pith. machine review for the scientific record. sign in

arxiv: 2604.04849 · v1 · submitted 2026-04-06 · 💻 cs.CY

Recognition: no theorem link

Latent Profiles of AI Risk Perception and Their Differential Association with Community Driving Safety Concerns: A Person-Centered Analysis

Amir Rafe , Anika Baitullah , Subasish Das
Authors on Pith no claims yet

Pith reviewed 2026-05-10 19:05 UTC · model grok-4.3

classification 💻 cs.CY
keywords AI risk perceptionlatent class analysisdriving safetyautonomous vehiclesperson-centered analysispublic attitudesworldview risk perceptionPew survey
0
0 comments X

The pith

AI risk perceptions form four latent classes that predict different levels of concern about driving safety.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper uses latent class analysis on nine AI risk-perception items from a large national survey to uncover distinct groups among U.S. adults. It identifies four profiles and demonstrates that membership in these groups is linked to significantly different views on nine community driving safety issues. Higher general AI concern tends to align with seeing driving hazards as more severe, though trust shapes the direct comparison between AI and human drivers. This approach treats risk attitudes as expressions of underlying worldviews that cut across technology and transportation domains. The resulting segmentation offers a way to understand how public views on AI may influence acceptance of new driving technologies.

Core claim

Latent class analysis applied to nine AI risk-perception indicators from a nationally representative Pew survey (n = 5,255) identifies four profiles—Moderate Skeptics (17.5%), Concerned Pragmatists (42.8%), AI Ambivalent (10.6%), and Extreme Alarm (29.1%)—that differ significantly on all nine driving-safety outcomes. Higher AI concern maps monotonically onto greater perceived driving-hazard severity, while the exception of comparative AI versus human driving evaluation is driven by trust rather than concern level. The cross-domain links supply person-level evidence for worldview-based risk structuring.

What carries the argument

Latent class analysis of nine AI risk-perception indicators, followed by Bolck-Croon-Hagenaars corrected distal outcome analysis to test class differences on driving-safety measures.

If this is right

  • Higher AI concern classes report greater perceived severity across most driving hazards.
  • All nine driving-safety outcomes show statistically significant differences by class.
  • Comparative judgments of AI versus human driving depend on trust levels separate from overall concern.
  • The four-class structure supplies a segmentation framework for communications about autonomous vehicles.
  • Demographic and ideological variables predict which AI risk profile an individual holds.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The profiles could be tested as predictors of support for specific AV regulations or usage intentions.
  • Similar person-centered methods might reveal parallel cross-domain links for other technologies such as gene editing.
  • If the classes prove stable over time, they could guide targeted interventions to address transportation risk perceptions.
  • The monotonic pattern suggests that general efforts to build AI trust might reduce perceived driving hazards.

Load-bearing premise

The nine AI risk-perception indicators from the Pew survey capture stable underlying worldviews that structure risk perceptions without substantial measurement error or confounding by unmeasured demographics and ideology.

What would settle it

A replication study using a different set of AI risk questions or additional controls for ideology that finds no significant differences across classes on the driving safety outcomes.

Figures

Figures reproduced from arXiv: 2604.04849 by Amir Rafe, Anika Baitullah, Subasish Das.

Figure 1
Figure 1. Figure 1: Analytic pipeline overview. Six stages proceed from data preparation and variable operationalization through weighted LCA enumeration and model selection (yellow, RQ 1), BCH-corrected distal outcome analysis (orange, RQ 2), and survey-weighted multinomial logistic regression (red, RQ 3), concluding with four robustness checks. The dashed sidebar on the left maps each research question to its corresponding … view at source ↗
Figure 2
Figure 2. Figure 2: Model fit indices for latent class solutions (𝐾 = 1–7). (A) BIC, SABIC, and AIC with the shaded region highlighting the 𝐾 = 4 elbow. (B) Classification entropy with the 0.80 adequacy threshold. (C) Log-likelihood trajectory. (D) Free parameters per model with BLRT and VLMR-LRT significance annotations. The red bar marks the recommended 𝐾 = 4 solution. from Cultural Theory (Douglas and Wildavsky, 1982). For… view at source ↗
Figure 3
Figure 3. Figure 3: Item-response probability profiles for the four-class solution (𝐾 = 4). Each panel displays one of the nine LCA indicator items; within each panel, the four classes are shown with response-category probabilities. Class labels and weighted prevalences: C1 = Moderate Skeptics (17.5%), C2 = Concerned Pragmatists (42.8%), C3 = AI Ambivalent (10.6%), C4 = Extreme Alarm (29.1%). membership. The Hausman–McFadden … view at source ↗
Figure 4
Figure 4. Figure 4: BCH-corrected class-specific means for nine driving-safety distal outcomes. Error bars denote sandwich-estimated standard errors. For driving-problem items (1–3 scale), lower bars indicate greater perceived severity. For road rage (1–5) and safety trend (1–6), higher bars indicate more pessimistic assessment and greater frequency, respectively. than ideological self-placement, was the stronger predictor. I… view at source ↗
Figure 5
Figure 5. Figure 5: Forest plot of odds ratios from the survey-weighted multinomial logistic regression, with Concerned Pragmatists (C2) as the reference class. Points indicate odds ratios; horizontal lines denote 95% Wald confidence intervals. The dashed vertical line at OR = 1.0 marks no effect. Filled points indicate statistically significant coefficients (𝑝 < 0.05). classification error correction. Configural measurement … view at source ↗
read the original abstract

