MISSMeas Instrum Soc SciMeasurement Instruments for the Social SciencesMeas. Instrum. Soc. Sci.2523-8930PsychOpenmiss.1985910.5964/miss.19859New Measurement InstrumentsDataCodeMaterialsOn ‘Doing Conspiracy Theory’. Introduction and Validation of the Conspiracy Practice Scale (CPS-6)On ‘Doing Conspiracy TheoryOn ‘Doing conspiracy theory’. Introduction and validation of the Conspiracy Practice Scale (CPS-6)https://orcid.org/0000-0001-8325-5874KnasmüllerFlorian*1https://orcid.org/0009-0008-6788-4310HenzeDustin1https://orcid.org/0000-0001-7864-2271BrunnerMarkus1https://orcid.org/0000-0002-6246-1692NachtweyOliver2https://orcid.org/0009-0001-0936-640XSpeckSarah34https://orcid.org/0000-0003-1428-9514KumkarNils C.5DavidovEldadFaculty of Psychology, Sigmund Freud University, Vienna, AustriaDepartement Gesellschaftswissenschaften, University of Basel, Basel, SwitzerlandInstitute for Social Research, Johann Wolfgang Goethe University, Frankfurt, GermanyKulturwissenschaftliche Fakultät, Europa-Universität Viadrina, Frankfurt (Oder), GermanySOCIUM Research Center on Inequality and Social Policy, University of Bremen, Bremen, GermanyUniversity of Cologne, Cologne, GermanyFreudplatz 1, 1020, Wien, Österreich. florian.knasmueller@sfu.ac.at2007202620268e1985928012026180520262026Knasmüller, Henze, Brunner et al.https://creativecommons.org/licenses/by/4.0/This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International License, CC BY 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Conspiracy theories have recently been met with a lot of public scrutiny. They are generally addressed as a form of belief system that is underpinned by psychological dispositions. In contrast, we introduce the Conspiracy Practice Scale (CPS-6) as a novel instrument focussing on the spread and consumption of conspiracy theories as a form of social practice. In a random split of a dataset collected in Austria, Germany, and Switzerland (N = 6,460), exploratory factor analysis was employed to determine dimensionality of the CPS-6 and discriminate it against conspiracy mentality (nEFA = 3,229). CPS-6 emerged as distinct single dimension with good overall reliability (α = .86, ω = .87). Results were confirmed using confirmatory factor analysis (CFA) on the second split (nCFA = 3,231). Multi-group CFA revealed measurement invariance across countries and convergent validity was indicated by correlations with related constructs. Overall, results show CPS-6 to be a reliable and valid short-scale to measure conspiracy practice as a distinct construct.
We are aware of the controversy surrounding the term conspiracy theories. However, we decided to use it nonetheless, because it is the most common term in this field of research (see e.g., Butter, 2020). We speak of conspiracy theories only when addressing the specific content of conspiracy beliefs. In the following, we will delineate our understanding of conspiracy practice as a distinct feature of conspiracy theorizing.
has spiked especially in the aftermath of the Covid-19 pandemic, rendering conspiracy beliefs more visible than ever before and spurring a sharpened public awareness of their impact on public discourse and political deliberation. Researchers in the social sciences have mainly dealt with conspiracy beliefs with regard to their epistemology and psychological underpinnings (see Cassam, 2023). While the former approach is concerned with the question if and under which circumstances conspiracy theories contain adequate explanations of the events in question, the latter aims at determining psychological proneness to and correlates of conspiracy beliefs (see Douglas & Sutton, 2023). There is a growing conviction among researchers in the field that a general susceptibility to conspiracy beliefs must be distinguished from beliefs in specific conspiracy theories (Imhoff et al., 2022; Strömbäck et al., 2024). However, both are understood in terms of their psychological predispositions to believe in such theories.
Strikingly, latest research showed that in absolute terms, conspiracy beliefs have not increased in recent decades (Uscinski et al., 2022), thus mirroring the notion that conspiracy theories are rather common and stable forms of social sense-making, albeit increasingly stigmatized ones (see Thalmann, 2019). The striking dissonance between the surge in public scrutiny of conspiracy beliefs and their apparent stagnation across time begs theoretical as well as empirical questions. If the shift in public discourse cannot be plausibly explained through a dramatic surge of belief in conspiracy theories, and we don’t want to dismiss it merely as an unfounded moral panic, explanations must lie beyond the scope of conspiracy beliefs (Kumkar et al., 2026). Drawing on research that brackets the role of the individual conviction of the factual accuracy of specific conspiracy theories (e.g., de Zeeuw & Gekker, 2023; Koper, 2025; Muirhead & Rosenblum, 2019) and that focusses on what conspiracy theorists actually do, such as organizing rallies and engaging in (online) discussions (e.g., Harambam, 2020; Kumkar, 2023), we proposed the theoretical concept of ‘doing conspiracy theory’ or ‘conspiracy practice’ (Kumkar et al., 2026). Assuming that partaking in the production and communication of conspiracy theories is, even where it involves acting on a preexisting belief, not reducible to it, we can conclude that it is unlikely that people who believe in conspiracy theories neatly correspond to those who communicate them. While, for example, belief in conspiracy theories on the origins, existence, and dangers of Covid-19 was widespread (Fotakis & Simou, 2023), the number of people that propagated those beliefs, for instance in the form of public anti-lockdown protests, was significantly smaller. Conversely, while conspiracy beliefs were generally highly prevalent within the protest crowds, not all participants indicated belief in specific Covid-19 related conspiracy theories (see e.g., Forschungswerkstatt Corona-Proteste, 2021; Lamberty et al., 2022; Nachtwey et al., 2020).
