supplementary Material.
In the rational condition, the text tackled the claims made in the first recording in a logically plausible manner, using numbers to support the objections, and pointing out the discrepancy between high influence and concealment. This speech pointed out the logical flaws of the first speech and corrected it with in-depth arguments regarding the link between the beliefs' objects and attributes. The goal of this condition was to emphasize the logical inconsistencies and to create a more complex and coherent relationship between the objects of the belief and the attributes.
In the ridiculing condition, the script addressed the same logical flaws, but reasoned against them differently: instead of focusing on certain details, it derided the logical inconsistencies and concentrated on those who believe in the CTs, picturing them as evidently ridiculous (e.g., mentioning the believers of Lizard Men). This text intended to increase the distance between the respondents' self and those who believe in CTs.
The empathetic condition contested the original text's claim in a different manner: instead of focusing on content or those who believe in the content, it placed the objects of the CTs in the center, and compassionately called attention to the dangers of demonizing and scapegoating, while also pointing out the human character of the CT objects (i.e., Jews face similar conspiracy theories and persecution nowadays that the Early Christians faced). This condition intended to reduce the distance between the respondent and the objects of CTs and to raise empathy toward these groups.
In the control condition participants listened to a simple weather forecast which was not related to the content of the CT. After the recording, respondents answered the same 11 questions.
An eight item scale was created specifically for the present experiment to measure the individual's attitudes toward conspiracies, assessing beliefs regarding conspiracies related to four aspects: (1) Hungary as a victim of conspiracy; (2) Jews as the leaders of the world; (3) the European Union is a parasitic formation without any function; and (4) the bankers as the leaders of the world. These specific topics were related to the CT audio they listened to previously. More precisely, they referred to the key elements of the text regarding these four topics. Respondents used an 11-point scale to indicate their level of agreement (0, Strongly Disagree; 10, Strongly Agree). This scale showed high levels of reliability with a Cronbach alpha value of 0.95 for the pre-manipulation and 0.96 for the post-manipulation.
In addition to the eight CAT items, one item measured the listener's level of agreement with the audio excerpts, using the response options of the previous eight items. Finally, two additional items were created: the first assessed the competence of the speaker on an 11-point scale (0, Not competent at all; 10, Completely competent), while the second item assessed the intelligence of the speaker on a 5-point scale (1, Far below average; 5, Far above average).
This 5-item measurement was developed by Bruder et al. (
The 40-item BIDR (Paulhus,
Personality-related dimensions were assessed by the Big Five thInventory (John and Srivastava,
Participants were asked to indicate the following demographic variables: age, gender (1, female; 2, male), place of residence (1, capital; 2, county town; 3, town; 4, village), and level of education (1, primary level education; 2, secondary level education; 3, higher level education).
Descriptive statistics of the above-mentioned questionnaires with ranges, summed scores, standard deviations, and intercorrelations can be seen in Table
1. Conspiracy Assessment Tool—pre | 8–88 | 45.15 | 25.88 | – | ||||||||
2. Conspiracy Assessment Tool—post | 8–88 | 43.28 | 26.11 | 0.94 |
– | |||||||
3. Conspiracy Mentality Questionnaire | 11–77 | 51.08 | 11.21 | 0.60 |
0.61 |
– | ||||||
4. Self-deception Enhancement factor of BIDR | 20–140 | 92.13 | 12.46 | −0.09 |
−0.07 | 0.02 | – | |||||
5. Impression Management factor of BIDR | 20–140 | 82.36 | 11.44 | −0.06 | −0.04 | −0.01 | 0.29 |
– | ||||
6. Extraversion | 3–15 | 10.62 | 2.67 | −0.01 | 0.00 | 0.04 | 0.20 |
0.04 | – | |||
7. Agreeableness | 3–15 | 11.42 | 2.13 | 0.07 | 0.08 |
0.11 |
0.20 |
0.20 |
0.25 |
– | ||
8. Conscientiousness | 3–15 | 10.34 | 2.53 | −0.01 | −0.03 | 0.04 | 0.33 |
0.34 |
0.09 |
0.14 |
– | |
9. Emotional stability | 3–15 | 9.93 | 2.73 | −0.05 | −0.04 | 0.01 | 0.43 |
0.12 |
0.11 |
0.32 |
0.14 |
– |
10. Openness | 3–15 | 11.35 | 2.92 | −0.13 |
−0.13 |
−0.05 | 0.17 |
0.16 |
0.16 |
0.22 |
0.11 |
0.11 |
No significant differences were found between the baseline (pre-test) measures considering the four conditions (all
Paired-samples
These results suggest that the manipulation in the rational and the ridiculing condition were successful. However, in the empathetic condition the extent of believing in conspiracy theories did not significantly decrease.