Public attitudes toward artificial intelligence (AI) and driving safety are typically studied in isolation using variable-centered methods that assume population homogeneity, yet risk perception theory predicts that these evaluations covary within individuals as expressions of underlying worldviews. This study identifies latent profiles of AI risk perception among U.S. adults and tests whether these profiles are differentially associated with community driving safety concerns. Latent class analysis was applied to nine AI risk-perception indicators from a nationally representative survey (Pew Research Center American Trends Panel Wave 152, n = 5,255); Bolck-Croon-Hagenaars corrected distal outcome analysis tested class differences on nine driving-safety outcomes, and survey-weighted multinomial logistic regression identified demographic and ideological predictors of class membership. Four classes emerged: Moderate Skeptics (17.5%), Concerned Pragmatists (42.8%), AI Ambivalent (10.6%), and Extreme Alarm (29.1%), with all nine driving-safety outcomes significantly differentiated across classes. Higher AI concern mapped monotonically onto greater perceived driving-hazard severity; the exception, comparative evaluation of AI versus human driving, was driven by trust rather than concern level. The cross-domain covariation provides person-level evidence for the worldview-based risk structuring posited by Cultural Theory of Risk and yields a four-class segmentation framework for AV communication that links AI risk orientation to transportation safety attitudes.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript applies latent class analysis (LCA) to nine AI risk-perception indicators drawn from the Pew Research Center American Trends Panel Wave 152 (n=5,255 U.S. adults) to identify four latent profiles: Moderate Skeptics (17.5%), Concerned Pragmatists (42.8%), AI Ambivalent (10.6%), and Extreme Alarm (29.1%). Bolck-Croon-Hagenaars (BCH) corrected distal outcome analysis is used to test class differences on nine community driving-safety outcomes, while survey-weighted multinomial logistic regression identifies demographic and ideological predictors of class membership. Results show significant differentiation across all nine outcomes, with a generally monotonic increase in perceived driving-hazard severity as AI concern rises (except for comparative AI-versus-human driving evaluations, which the authors attribute to trust). The work interprets these patterns as person-level evidence for worldview-based risk structuring under Cultural Theory of Risk and proposes a four-class segmentation for autonomous-vehicle communication.

Significance. If the associations survive proper covariate adjustment, the study supplies useful descriptive evidence that AI risk perceptions and driving-safety concerns covary within individuals rather than across independent domains, extending Cultural Theory of Risk into a new substantive area. The nationally representative sample and standard application of LCA plus BCH methods give the descriptive segmentation reasonable credibility. The resulting four-class framework could inform targeted safety messaging, though the current lack of adjustment limits stronger claims about stable individual differences.