Thus, we propose ‘doing conspiracy theory’ as a distinct dimension of contemporary “conspiracy culture” (Knight, 2000) that has been largely neglected to date and hold that it allows for a broader understanding of this timely matter, without dismissing conspiracy beliefs as an important factor. In the following, we introduce the conspiracy practice scale (CPS-6) as a simple tool to measure ‘doing conspiracy theory’ using just six items. We validate it in a sample comprised of data from Austria, Germany and Switzerland. Accounting for and measuring conspiracy practice could help to further our understanding of conspiracy beliefs by filling existing blind spots. CPS-6 is designed as a complementary tool alongside established scales focusing on susceptibility to conspiracy beliefs such as the “Conspiracy Mentality Questionnaire” (Bruder et al., 2013) and “Generic Conspiracist Beliefs Scale” (Brotherton et al., 2013). We hope that this exploratory approach taking into account conspiracy practice will prove valuable for advancing the scholarship of conspiracy theories.
Conspiracy Practice Scale (CPS-6)
In order to measure ‘doing conspiracy theory’ (Kumkar et al., 2026), i.e. conspiracy practice, we developed a brief measurement instrument comprised of six items. Item development was guided both by theoretical considerations and experience from field work as well as interviews with people involved in conspiracy milieus. The authors thus drew on experience from years of extensive research in the field of conspiracy theories (Amlinger & Nachtwey, 2021; Forschungswerkstatt Corona-Proteste, 2021; Kumkar, 2018). The scale broadly consists of three facets of conspiracy practice, i.e. 1) protest participation, 2) online activism in so called ‘alternative media’, and 3) convincing people of hidden truths. However, we assumed a single factor construct that is conspiracy practice. Items were designed to either directly address how practices relate to notions of ongoing conspiracies or used indirect cues that indicate mistrust or opposition to ‘mainstream’ accounts, that are key features of conspiracy theories (The original items are in German—see Table 1 in the Appendix, translation: the authors):
I have participated in protests that were publicly portrayed as conspiracy-minded.
I research alternative media to stay up to date on what is really happening in the world.
I discuss on social media what is being concealed in the mainstream media.
I try to educate people that they cannot trust public opinion.
In discussions I am often referred to as a conspiracy theorist.
I try to rouse people with protests so that they finally realize that they are being deceived.
Participants were asked to indicate their approval through a Likert-scale ranging from (1) do not agree at all to (6) completely agree. A middle response option was deliberately omitted to reduce neutral responding and encourage respondents to express a directional evaluation. However, participants were permitted to abstain from answering any given item.
Objectives of the Present Research
In our study, we aimed at developing and validating a scale that allows us to effectively and economically measure ‘doing conspiracy theory’ as outlined in the introduction. CPS-6 addresses a specific set of behaviors (online and offline) and allows for gauging the relationship between conspiracy beliefs and practices. While it is entirely plausible that in some cases conspiracy practices follow specific beliefs, in other cases people might engage in conspiracy practice independent of belief in specific theories.
Currently, several different instruments are available to measure proneness to conspiracy beliefs, such as the “Conspiracy Mentality Questionnaire (CMQ)” (Bruder et al., 2013) and “Generic Conspiracist Beliefs Scale” (Brotherton et al., 2013). Notwithstanding that these scales differ in their approach to measuring conspiracy beliefs and in their underlying assumptions on how they come about, they align in their treatment of the issue as something individuals do or do not believe in. Other instruments are designed to assess protest participation (Kelloway et al., 2007), alternative media use (Prochazka & Schweiger, 2019) or susceptibility to fake news (Maertens et al., 2024), thus each touching on individual facets of CPS-6. However, our instrument is a comprehensive scale, covering different facets at once, while still maintaining its efficiency.
In our exploratory study, we introduce CPS-6 to the armory of the studies in conspiracy theories, by assessing the additional value of considering conspiracy practice as a feature of contemporary conspiracy culture. Thus, we developed a scale that covers different facets of “conspiracy practice”. The objectives of the research are the following:
Validating the internal structure of the theoretically derived instrument, testing the assumption that CPS-6 is a unidimensional factor, and discriminating CPS-6 against “Conspiracy Mentality”.
Examining measurement invariance across the three German speaking countries (Austria, Germany, Switzerland).
Testing convergent validity through correlations with the adjacent dimensions of belief in specific conspiracy theories, authoritarianism, and right-wing extremism. Although we assume that CPS-6 acts as a distinct dimension, we still predict that it is positively correlated to said attitudes.
