Introduction
In the weeks after the start of the COVID-19 pandemic, we presented data in this journal that conspiracy beliefs may have become mainstream (Freeman, Waite et al., Reference Freeman, Waite, Rosebrock, Petit, Causier, East and Lambe2022). Subsequent evidence is broadly consistent with such a view (Fotakis & Simou, Reference Fotakis and Simou2023). Conspiracy beliefs are one form of excessive mistrust. A related, but distinct (Alsuhibani et al., Reference Alsuhibani, Shevlin, Freeman, Sheaves and Bentall2022; Martinez et al., Reference Martinez, Shevlin, Valiente, Hyland and Bentall2022), type of excessive mistrust is paranoia: thinking that others are targeting you when they are not (Freeman & Garety, Reference Freeman and Garety2000). Persecutory ideation is common in the general population (e.g. Bebbington et al., Reference Bebbington, McBride, Steel, Kuipers, Radovanoviĉ, Brugha and Freeman2013; Bird et al., Reference Bird, Evans, Waite, Loe and Freeman2019; Neidhart, Mohnke, Vogel, & Walter, Reference Neidhart, Mohnke, Vogel and Walter2024). A quarter of the UK population describes themselves as mistrustful of other people (Freeman & Loe, Reference Freeman and Loe2023). Another distinct form of mistrust may be jealousy in relationships, which may range too in severity from normal and understandable to excessive (e.g. Kupfer et al., Reference Kupfer, Sidari, Zietsch, Jern, Tybur and Wesseldijk2022; Lima et al., Reference Lima, Köhler, Stubbs, Quevedo, Hyphantis, Koyanagi and Carvalho2017). There are of course real dangers that mean mistrust is not always misplaced; it can be absolutely justified and right to be wary. There is however an appropriate balance to be made between trust and mistrust, that will vary by person, place, time, and circumstance. But excessive mistrust, when the fear has become clearly inaccurate and disproportionate, can lead to negative behavioral consequences, such as not taking up suitable medical procedures to the detriment of health (Bierwiaczonek, Gundersen, & Kunst, Reference Bierwiaczonek, Gundersen and Kunst2022; Coelho, Foster, Nedri, & Marques, Reference Coelho, Foster, Nedri and Marques2022; Freeman, Loe et al., Reference Freeman, Loe, Chadwick, Vaccari, Waite, Rosebrock and Lambe2022), the breakdown of relationships, and isolation. If excessive mistrust is widespread, then extreme examples should be noticed in the people around us. We report here the first survey on this prediction.
Paranoia is the form of excessive mistrust most commonly studied, understood, and treated within a mental health context (Freeman, Isham, & Waite, Reference Freeman, Isham and Waite2025). Its most common clinical presentation is persecutory delusions in the context of severe mental health conditions such as psychosis. Though it has a precise meaning in clinical settings, the term is used much more widely in the general population. Common uses highlighted by psychoeducational websites are: ‘Some people might describe a paranoid thought as a type of anxious thought’ (Mind, 2024). We were interested to find out the degree to which the general public’s definition of paranoia aligns with that of clinical researchers. Mental health literacy positively influences attitudes towards mental health help-seeking (Jung, von Sternberg, & Davis, Reference Jung, von Sternberg and Davis2017). Furthermore, it is plausible that how the general public perceives the behaviors of people with severe paranoia may impact outcomes. Negative responses may worsen the interactions of the mistrustful person and entrench the fears further; positive responses may increase the likelihood that trust can re-emerge and the paranoia lessen. These responses are likely to be partly determined by levels of knowledge about paranoia. Understanding in which areas public knowledge about paranoia is limited could inform future psychoeducational interventions to the benefit of patients and carers.