The CONDITION * TIME ANCOVA controlling for BIDR-SE predicting the change in the extent of believing in conspiracy theories controlling for the self-deceptive enhancement factor of social desirability revealed significant main effect of TIME,
The CONDITION * TIME ANCOVA controlling for BIDR-IM predicting the change in the extent of believing in conspiracy theories controlling for the impression management factor of social desirability revealed significant main effect of TIME,
The CONDITION * TIME ANCOVA controlling for COMPETENCY predicting the change in the extent of believing in conspiracy theories controlling for the perceived competency of the speaker revealed only marginally significant main effect of TIME,
The CONDITION * TIME ANCOVA controlling for INTELLIGENCE predicting the change in the extent of believing in conspiracy theories controlling for the perceived intelligence of the speaker revealed no significant main effect of TIME,
Using Big Five traits as covariates in the CONDITION * TIME ANCOVAs did not show either the main effect or the interaction of personality traits, except for Openness which had a significant main effect,
The CONDITION * TIME ANCOVA controlling for CMQ predicting the change in the extent of believing in conspiracy theories controlling for the level of engaging in conspiracy-related ideations revealed significant main effects of TIME,
Most of the existing studies regarding the nature of conspiracy theories are descriptive; moreover, experimental research exploring the possibility of CT belief change is very rare (Banas and Miller,
Previously, it was assumed that CTs resisted rational arguments, because only those pieces of information that were consistent with them were processed, while the incompatible ones were rejected (Bartlett and Miller,
In the ridiculing condition, the arguments focused on the deficiencies of CT believers' thinking. This belief change can be expected as a result of the emerging distance between the self and those who hold CT beliefs. Ridiculing arguments are threatening for those who hold strong conspiracy beliefs, but among those who do not have very strong CT beliefs it can lead to disidentification from the group of CT believers. The stronger the belief in conspiracy, the stronger cognitive dissonance can be expected, and the results can be apparent in attitudes and behavior. We only examined beliefs, and in this field, cognitive dissonance can be resolved in two main ways: CT beliefs can be reduced and the new information regarding CT believers can be rejected. Considering the full sample, this strategy reduced CT beliefs. However, when we examined the effect of this condition among those who held strong vs. relatively weak CT beliefs separately, we found no significant difference between the effects of rational arguments: it was effective in both groups. Further, examination is needed in order to identify why ridiculing strategy did not have an overwhelming effect.
Among the experimental manipulations, the empathetic condition was the least effective. Previous studies found that perspective taking can effectively reduce CT beliefs (van Prooijen and van Dijk,
On the basis of these results, two belief changing strategies appeared to be effective: the rational and the ridiculing ones. From the perspective of the different theoretical backgrounds, there are alternative explanations for these findings. On the basis of the functional attitude theories (Katz,
Thanks to the rich literature regarding the links between individual differences in terms of personality (Abalakina-Paap and Stephan,
The present study is not without limitations. The effect sizes were not large. However, measuring the effectiveness of different reasoning or convincing strategies is not easy. In the present study, the number of arguments was balanced, but the length of the audio recordings was different in the different conditions. Further, studies should balance the number of arguments, their length and pretest the effectiveness of each argument. Needless to say that it is a time consuming task. If we consider the present study as an intervention, it can first be said that this is not a wise one, as direct and confronting strategies were used to convince individuals regarding CT reduction. Second, this experiment did not have the very solid theoretical background that a good intervention requires. Third, this study only measured the short-term effects of different CT reduction strategies. Fourth, it targeted a general population instead of a specific subgroup of individuals. Fifth, the timing of the experiment was not related to a big CT-related scandal, which could have influenced the effectiveness of the conditions.
Besides these numerous deficiencies, the present study shows that rational arguments can reduce CT beliefs, while ridiculing also appears to be somewhat effective. Future studies are needed in order to explore the boundaries of these results. But, after careful investigation of these conditions (culture, timing, different groups with different characteristics, different speakers, etc.), media campaigns can be designed and in collaboration with competent public speakers, different CT reduction strategies can be tested.
Despite the extensive knowledge about the harmful effects of having CT beliefs, the reduction of CT beliefs with experimental methods is a relatively neglected topic of scientific investigation. In the present study, three convincing strategies were tested in order to reduce CT beliefs: rational arguments, ridiculing of CT believers, and expressing empathy toward the objects of CT beliefs. Providing rational arguments was found as being an effective strategy, along with providing ridiculing arguments, which could also reduce CT beliefs. Only very weak, or even non-significant links were found between conspiracy theory-related variables and individual differences. Considering these results and previous studies focusing on the benevolent effects of analytic thinking in CT belief reduction, it can be assumed that uncovering arguments regarding the logical inconsistencies of CT beliefs can be an effective way to discredit them. Our findings on the efficiency of rational argumentation go against the mainstream of the communication literature and “common wisdom,” as well as the current affective wave of social psychology emphasizing that emotions constitute the most important factor behind shaping beliefs and attitudes. Considering the modest effect sizes, we assume that rationality has a bigger impact on shaping (sometimes irrational) beliefs than previously expected, given that in the current communication environment, people are overloaded with emotional messages coming from ads, political and social campaigns. Future studies should also investigate the role of rationality and the “rationality heuristic” in belief change.
GO contributed to the study design, manuscript writing, and data analyses; PK contributed to the study design, data gathering and interpretation, and literature review; BP contributed to the literature review and interpretation of the results and writing the manuscript; IT and BB contributed to the data analysis and interpretation, writing the manuscript, and the literature review; CR contributed to the manuscript writing and literature review. All authors commented on the draft and contributed to the final version, approved the publication of the manuscript, and agreed to be accountable for all aspects of the work.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The first author (GO) was supported by the Hungarian Research Fund (NKFI PD 106027, 116686) and the Hungarian Academy of Sciences (Lendület Project LP2012-36).
The Supplementary Material for this article can be found online at:
1According to the theory of planned behavior, a behavior occurs more probably if (a) the individual has a positive attitude toward, (b) has a sense of being able to control the given behavior, and (c) the subjective norms support this particular behavior. This combination of attitudes, beliefs, and subjective norms leads to increased behavioral intention, which in turn results in higher probability of the actual performance.