Materials and MethodsParticipants and Data Collection
Data collection was conducted via the online panel provider Bilendi/Respondi in all three countries of the DACH-region, i.e. Austria (AT), Germany (DE) and Switzerland (CH) between October 1st, 2024, and November 15th, 2024. Only the German-speaking population in each country was targeted. We used SoSci Survey and employed survey quota to ensure that the sample resembles the populations of the respective countries. Quota were assigned for gender, age, region and education (low/medium/high). Due to recruitment difficulties not all quotas could be met (see Table 2 in the Appendix for sample characteristics).
A minimum age of 18 years was required to take part in the survey. Only participants that provided their informed consent were eligible for participation. The survey included a sociodemographic section before presenting participants with items regarding political attitudes such as conspiracy mentality, authoritarianism, and right-wing extremism.
To account for limitations of data collection via online panels, two types of quality checks were implemented. First, three attention checks were incorporated. All participants that failed to give the intended answer were automatically excluded from further participation. Second, a check ensured that participants who finished the survey 50% faster than the median completion time of the pilot study were also automatically excluded from the dataset. Third, after data collection was completed, we excluded 52 participants with a relative speed index (RSI) greater than 2 (Leiner, 2019). Fourth, 1513 participants with missing values in items of the CPS-6 or the CMQ were excluded.2
To identify possible patterns underlying missing values, we conducted a non-responder Analysis for each factor used in the following analysis, drawing on one item (with the highest item discrimination) per factor and using chi-squared tests (see Table 8 in the Appendix). In general, female respondents as well as those with lower education and social status are overrepresented amongst non-responders. Non-response thus mirrors findings from comparable studies on political attitudes (see Koch & Blohm, 2009). Additionally, we checked for item-level pattens of missingness in the CPS-6 (see Table 9 in the Appendix). Item 4 and 6 showed a higher number of missing values than the other four items but overall, the number of missing responses in the CPS-6 items was no higher than in the established constructs. We therefore conclude that there is no pattern of missing values that warrants interpretation.
The final sample of each country consisted of n = 2,125 (AT), n = 2,164 (GER) and n = 2,171 (CH).
Measures
In addition to the CPS-6, four other instruments were incorporated to determine convergent validity and its role within the broader matrix of research on conspiracy mentality, belief in specific conspiracy theories, authoritarianism, CMQ, and right-wing extremism. The relationship between these psychological properties has been extensively researched. While conspiracy mentality is defined as an element of authoritarian dispositions (Schließler et al., 2020), i.e. the psychological disposition to succumb to fascist ideologies, research found that conspiracy beliefs predict right-wing extremism (see van Prooijen et al., 2015). CPS-6 serves as an additional dimension to this theoretical and empirical complex, focusing on ‘doing conspiracy theory’ as a form of social practice. We thus expect CPS-6 to be positively correlated with all four instruments, while still acting as an independent construct.
Conspiracy Mentality Questionnaire (CMQ)
CMQ (Bruder et al., 2013) is one of the most commonly used measurements to determine susceptibility to conspiracy theories. It comprises five items (e.g., “I think that many very important things happen in the world, which the public is never informed about”). We adapted the original rating scale to a six-point Likert-scale without a middle category. In our sample, internal consistency was good, α = .879 and ω = .882.
Belief in Specific Conspiracy Theories
We asked participants to indicate belief in six specific conspiracy beliefs using one item each with a six-point Likert response scale regarding 1) Climate change denial, 2) Conspiracy theory on the origins of the Ukraine-Russia war, 3) Deep state theory, 4) Covid-19 related conspiracy, 5) Great Replacement Theory and 6) “Homo-Lobby” theory (see Table 3 in the Appendix for item wording). We established a mean index of belief in conspiracy theories. Higher scores indicate greater overall belief in specific conspiracy theories.
Authoritarianism (KSA-3)
We drew on the KSA-3 scale by Beierlein et al. (2015) to determine authoritarian dispositions. The scale follows the tradition of the F-scale developed by Adorno et al. (1950) and its further development by Altemeyer (1981). It measures susceptibility to fascist ideologies rather than overt approval of them. It is comprised of three subdimensions with three items each: authoritarian aggression (e.g., “Social rules should be enforced without compassion”; α = .74 and ω = .74), authoritarian submission (e.g., “We need strong leaders so that we can live safely in society”; α = .69 and ω = .69), and conventionalism (e.g., “Proven practices should not be questioned”; α = .75 and ω = .76), but is also applicable as a global factor across the subdimensions (α = .84 and ω = .84).
Right-Wing Extremism (FR – LF)
We used a shortened version of the “Questionnaire on right-wing extremist attitudes in the Leipzig form” (FR – LF) (Decker et al., 2013), that consisted of six items in total (e.g., “When jobs become scarce, foreigners should be sent back to their home countries”). The original version of the scale comprises six dimensions (i.e. “Approval of authoritarian dictatorships”, “Chauvinism”, “Xenophobia”, “Antisemitism”, “Social Darwinism” and “Trivialization of National Socialism”) with three items each. When selecting items, we followed the suggestions of a group of experts (see Kreis, 2007); each dimension is represented by one item. Tests on internal consistency indicate good reliability in our sample (α = .83 and ω = .83).