Paranoia has increasingly become a topic of study in its own right, with advances in understanding and treatment. Paranoia typically develops in adolescence or young adulthood (Bird, Freeman, & Waite, Reference Bird, Freeman and Waite2022), and decreases over time (Della Libera et al., Reference Della Libera, Larøi, Raffard, Quertemont and Laloyaux2020; Greenburgh & Raihani, Reference Greenburgh and Raihani2022; Saarinen, Granö, & Lehtimäki, Reference Saarinen, Granö and Lehtimäki2021), with a slight increase in older age (Östling, Reference Östling2004). A range of social, cognitive, environmental, and biological processes have been linked to paranoia. Paranoia has been associated with anxiety, worry, and depression (Freeman et al., Reference Freeman, Stahl, McManus, Meltzer, Brugha, Wiles and Bebbington2012), as well as to social stress (Veling et al., Reference Veling, Pot-Kolder, Counotte, van Os and van der Gaag2016). In a recent survey of the adult general population, two-thirds of the variance in paranoia was explained by a model containing a range of cognitive and social processes such as within-situation defense behaviors, negative images, negative self-beliefs, discrimination, dissociation, aberrant salience, agoraphobic distress, decreased social support, agoraphobic avoidance, decreased analytical reasoning, and alcohol use (Freeman & Loe, Reference Freeman and Loe2023). Heritability of paranoia has been estimated at 50% (Zavos et al., Reference Zavos, Freeman, Haworth, McGuire, Plomin, Cardno and Ronald2014), and a range of genes may be associated with paranoia (Crespi, Read, Salminen, & Hurd, Reference Crespi, Read, Salminen and Hurd2018; Sieradzka et al., Reference Sieradzka, Power, Freeman, Cardno, Dudbridge and Ronald2015). Taken together, the evidence indicates that the cause of paranoia is multi-factorial (Freeman et al., Reference Freeman, Isham and Waite2025). For severe clinical presentations, new psychological treatments have been developed from this perspective that have shown high efficacy (Freeman et al., Reference Freeman, Emsley, Diamond, Collett, Bold, Chadwick and Waite2021a), while antipsychotic medications can also clearly reduce paranoia (Leucht et al., Reference Leucht, Leucht, Huhn, Chaimani, Mavridis, Helfer and Davis2017).
The overall aim of the present study was to describe for the first time the experiences and understanding of the general public regarding excessive mistrust occurring in other people, with a particular focus on paranoia. We set out to answer three specific questions: Do people observe others being excessively mistrustful? What do people mean when they use the term paranoia? How does the general public understand the particular form of mistrust that is paranoia (inaccurate fears that others intend harm)?
Methods
Participants
An online survey was conducted with a quota-sampled UK participant group of 1,128 adults (aged 18 years old or over), representative of the overall UK population for age, gender, ethnicity, income, and region. Participant recruitment was conducted from the 20th of May 2024 to the 29th of May 2024 via Cint, a market research company. Respondents were recruited via a variety of methods: ads and promotions on digital networks, word of mouth, social networks, affiliate marketing, banner ads, online and mobile games, and offline recruitment with mail campaigns. All participants recruited through Cint had opted in to being panel members for the market research company. The study was approved by the Medical Sciences Interdivisional Research Ethics Committee at the University of Oxford (reference R93164/RE001). The survey questions were not accessible until informed consent was given. The survey was completed online via Qualtrics, either on a computer or mobile browser, and took on average 24 minutes to complete.
Sixteen people were excluded for completing the survey in less than one-third of the median completion time; 36 people were excluded for providing answers that were not serious (e.g. nonsense words) to the open-ended questions; six people were excluded for selecting the same answer for every item on at least three different sections; and 34 people were excluded for both non-serious responses to the open-ended questions and selecting the same answer repeatedly across sections. This resulted in a final sample of 1,036 people.
Assessments
Demographics
In the first section of the survey participants indicated their age, gender, level of education, ethnicity, socioeconomic status, marital status, employment status, and religious or spiritual affiliation. These factors have been associated with community attitudes towards individuals with mental illness (Angermeyer & Matschinger, Reference Angermeyer and Matschinger1997; Bhugra, Reference Bhugra1989; Lam, Chan, & Chen, Reference Lam, Chan and Chen1996).
Experiences of excessive mistrust in other people
In the second section of the survey participants reported any experiences of other people being excessively mistrustful. For nine different examples of extreme mistrust (e.g. ‘Saying that the COVID-19 virus is not real and that the pandemic was a hoax,” ‘Thinking that others are targeting them in order to bully or exploit them, and so isolating from the world and refusing to leave their home.’), they were asked whether they had interacted with anyone in the past year who had shown such behavior. They could respond: No one, One or two people, Quite a few people, or Most people. If they had noticed such behavior in others they were then asked about the nature of their relationship with the mistrustful person (e.g., friend, acquaintance, partner).
Understanding paranoia
In the third section of the survey participants were asked about their conceptions and understanding of paranoia. They were first asked an open-ended question: ‘How would you define paranoia, i.e., what do you think it is?’. Participants gave a free text response for their definition of the word paranoia. Participants were then given the following definition based on Freeman and Garety’s (Reference Freeman and Garety2000) definition of persecutory delusions:
We define paranoia as having an enduring belief that other people (or organisations) are deliberately trying to harm you when they are not. It is excessive mistrust of other people. People may think others are trying to harm them socially, physically, or financially when they are not.
Based on this definition participants were asked if they knew anyone who experiences paranoid thoughts (answer choices: ‘No, I do not know anyone who has these thoughts’; ‘Yes, I know someone who previously had these thoughts’; ‘Yes, I know someone who currently has these thoughts’), and, if so, what their relationship was to that person (‘Partner’; ‘Someone at work’; ‘Close friend’; ‘Friend’; ‘Family member’; ‘Acquaintance’; ‘Stranger’), and how they realized this person was experiencing paranoid thoughts (‘They told me’; ‘I noticed it in their behavior or something they said’; ‘Someone else pointed it out to me’). Participants were then asked to rate their level of understanding of paranoia (‘I do not understand it at all’; ‘I understand a little why people may get paranoid’; ‘I have some understanding of why people get paranoid,” ‘It is very understandable why people get paranoid’).
Participants were asked about causes of paranoia. They were asked to indicate on a 5-point scale (Strongly disagree, Disagree, Neither agree nor disagree, Agree, Strongly agree) to what extent they believed each of 34 different factors play a role in causing paranoia (e.g., ‘chemical imbalance in the brain’, ‘Social isolation’). Participants were also asked to indicate what they believed to be the main cause of paranoia out of the items presented. The items were derived from Part C of the Trauma-Voice Association Questionnaire (van den Berg, Hardy, & Staring, Reference van den Berg, Hardy and Staring2015), a local NHS service intake questionnaire, Freeman’s (Reference Freeman2016) model of persecutory delusions, and from discussions within the authorship team (which included a lived experience researcher).
Participants were then asked questions about their ideas concerning the onset, timeline, and potential gender differences in paranoia.
Finally, participants were asked to indicate on a 5-point scale (Strongly disagree, Disagree, Neither agree nor disagree, Agree, Strongly agree) to what extent they believed each of 13 approaches could help someone with paranoia (e.g. ‘Getting more sleep’, ‘Seeing a therapist’).
Revised Green et al. paranoid thoughts scale (R-GPTS; Freeman, Loe et al., Reference Freeman, Loe, Kingdon, Startup, Molodynski, Rosebrock and Bird2021 )
The R-GPTS was used to measure the participants’ own levels of paranoia. This scale comprises eight items concerning ideas of reference (Part A; e.g. ‘I spent time thinking about friends gossiping about me’) and 10 items concerning ideas of persecution (Part B; e.g. ‘People have been hostile towards me on purpose’). Each item is rated for the past month on a 5-point scale (from 0 = not at all to 4 = totally). Higher scores on each scale indicate higher levels of paranoia.
Generic conspiracist beliefs (GCB) scale (Brotherton, French, & Pickering, Reference Brotherton, French and Pickering2013)
The GCB was used to assess participants’ own levels of conspiracy belief endorsement. This is a 15-item scale where each item is rated on a 5-point Likert scale (1 = definitely not true; 2 = probably not true; 3 = not sure/cannot decide; 4 = probably true; 5 = definitely true). Example items include: ‘The spread of certain viruses and/or diseases is the result of the deliberate, concealed efforts of some organisation’, or ‘A lot of important information is deliberately concealed from the public out of self-interest’. A higher score indicates greater levels of conspiracy-type assumptions and beliefs.
Analysis
Statistical analyses were conducted using SPSS version 29 (IBM Corp., 2023) and R version 4.2.1. (R Core Team, 2021). The analysis was largely descriptive in nature, showing frequency responses for items. Due to the ordinal nature of the data, polychoric correlation analyses were carried out between the nine mistrust examples. A summed total score was also calculated for the mistrust examples, which was then correlated with paranoia and conspiracy beliefs scores.