Statistical Analysis
First, we examined the distributional properties of the CPS-6 items and assessed the suitability of the sample for exploratory factor analysis by using the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and conducting Bartlett’s test of sphericity on the full dataset. In order to obtain two independent samples for exploratory and confirmatory analysis, we split the complete dataset. To ensure similar sub-sample-sizes for each country in both halves, we randomly split all three sub samples and reassigned each of these country-specific halves to either one of the two final samples. Sample 1 finally consisted of nEFA = 3,229 and Sample 2 of nCFA = 3,231 cases.
We then conducted an exploratory factor analysis in Sample 1 to test whether the items of the CPS-6 can be differentiated from the existing construct of conspiracy mentality and form an independent unidimensional factor. For this purpose, both the newly developed items of the CPS-6 and the established CMQ were included in the analysis.
Second, to further explore the dimensionality and construct validity of the CPS-6, we conducted three confirmatory factor analyses (CFAs) in Sample 2. The first CFA included a model with only the newly developed CPS-6. The second CFA included a model with the two scales (CPS-6 and CMQ) as separate factors. The third CFA included a model with CPS-6 and CMQ Items as a single factor. Following recent recommendations not to rely solely on Cronbach’s α (Hayes & Coutts, 2020), we determined Cronbach’s α and McDonald’s ω and item-total correlations for each item to assess internal consistency of CPS-6 in Sample 2. According to common recommendations, we considered a α and ω > .70 as acceptable and > .80 as good (Cortina, 1993).
To examine measurement invariance across the three countries, we conducted a multi-group confirmatory factor analysis in Sample 2 by comparing increasingly restricted, nested models (configural, metric, scalar, strict and mean invariance, see Leitgöb et al., 2023).
Finally, we tested convergent validity through correlations with the adjacent dimensions of conspiracy mentality, belief in specific conspiracy theories, authoritarianism, and right-wing extremism in the complete dataset.
All statistical analysis was performed with R (Version 4.4.1) with the graphical user interface RStudio (Version 2024.12.1:Build 563) using several packages from the Comprehensive R Archive Network (See Table 4 in the Appendix for a list of the used packages).
ResultsItem-Level Analysis
According to common cut-offs of < 2 in skewness and kurtosis (Pituch & Stevens, 2016), item 1 (“I have participated in protests that were publicly portrayed as conspiracy-minded“) and item 5 (“In discussions I am often referred to as a conspiracy theorist”) of the CPS-6 showed a problematic distribution in the full dataset (see Table 1 for detailed distributional properties, descriptive statistics, skewness and kurtosis).
Descriptive Statistics of the CPS-6 Items in the Complete Dataset
Item
n
M
SD
Min
Max
Skew
Kurtosis
I have participated in protests that were publicly portrayed as conspiracy-minded.
6,460
1.48
1.12
1
6
2.65
6.43
I research alternative media to stay up to date on what is really happening in the world.
6,460
2.77
1.63
1
6
0.47
-0.97
I discuss on social media what is being concealed in the mainstream media.
6,460
1.99
1.38
1
6
1.36
0.90
I try to educate people that they cannot trust public opinion.
6,460
2.44
1.45
1
6
0.74
-0.43
In discussions I am often referred to as a conspiracy theorist.
6,460
1.77
1.26
1
6
1.72
2.20
I try to rouse people with protests so that they finally realize that they are being deceived.
6,460
2.05
1.31
1
6
1.20
0.66
Note. M = Mean; SD = Standard Deviation.
A Shapiro-Wilk normality test indicated deviation of normal distribution for each of the six items. The Mardia test for multivariate normality also revealed significant deviations, Skewness = 15828.813, p < .001; Kurtosis = 125.986, p < .001. All CPS-6 items showed pronounced floor effects, with item means ranging from 1.48 to 2.77. However, this distribution was theoretically expected, as only a small proportion of the general population is likely to endorse such practices. Accordingly, all subsequent exploratory and confirmatory factor analyses were estimated using robust methods, with robust maximum likelihood estimation (MLR) applied to account for non-normal item distributions and floor effects.
Dimensionality
Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity tests confirmed suitability of the data for EFA. The Items exhibited sufficient shared variance with an overall Measure of Sampling Adequacy (MSA) of .93—which exceeds the recommended threshold of .60 (Field, 2017). Bartlett’s test was significant with χ2(55) = 18130.76, p < .001, conforming adequate homogeneity of variances (Watson, 2017).
As theoretically expected, parallel analysis suggested that the number of factors is 2. Following the Kaiser Guttman criterion, the results of the EFA confirmed a two-factor solution, Eigenvalue > 3 in both factors, with the items clearly loading on two distinct dimensions corresponding to the CPS-6 and the CMQ. Each factor accounted for approximately 28% of the total variance. The pattern of factor loadings is presented in Table 5 in the Appendix.
Reliability
The results of the reliability tests of the CPS-6 in Sample 2 showed very good internal consistency with α = .86 and ω = .87. Corrected total-item correlations ranged from .59 to .70 (see Table 6 in the Appendix), indicating that all items contributed meaningfully to the overall scale. The scale showed strong internal homogeneity with an average inter-item correlation of r = .518. Although this relatively high average could suggest potential item redundancy, reliability did not improve if any individual item was removed. Overall, the scale demonstrated strong internal consistency across all three countries, with reliability estimates remaining consistently high in Austria (Cronbach’s α = .85, McDonald’s ω = .87), Germany (α = .88, ω = .89), and Switzerland (α = .85, ω = .85).