A content analysis was conducted on participants’ free text responses to the question: ‘How would you define paranoia, i.e., what do you think it is?’ Participants’ answers were coded by a researcher (GS). The following steps were undertaken in conducting the content analysis: familiarisation with the data, initial coding, grouping codes into themes, refining themes through discussion sessions with the research group, and quantifying the themes by counting how many participant answers fell into each theme. Each response was only assigned one code. A code hierarchy was built based on the closeness of the definition to our definition of paranoia. Coding a definition into the ‘Having ideas that others are trying to harm them’ code always took priority over other codes. Inter-rater reliability was assessed by selecting 100 definitions at random, which were independently coded by two raters, GS and MA (research assistants), using the list of codes and the code hierarchy. Cohen’s Kappa coefficient was calculated to quantify the inter-rater reliability of this coding.
An exploratory factor analysis (EFA) was conducted using R version 4.2.1. (R Core Team, 2021) to identify the underlying structure of causes for paranoia endorsed by the sample. A confirmatory factor analysis (CFA) was then conducted to confirm the structure identified by the EFA. The total sample (n = 1036) was first randomly split into two subsamples: a derivation sample (n = 725) and a confirmatory sample (n = 311). The EFA was conducted on the derivation sample (n = 725). Prior to EFA, we employed the Kaiser–Meyer–Olkin measure of sampling adequacy and Bartlett’s test for sphericity to assess the suitability of factor recovery using the observed dataset. Parallel analysis was used as the criterion for factor retention for subsequent rotation using the oblimin algorithm. Items with low communalities (<0.30), items that had low loadings on all factors (<0.30), and items that loaded onto multiple factors were excluded from the final model. Seven factors were retained. Subsequently, a CFA was conducted to test the seven-factor structure identified by the EFA. The maximum likelihood estimation method was applied to both the derivation sample (n = 725) and confirmatory sample (n = 311). Model fit was evaluated using a Comparative Fit Index (CFI) and Tucker–Lewis index (TLI) of >0.95, a Root Mean Square Error of Approximation (RMSEA) of <0.06, and a Standardized Root Mean Square Residual (SRMR) of <0.08 (Hu & Bentler, Reference Hu and Bentler1999).
Exploratory analyses were conducted to examine whether paranoia is perceived and understood differently across demographic groups (see Supplementary materials B). Perceived causes of paranoia were grouped according to the factor structure identified in the EFA, with factor scores computed by summing the scores of the items that constituted each factor. To compare factor scores across male and female genders, independent sample t-tests were conducted. Multivariate Analyses of Variance (MANOVA) were conducted to compare factor scores across age, ethnicity, and level of education. Following a significant MANOVA, one-way ANOVAs were conducted to identify where group differences occurred, followed by post-hoc comparisons. A full description of the analysis can be found in Supplementary materials B: Analysis of demographic factors.
Results
Participant group
A summary of the socio-demographic information of the final participant group (N = 1,036) can be seen in Table 1. The mean age of the participant group was 48.2 (SD = 16.8) years old. The mean R-GPTS Part A score was 7.3 (SD = 8.1) and Part B score was 7.5 (SD = 10.3). These R-GPTS scores are a little lower than the UK representative group of 10,382 reported by Freeman and Loe (Reference Freeman and Loe2023) (mean R-GPTS Part A score = 9.4 (SD = 9.1) and Part B score = 9.7 (SD = 11.5). The mean GCB score was 35.9 (SD = 14.6).
Table 1. Socio-demographic information about the participant group
Experiences of excessive mistrust in others
A summary of participants’ reported encounters with excessive mistrust behaviors in others is provided in Figure 1 (and see the full frequency table in Supplementary materials Table SC1). The most commonly encountered example of excessive mistrust was a person ‘Saying that they would not get a COVID-19 vaccination because of concerns about the real motivation behind the vaccine rollout,” with 698 (67.4%) participants reporting that they knew at least one person who exhibited this behavior. The least commonly encountered example was a person ‘Thinking that others are targeting them in order to bully or exploit them, and so isolating from the world and refusing to leave their home,” with 328 (31.7%) participants knowing at least one person who showed this behavior. Participants recognized excessive mistrust in other people across a range of relationships. A summary of the categories of people in which people recognized each mistrust example can be found in Figure 2 (and see the full frequency table in Supplementary materials Table SC1). On average (weighted average among those who recognized mistrust in at least one person), 155 (27.6%) participants recognized excessive mistrust in a friend, 135 (25.7%) participants in an acquaintance, 92 (17.9%) participants in a stranger, and 44 (8.6%) participants in a partner.
Figure 1. Observing mistrustful behaviors in other people.
Figure 2. Nature of the relationship with the person who was mistrustful.
Out of the 1,036 respondents, 182 (17.6%) had not observed any of the nine excessive mistrust examples in other people; 85 (8.2%) participants had observed one example; 81 (7.8%) participants had observed two examples; 102 (9.8%) had observed three examples; 84 (8.1%) had observed four examples; 92 (8.9%) had observed five examples; 92 (8.9%) had observed six examples; 91 (8.8%) had observed seven examples; 81 (7.8%) had observed eight examples; and 146 (14.1%) participants had observed all nine excessive mistrust examples.
There were moderate associations between all the nine examples of observing mistrustful behavior (see Supplementary materials Table SA1; ρ range between .49 and .76). People were more likely to have observed mistrust behaviors if they had higher levels of paranoia themselves (R-GPTS Part A: ideas of reference subscale, r = 0.43, p < .001, N = 1036; R-GPTS Part B: ideas of persecution subscale, r = 0.47, p < 0.001, N = 1036) or higher general conspiracy belief scores (GCB score, r = 0.48, p < .001, N = 1036).
Definitions of paranoia
Eighteen codes were developed from the content analysis of participants’ free text definitions of paranoia. Cohen’s Kappa was calculated to assess inter-rater reliability of the coding. The Kappa value indicated an almost perfect agreement between the two raters (κ = 0.87). These codes are presented in Figure 3 (see Supplementary materials Table SC2 for a full frequency table with example definitions from the sample). Three hundred and thirteen (30.2%) participants defined paranoia similarly to the technical definition of paranoia (having inaccurate fears that others are trying to harm them). The next most common definition was of paranoia meaning worry or fear (n = 206, 19.9%).
Figure 3. Coding of participants’ definitions of the word paranoia.
Experiences and understanding of paranoia
A summary of responses to questions about paranoia, after the provision of a definition (an enduring belief that other people (or organizations) are deliberately trying to harm you when they are not), is presented in Figure 4 (see Supplementary materials Table SC3 for a full frequency table). Four hundred and thirty-six (42.1%) participants knew someone who previously or currently experienced paranoid thoughts. Paranoia, similarly to the other examples of excessive mistrust, was also recognized in others across a range of relationships. Paranoia was least commonly recognized in strangers (n = 19, 4.4%).
Figure 4. Noticing paranoia in other people and how paranoia is understood.
Over half of the participant group (n = 600, 57.9%) reported that either they did not understand why people may get paranoid (‘I do not understand it at all’) or understood only a little (‘I understand a little why people may get paranoid’). The most endorsed opinions about the initial occurrence of paranoia were that it could either develop around ‘adolescence’ and ‘young adulthood’ (n = 442; 42.6%) or ‘at any age’ (n = 366; 35.5%). Around half of participants thought that paranoia levels increase or decrease depending on what goes on in someone’s life (‘It depends on what goes on in your life’) (n = 544, 52.5%). Almost two-thirds of participants (n = 674; 65.1%) thought that men and women experienced similar levels of paranoia.
Participants endorsed a range of different causes for paranoia (see Figure 5; or Supplementary materials Table SC4 for a full frequency table). On average, participants endorsed 20 causes (SD = 7) out of the 34 presented. The five most ‘disagreed’ with (disagree + strongly disagree) causes for paranoia were: ‘God’s will’ (n = 581; 56.1%), ‘diet’ (n = 455; 44.0%), ‘environmental pollution’ (n = 444; 42.9%), ‘a germ or virus’ (n = 404; 39.0%), and ‘low intelligence’ (n = 346; 33.4%). The five potential paranoia causes that participants neither agreed nor disagreed with the most were: ‘genetics’ (n = 481; 46.4%), ‘chance, random’ (n = 424; 40.9%), ‘diet’ (n = 383; 37.0%), ‘environmental pollution’ (n = 382; 36.9%), ‘low intelligence’ (n = 370; 35.7%) and ‘biased reasoning’ (n = 370; 35.7%). The five most ‘agreed’ with (agree + strongly agree) causes for paranoia were: ‘excessive worry’ (n = 884; 85.3%), ‘taking illicit drugs’ (n = 881; 85.0%), ‘stress’ (n = 850; 82.0%), ‘anxiety’ (n = 842; 81.3%), and ‘depression’ (n = 834; 80.5%).