Confirmatory Factor Analysis
Confirmatory factor analyses (CFAs) were estimated using the robust maximum likelihood estimator (MLR). Model fit was evaluated using commonly recommended cut-off criteria (Hu & Bentler, 1999), with Comparative Fit Index (CFI) and Tucker–Lewis Index (TLI) values ≥ .95, Standardized Root Mean Square Residual (SRMR) ≤ .08, and Root Mean Square Error of Approximation (RMSEA) ≤ .06 indicating good model fit. All reported chi-square statistics and fit indices are based on robust (scaled) estimates.
The CFA of the first model including only the CPS-6 items showed good fit to the data, χ2(9) = 114.693, p < .001, CFI = .977, TLI = .961, RMSEA = .080, SRMR = .027. As expected, given the large sample size, the chi-square test was significant. Standardized factor loadings (λ) ranged from .651 to .758 and were all statistically significant (p < .001), indicating that the items adequately represent the latent construct of conspiracy practice. These results provide initial evidence for the structural validity of the CPS-6.
The CFA of the second model, specifying CPS-6 and CMQ as two correlated latent factors, demonstrated acceptable model fit, χ2(43) = 762.900, p < .001, CFI = .950, TLI = .936, RMSEA = .082, SRMR = .051. A significant covariance between the two latent factors was observed (b = .055, SE = .028, p < .001), corresponding to a standardized correlation of r = .743, indicating a strong positive relationship between the constructs.
In contrast, a third CFA specifying CPS-6 and CMQ items as indicators of a single common factor demonstrated poor model fit, χ2(44) = 2232.925, p < .001, CFI = .842, TLI = .803, RMSEA = .143. A chi-square difference test comparing the one-factor model to the two-factor model indicated that the two-factor model fit the data significantly better with a Δχ2(1) = 661.65, p < .001, ΔCFI = .108 and ΔRMSEA = .061, exceeding the recommended cut-off criteria proposed by Chen (2007) of ΔCFI ≥ .010 and ΔRMSEA ≥ .015. This indicates that the items do not adequately reflect a single underlying construct. Taken together, these results suggest that CPS-6 and CMQ represent two distinct but related latent constructs rather than a single common factor.
To test the robustness of the results in view of the numerous cases excluded due to missing values, we also conducted a CFA of Model 1 and 2 in the complete dataset—including all cases with missing values in either CPS-6 or CMQ—using Full Information Maximum Likelihood Method (FIML). The analysis showed only small differences compared to the aforementioned results. The results can be found in Table 7 in the appendix. These findings indicate that the reported CFA results are robust with respect to the handling of missing data, as substantively equivalent model fit and parameter estimates were obtained when missing values were addressed using full information maximum likelihood estimation.
Multi Group Confirmatory Factor Analysis
To assess measurement invariance across the three countries of the DACH-region, multi-group confirmatory factor analysis with a model consisting only of CPS-6 as a factor using the robust maximum likelihood estimator (MLR) was used with increasingly restrictive models: first, configural invariance was analyzed to ensure adequate fit in all three samples. Second, loadings were restricted to be equal across groups (metric invariance). Third, intercepts were restricted to be equal across groups as well (scalar invariance) and deterioration of model fit was observed for each step. Fourth, item residuals were restricted to be equal across groups (strict invariance).
Although the scaled chi-squared difference tests were significant at all levels of measurement invariance, the changes in fit indices did not exceed the beforementioned recommended cut-off criteria proposed by Chen (2007). Thus, metric, scalar, and strict measurement invariance can be assumed across all three countries.
After establishing measurement invariance, latent mean comparisons were conducted by constraining factor means across groups. Constraining latent means did not meaningfully reduce model fit, indicating no substantial evidence for cross-national latent mean differences. These results indicate that the CPS-6 allows for meaningful cross-national comparisons and that levels of conspiracy practice are comparable across the three German-speaking countries. An overview of model fits can be found in Table 2.
Tests for Measurement Invariance and Latent Mean Comparisons
Invariance Model
χ2(df)
Δχ2(Δdf)
p
CFI
ΔCFI
RMSEA
ΔRMSEA
Configural
128.760 (27)
.000
0.977
0.078
Metric
151.903 (37)
23.143 (10)
.000
0.974
0.003
0.070
0.008
Scalar
183.619 (47)
31.716 (10)
.000
0.974
0.000
0.064
0.006
Strict
217.611 (59)
33.992 (12)
.000
0.967
0.007
0.064
0.000
Latent Mean Model
189.046 (49)
.000
0.973
0.063
Note. df = degrees of freedom; CFI = comparative-fit-index; RMSEA = Root Mean Square Error of Approximation. The latent mean model is based on scalar invariance and is not part of the hierarchical measurement invariance sequence.