Figure 5. Endorsement rates for perceived causes of paranoia.
When participants were asked to indicate what they thought the main cause of paranoia was, 227 (21.9%) participants reported that ‘taking illicit drugs’ was the main cause of paranoia. 103 (9.9%) participants reported that ‘excessive worry’ was the main cause of paranoia. Ninety-seven (9.4%) reported that the main cause of paranoia was ‘stress.” Ninety-seven (9.4%) reported that the main cause of paranoia was ‘a chemical imbalance in the brain’ and 84 (8.1%) participants reported that trauma was the main cause of paranoia.
Factor analysis was used to group the responses to the potential causes of paranoia. During the EFA of the 34 causes, one cause was deleted due to low communality (<0.30) (‘taking prescription drugs’), three were deleted because they did not load onto any factor (loadings <0.30) (‘dementia,” ‘social media,” ‘lack of sleep’), and four were deleted due to cross loadings (‘someone’s upbringing,” ‘people having done bad things to them,” ‘excessive worry,” ‘stress’). With the remaining 26 causes, a seven-factor structure was identified that explained 52% of the variance: negative affect, external and uncontrollable factors, being mistreated, flawed cognition, substance use, family influences, and internal/personal vulnerabilities. Correlations between the factors were r = 0.10–0.46. Factor loadings and communalities for the 7-factor oblimin-rotated model derived from the EFA are presented in Table 2.
Table 2. Factor loadings and communalities for the oblimin rotated 7-factor structure for the causes of paranoia
Note. This table summarizes the final factor structure from the EFA (n = 725).
CFA in the derivation sample (n = 725) showed that the seven-factor model had an acceptable fit based on SRMR and RMSEA values, but the CFI and TFI suggested the model could be improved further (χ2 = 1229.65, df = 278, p < 0.001, SRMR = 0.070, RMSEA = 0.069, CFI = 0.878, TLI = 0.857). CFA in the confirmatory sample (n = 311) demonstrated a similar model fit (χ2 = 763.34, df = 278, p < 0.001, SRMR = 0.076, RMSEA = 0.075, CFI = 0.864, TLI = 0.841).
A summary of participants’ endorsements (or lack thereof) of potential sources of help for someone experiencing paranoia can be found in Figure 6 (see Supplementary materials Table SC4 for a full frequency table). A majority of participants neither strongly agreed nor strongly disagreed with any of the proposed factors that could help with paranoia (i.e. there was limited certainty in answers). The three most agreed with (agree + strongly agree) sources of help for paranoia were: ‘seeing a therapist’ (n = 846; 81.6%), ‘getting more sleep’ (n = 749; 72.3%), and ‘going to the doctors’ (n = 741; 71.5%). The three most disagreed with (disagree + strongly disagree) sources of help were: ‘they don’t need help’ (n = 785; 75.7%), ‘they cannot be helped’ (n = 716; 69.1%), and ‘visiting a religious minister or spiritual guide’ (n = 450; 43.4%). The three sources of help for paranoia that participants endorsed ‘neither agree nor disagree’ (that is may have had no view) with the most were: ‘a better diet’ (n = 470; 45.4%), ‘visiting a place of worship’ (n = 364; 35.1%), and ‘visiting a religious minister or spiritual guide’ (n = 354; 34.2%).
Figure 6. Endorsement rates for potential ways to reduce paranoia.
In exploratory analyses, scores for perceived causes of paranoia were compared across gender, age, ethnicity, and level of education (see Supplementary materials B: Analysis of demographic factors). A summary of mean factor scores for the total sample, as well as a breakdown by gender, age, and ethnicity can be found in Supplementary Table SB1.
Females scored significantly higher than males on sum scores for Factors 1 (negative affect), 3 (being mistreated), 5 (substance use), 6 (family influences), and 7 (internal/personal vulnerabilities), indicating greater endorsement of the corresponding items (see Supplementary Table SB2).
A MANOVA indicated that age may affect what participants endorse as causes of paranoia (F(28, 4112) = 3.96, p < 0.001). There was a significant effect of age on factor scores for the following factors (see Supplementary Table SB3): ‘external and uncontrollable factors’, ‘being mistreated’, ‘flawed cognition’, and ‘family influences’. Post-hoc Tukey’s t-tests revealed indicated that 25–39 year olds endorsed significantly more highly ‘external and uncontrollable factors’ and ‘family influences’ as causes for paranoia than 55–69 and 70+ year olds, and endorsed the factor ‘being mistreated’ significantly more highly than 55–69 year olds. 18–24 year olds endorsed ‘flawed cognition’ items significantly less highly than 40–54 and 70+ year olds.
A MANOVA also indicated that ethnicity may affect endorsement of causes of paranoia (F(28, 4068) = 2.55, p < 0.001). Post-hoc Tukey’s t-tests for Factor 2 (external and uncontrollable factors) revealed significant differences between responses for White and Asian ethnicities, with those of White ethnicity endorsing items in factor 2 (external and uncontrollable factors) significantly less often than those of Asian ethnicity.
There were no statistically significant differences in factor scores across levels of education.
Discussion
The overwhelming majority of people who took part in the survey, representative of the general UK population on a number of basic socio-demographic factors, had encountered excessive mistrust in other people during the past year. We chose examples that were potentially observable from behavior (e.g. people talking about their fears), that were likely excessive (e.g. not taking potentially appropriate medical procedures) or false (e.g. climate change is a hoax). We asked about different domains of excessive mistrust, which were found to be moderately associated. The most commonly observed examples of excessive mistrust were related to the pandemic, which is another example of its significant impact on trust. Mistrust was primarily recognized in people the respondents knew, rather than strangers. If the respondent had higher levels of paranoia or conspiracy-type thinking, then they were more likely to report noticing such thinking in the people around them. Broadly, the results support the emerging evidence that excessive mistrust is fairly ubiquitous in the population and likely affecting behavior.
The survey had a significant focus on paranoia, given its importance in mental health. The general public’s use of the word paranoia is heterogeneous: 18 distinct categories of definitions were identified. The mental health definition of paranoia was the definition most commonly used by the participants but was still only used by just less than one-third of people. When the mental health definition of paranoia was provided to participants, the respondents were not particularly confident in their knowledge about the issue. It could well be valuable to have engaging, accessible, informational resources on the topic for the general public. However, the respondents’ views did align with the research evidence on paranoia in an important way: recognition that paranoia is mostly caused by a large number and range of factors. There were also accurate views about the emergence of paranoia in adolescence and early adulthood and that men and women experience similar levels of paranoia. Overall, participants generally disagreed with the statement that people with paranoia cannot be helped or do not need help. The importance in overcoming paranoia of both psychological therapy and medication were recognized. The significance of sleep was also highlighted, which is likely ahead of the awareness in mental health treatment provision (Freeman et al., Reference Freeman, Sheaves, Goodwin, Yu, Nickless, Harrison and Espie2017).
Our study has several limitations. First, this was a cross-sectional, anonymous, self-report study conducted online. Therefore, while we were able to check for data quality, there is no way of checking whether participants were responding honestly to the questions, and responses might be affected by a variety of self-report biases. Second, we employed a non-probability online quota sampling method, meaning that while the sample was generally representative of the UK population, the individuals may not be representative. Fourth, individuals from Asian and Asian British ethnicities were underrepresented in our study. Fifth, our assessment of participants’ recognition of excessive mistrust was limited to nine discrete examples. Although these examples captured some commonly held excessively mistrustful opinions, they are not exhaustive, and we may have overlooked other frequent or less common instances of excessive mistrust. Lastly, the survey was only conducted in one country, and results are likely to vary by location. Nevertheless, the survey provides an intriguing first snapshot of how excessive mistrust may be frequently witnessed in other people and how sense is made of it that potentially affects responses. The survey not only moves beyond the study of individual experiences of excessive mistrust, but also informs about the social world in which it is happening.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291725102365.
Acknowledgements
The project was funded by a NIHR Senior Investigator Award to DF (NIHR202385). DF and TK are supported by the NIHR Oxford Health Biomedical Research Centre. SL is supported by an NIHR Doctoral Fellowship (NIHR301483). The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.
Competing interests
The authors declare none.
Open access