Convergent Validity
As shown in Table 3, Spearman’s rank-order correlations indicated significant positive associations between the CPS-6 and all related constructs, thus confirming our initial assumption. The strongest correlations were observed regarding belief in specific conspiracy theories, conspiracy mentality, and right-wing extremism. Moderate correlations were found with authoritarianism, and two of its subdimensions: aggression and conventionalism. In contrast, only a weak correlation was observed for the submission subdimension (all indicators can be found in Table 3). All correlations were statistically significant, p < .001. Confidence intervals (CI) were derived via bootstrapping with 5,000 resamples. Overall, our theoretical expectations were met by our empirical model: While it is conceptually related to adjacent constructs, the scale measures a construct distinct from classical authoritarian attitudes, conspiracy mentality and beliefs in specific conspiracy theories.
Convergent Validity of CPS-6 Using Spearman’s Rank Correlation
Variable
rₛ
CI
n
Conspiracy mentality
0.65
[0.64, 0.67]
6,460
Belief in specific conspiracy theories
0.73
[0.72, 0.74]
5,426
Authoritarianism
0.38
[0.35, 0.40]
5,781
Authoritarian aggression
0.31
[0.29, 0.34]
6,066
Authoritarian submission
0.25
[0.22, 0.27]
6,202
Conventionalism
0.34
[0.31, 0.36]
6,241
Right-wing extremism
0.61
[0.59, 0.63]
5,540
Note. rₛ = Spearman's rank correlation coefficient; CI = 95% bootstrap confidence interval; n = sample size.
Discussion and Limitations
The present study introduced and validated the novel Conspiracy Practice Scale (CPS-6) as a brief instrument to assess ‘doing conspiracy theory‘. We understand CPS-6 as an additional, insightful tool in the study of conspiracy theories, focusing on what people do rather than what they believe in.
Our findings provide promising evidence for the reliability, structural validity, and applicability of the scale across the three German-speaking countries. First, exploratory factor analysis supported the expected one-factor structure of the CPS-6, with all items demonstrating substantial and statistically significant factor loadings. Notably, CPS-6 emerged as a distinct factor that is empirically separable from the Conspiracy Mentality Questionnaire (CMQ), indicating that the instrument captures participation in conspiracy practices rather than susceptibility to conspiracy beliefs. This provides strong support for the structural validity of the scale and initial evidence for its construct validity. Second, CPS-6 exhibited good internal consistency across all three countries. Moreover, metric, scalar, and strict measurement invariance across countries was supported, and latent mean comparisons indicated comparable levels across groups. These findings suggest that the CPS-6 allows for meaningful cross-national comparisons and that similar levels of conspiracy practice can be observed across the three German-speaking countries. Third, convergent validity was demonstrated by high correlations with conspiracy mentality, belief in specific conspiracy theories, and right-wing extremism. Additionally, moderate associations were observed with authoritarianism and its subdimensions, except for authoritarian submission, which showed only a weak correlation. Overall, these results indicate that conspiracy practice is conceptually as well as empirically related to adjacent constructs, while still constituting an independent factor.
Despite these strengths, the following limitations must be acknowledged. All CPS-6 items show a floor effect, which may reflect the generally low prevalence of participation in conspiracy practice in the general population. Although this distribution was to be expected in the population, the effects of social desirability must also be considered a contributing factor. What participants actually do does not necessarily coincide with practices they are willing to admit to in an online survey. While this is the case for all inquiries of socially divisive topics, items addressing actions rather than attitudes or beliefs might be affected differently. Thus, further research must put more emphasis on discerning conspiracy practice from mere claims of ‘doing conspiracy theory’. Furthermore, potential self-selection effects must be factored in, especially since conspiracy beliefs are associated with mistrust in science (see Steger & Schütz, 2025).
Moreover, the present study focused exclusively on the CMQ as a comparative measure; future research should include a broader range of theoretically related constructs such as the Generic Conspiracist Beliefs Scale” (Brotherton et al., 2013) and scales on political activism or civic engagement to test convergent validity more thoroughly. This would also allow us to further empirically explore the general gap between attitudes and behavior often assumed in social psychology (e.g., Fishbein & Ajzen, 1975), while also relating it to the question at hand: i.e. the relationship between doing and beliefs in conspiracy theories.
Additionally, the current version of the CPS-6 does not allow for differentiation between various forms of conspiracy practices, such as online versus offline practices. Future work should aim to further refine the scale by developing subdimensions that capture the heterogeneity of ‘doing conspiracy theory’. Lastly, CPS-6 should be used to map the nexus of conspiracy mentality, specific conspiracy beliefs and practices in future research, e.g., through Structural Equation Modeling (SEM) that allow for examining the direction of the relationships between the variables.
In summary, the CPS-6 offers a concise and sound instrument to assess conspiracy practices. It constitutes an important step toward understanding the relationship between conspiracy mentality, belief and practices and provides a valuable tool for future research on the social and political consequences of conspiracy theories. However, our model should be replicated in other samples to corroborate its validity and applicability.
Ethics Statement
Ethical approval for this study series was obtained from the Ethics Commission of the University of Basel (Reference number: 159)
This research was funded in whole or in part by the Austrian Science Fund (FWF) [10.55776/I6577], the German Research Foundation [522282012] and the Swiss National Science Foundation [10001AE_216039]. For open access purposes, the authors have applied a CC BY public copyright license to any author accepted manuscript version arising from this submission.
Acknowledgements
We would like to thank Dr. Ayline Heller for her helpful comments on this manuscript.
ReferencesAdorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., and Sanford, R. N. (1950). The authoritarian personality. Harper & Brothers.Altemeyer, B. (1981). Right-wing authoritarianism. University of Manitoba Press.Beierlein, C., Asbrock, F., Kauff, M., & Schmidt, P. (2015). Kurzskala Autoritarismus (KSA 3). Zusammenstellung sozialwissenschaftlicher Items und Skalen (ZIS). 10.6102/zis228Amlinger, C., & Nachtwey, O. (2021). Sozialer Wandel, Sozialcharakter und Verschwörungsdenken in der Spätmoderne.Aus Politik und Zeitgeschichte, 71(35-36), 13–19.Brotherton, R., French, C. C., & Pickering, A. D. (2013). Measuring belief in conspiracy theories: The Generic Conspiracist Beliefs scale.Frontiers in Psychology, 4, 279. 10.3389/fpsyg.2013.0027923734136Bruder, M., Haffke, P., Neave, N., Nouripanah, N., & Imhoff, R. (2013). Measuring individual differences in generic beliefs in conspiracy theories across cultures: Conspiracy Mentality Questionnaire.Frontiers in Psychology, 4, 225. 10.3389/fpsyg.2013.0022523641227Butter, M. (2020). The nature of conspiracy theories. Polity.Cassam, Q. (2023). Conspiracy theories.Society, 60, 190–199. 10.1007/s12115-023-00816-136776684Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance.Structural Equation Modeling, 14(3), 464–504. 10.1080/10705510701301834Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and applications.The Journal of Applied Psychology, 78(1), 98–104. 10.1037/0021-9010.78.1.98Decker, O., Hinz, A., Geißler, N., & Brähler, E. (2013). Fragebogen zur rechtsextremen Einstellung – Leipziger Form (FR-LF). In O. Decker, J. Kiess, & E. Brähler (Hrsg.), Rechtsextremismus der Mitte. Eine sozialpsychologische Gegenwartsdiagnose (pp. 197–212). Psychosozial.de Zeeuw, D., & Gekker, A. (2023). A god-tier LARP? QAnon as conspiracy fictioning.Social Media + Society,9(1). 10.1177/20563051231157300Douglas, K. M., & Sutton, R. (2023). What are conspiracy theories? A definitional approach to their correlates, consequences, and communication.Annual Review of Psychology, 74(1), 271–298. 10.1146/annurev-psych-032420-03132936170672Field, A. (2017). Discovering statistics using IBM SPSS statistics (5th ed.). Sage.Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Longman Higher Education.Forschungswerkstatt Corona-Proteste. (2021). Corona-Protest-Report. Narrative – Motive – Einstellungen. SocArXiv. 10.31235/osf.io/25qb3Fotakis, E. A., & Simou, E. (2023). Belief in COVID-19 related conspiracy theories around the globe: A systematic review. Health Policy, 137, 104903. 10.1016/j.healthpol.2023.10490337688953Harambam, J. (2020). Contemporary conspiracy culture: Truth and knowledge in an era of epistemic instability. Routledge.Hayes, A. F., & Coutts, J. J. (2020). Use omega rather than Cronbach’s alpha for estimating reliability. But….Communication Methods and Measures, 14(1), 1–24. 10.1080/19312458.2020.1718629Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives.Structural Equation Modeling, 6(1), 1–55. 10.1080/10705519909540118Imhoff, R., Bertlich, T., & Frenken, M. (2022). Tearing apart the “evil” twins: A general conspiracy mentality is not the same as specific conspiracy beliefs.Current Opinion in Psychology, 46, 101349. 10.1016/j.copsyc.2022.10134935537265Kelloway, E. K., Francis, L., Catano, V. M., & Teed, M. (2007). Predicting protest.Basic and Applied Social Psychology, 29(1), 13–22. 10.1080/01973530701330884Knight, P. (2000). Conspiracy culture: From the Kennedy assassination to the X-Files. Routledge.Koch, A., & Blohm, M. (2009). Item nonresponse in the European Social Survey.Research & Methods, 18(1), 45–65.Koper, A. J. (2025). More than a theory: Conspiracy theorising as a practical tradition.Political Studies Review, 23(4), 994–1010. 10.1177/14789299241292573Kumkar, N. C., Speck, S., Brunner, M., Knasmüller, F., Kreienbaum, S. & Nachtwey, O. (2026). Doing conspiracy theory: Reconstructing the social production of a specific form of ‘social critique’.Journal of Political Sociology, 4(1), 23–41. 10.54195/jps.19471Kumkar, N. C. (2023). Collaborative mistrust: The communicative function of alternative facts in social media interactions.Ethnologia Europaea, 53(2), 1–21.Kumkar, N. C. (2018). The Tea Party, Occupy Wall Street, and the Great Recession, Critical Political Theory and Radical Practice. Palgrave Macmillan. 10.1007/978-3-319-73688-4Kreis, J. (2007). Zur Messung von rechtsextremer Einstellung: Probleme und Kontroversen am Beispiel zweier Studien. Arbeitshefte aus dem Otto-Stammer-Zentrum Nr. 12. https://refubium.fu-berlin.de/bitstream/handle/fub188/19527/ahosz12.pdf?sequence=1&isAllowed=yLamberty, P., Holnburger, J., & Goedeke Tort, M. (2022, 17. Februar). CeMAS-Studie: Das Protestpotential während der COVID-19-Pandemie. CeMAS. https://cemas.io/blog/protestpotential/Leiner, D. J. (2019). Too Fast, too Straight, too Weird: Non-Reactive Indicators for Meaningless Data in Internet Surveys.Survey Research Methods, 13(3), 229–248. 10.18148/srm/2019.v13i3.7403Leitgöb, H., Seddig, D., Asparouhov, T., Behr, D., Davidov, E., De Roover, K., Jak, S., Meitinger, K., Menold, N., Muthén, B., Rudnev, M., Schmidt, P., & van de Schoot, R. (2023). Measurement invariance in the social sciences: Historical development, methodological challenges, state of the art, and future perspectives.Social Science Research, 110, 102805. 10.1016/j.ssresearch.2022.10280536796989Maertens, R., Götz, F. M., Golino, H. F., Roozenbeek, J., Schneider, C. R., Kyrychenko, Y., Kerr, J. R., Stieger, S., McClanahan, W. P., Drabot, K., He, J., & van der Linden, S. (2024). The Misinformation Susceptibility Test (MIST): A psychometrically validated measure of news veracity discernment.Behavior Research Methods, 56, 1863–1899. 10.3758/s13428-023-02124-237382812Muirhead, R., & Rosenblum, N. L. (2019). A lot of people are saying: The new conspiracism and the assault on democracy. Princeton University Press. 10.2307/j.ctv941trnNachtwey, O., Schäfer, R., & Frei, N. (2020). Politische Soziologie der Corona-Proteste. SocArXiv. 10.31235/osf.io/zyp3fPituch, K. A., & Stevens, J. P. (2016). Applied Multivariate Statistics for the Social Sciences: Analyses with SAS and IBM’s SPSS. RoutledgeProchazka, F., & Schweiger, W. (2019). How to measure generalized trust in news media? An adaptation and test of scales.Communication Methods and Measures,13(1), 26–42. 10.1080/19312458.2018.1506021Schließler, C., Hellweg, N., & Decker, O. (2020). Aberglaube, Esoterik und Verschwörungsmentalität in Zeiten der Pandemie. In E. Brähler & O. Decker (Eds.), Autoritäre Dynamiken: Alte Ressentiments – neue Radikalität (pp. 283–308). Psychosozial-Verlag.Steger, D., & Schütz, A. (2025). Assessing trust in science: Development and validation of a short scale for adolescents and adults.Measurement Instruments for the Social Sciences, 7, e16225. 10.5964/miss.16225Strömbäck, J., Broda, E., Tsfati, Y., Kossowska, M., & Vliegenthart, R. (2024). Disentangling the relationship between conspiracy mindset versus beliefs in specific conspiracy theories.Zeitschrift für Psychologie, 232(1), 18–25. 10.1027/2151-2604/a000546Thalmann, C. (2019). The stigmatization of conspiracy theory since the 1950s: "A plot to make us look foolish". Routledge.Uscinski, J., Enders, A., Klofstad, C., Seelig, M., Drochon, H., Premaratne, K., & Murthi, M. (2022). Have beliefs in conspiracy theories increased over time?PLoS One, 17(7), e0270429. 10.1371/journal.pone.027042935857743van Prooijen, J.-W., Krouwel, A. P. M., & Pollet, T. V. (2015). Political extremism predicts belief in conspiracy theories.Social Psychological & Personality Science, 6(5), 570–578. 10.1177/1948550614567356Watson, J. C. (2017). Establishing evidence for internal structure using exploratory factor analysis.Measurement & Evaluation in Counseling & Development, 50, 232–238. 10.1080/07481756.2017.1336931Data Availability
For this article, data is freely available (see Knasmüller et al., 2026a).
Supplementary Materials
For this article, the following Supplementary Materials are available:
the original anonymized dataset including all relevant variables as well as a codebook to facilitate navigation of the dataset (see Knasmüller et al., 2026a),
additional tables referenced in the article that provide futher information (see Knasmüller et al., 2026b), and
the R scripts of all statistical analyses required to replicate our findings (see Knasmüller et al., 2026c).
KnasmüllerF.HenzeD.BrunnerM.NachtweyO.SpeckS.KumkarN. C. (2026a). Dataset for: On ‘Doing conspiracy theory’. Introduction and validation of the Conspiracy Practice Scale (CPS-6)[Dataset]. PsychArchives. 10.23668/psycharchives.22310KnasmüllerF.HenzeD.BrunnerM.NachtweyO.SpeckS.KumkarN. C. (2026b). Tables for: On ‘Doing conspiracy theory’. Introduction and validation of the Conspiracy Practice Scale (CPS-6)[Tables]. PsychArchives. 10.23668/psycharchives.22312KnasmüllerF.HenzeD.BrunnerM.NachtweyO.SpeckS.KumkarN. C. (2026c). Syntax for: On ‘Doing conspiracy theory’. Introduction and validation of the Conspiracy Practice Scale (CPS-6)[Code]. PsychArchives. 10.23668/psycharchives.22311
The authors have declared that no competing interests exist.