major comments (2)
  1. [Methods (BCH analysis) and Results (distal outcomes)] Methods, BCH distal outcome analysis: The BCH-corrected tests are described as examining raw class differences on the nine driving-safety outcomes. The manuscript separately reports survey-weighted multinomial logistic regression showing that age, education, political ideology, and other covariates predict class membership. Because these same covariates are established correlates of both AI risk perception and safety-hazard ratings, the reported significant differentiations and monotonic mappings may be confounded. This issue is load-bearing for the central claim that the profiles reflect stable, worldview-driven individual differences whose associations with driving outcomes are not artifacts of demographics.
  2. [Results (latent class analysis)] Results, LCA model selection: The manuscript reports the four-class solution and class sizes but does not present standard fit statistics (BIC, AIC, entropy, Lo-Mendell-Rubin likelihood-ratio test) or robustness checks (alternative class numbers, split-sample replication). Without these, it is impossible to evaluate whether the four-class structure is optimal or stable, which directly affects the validity of all subsequent BCH and regression results.
minor comments (2)
  1. [Abstract] The abstract states that 'all nine driving-safety outcomes significantly differentiated across classes' but supplies neither p-values nor effect-size information, making it difficult for readers to judge practical significance.
  2. [Methods (measures)] Variable coding details for the nine AI risk-perception indicators and the nine driving-safety outcomes are not fully reported (e.g., exact item wording, response scales, or handling of missing data), which would aid reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which highlight important issues for strengthening the manuscript. We address each major comment point by point below, indicating planned revisions where appropriate.

read point-by-point responses

  1. Referee: Methods (BCH analysis) and Results (distal outcomes)] Methods, BCH distal outcome analysis: The BCH-corrected tests are described as examining raw class differences on the nine driving-safety outcomes. The manuscript separately reports survey-weighted multinomial logistic regression showing that age, education, political ideology, and other covariates predict class membership. Because these same covariates are established correlates of both AI risk perception and safety-hazard ratings, the reported significant differentiations and monotonic mappings may be confounded. This issue is load-bearing for the central claim that the profiles reflect stable, worldview-driven individual differences whose associations with driving outcomes are not artifacts of demographics.

    Authors: We acknowledge that the BCH approach yields unadjusted distal outcome differences and that the covariates identified in the multinomial regression (age, education, political ideology) are known correlates of both AI risk perceptions and driving safety ratings. This creates a legitimate possibility of confounding for the reported class differences. We will revise the manuscript to explicitly discuss this limitation, reframe the primary interpretations as descriptive patterns of covariation rather than evidence of unconfounded stable individual differences, and add supplementary analyses that incorporate key covariates into the distal outcome models where methodologically feasible. These changes will clarify the scope of the claims while preserving the person-centered descriptive contribution of the profiles. revision: partial

  2. Referee: Results (latent class analysis)] Results, LCA model selection: The manuscript reports the four-class solution and class sizes but does not present standard fit statistics (BIC, AIC, entropy, Lo-Mendell-Rubin likelihood-ratio test) or robustness checks (alternative class numbers, split-sample replication). Without these, it is impossible to evaluate whether the four-class structure is optimal or stable, which directly affects the validity of all subsequent BCH and regression results.

    Authors: We agree that standard LCA fit indices and robustness checks are necessary to justify the four-class solution. These statistics were generated during model estimation but were omitted from the submitted version. In the revision we will insert a new table (or expanded methods/results subsection) reporting AIC, BIC, entropy, and Lo-Mendell-Rubin test values for the 1- through 6-class solutions, together with our model-selection rationale. We will also add split-sample replication results demonstrating class stability. These additions will directly resolve the concern and improve transparency. revision: yes

Circularity Check

0 steps flagged

No circularity: standard LCA on external survey data with no self-referential reductions

full rationale

The paper applies latent class analysis to nine indicators from an independent, publicly referenced Pew survey (n=5,255) and uses Bolck-Croon-Hagenaars correction plus multinomial logistic regression for associations and predictors. No equations, fitted parameters, or results are shown to reduce by construction to quantities defined within the paper itself. No self-citations appear as load-bearing premises, and the four-class solution plus monotonic associations are direct empirical outputs rather than tautological renamings or imported uniqueness claims. The derivation chain is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The claim depends on the validity of the latent class model fitted to nine survey items and the assumption that these items measure worldview-structured risk perceptions. No new physical or theoretical entities are introduced.

free parameters (1)
  • Number of latent classes
    The four-class solution was selected from the data; the abstract does not detail the model comparison process that fixed this number.
axioms (1)
  • domain assumption AI risk perceptions covary within individuals as expressions of underlying worldviews rather than independent variables
    This premise is stated explicitly as the justification for using person-centered rather than variable-centered methods.

pith-pipeline@v0.9.0 · 5550 in / 1427 out tokens · 67995 ms · 2026-05-10T19:05:04.398340+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 47 canonical work pages

  1. [1]

    Demographicinfluencesontrustinartificialintelligenceacross cognitive domains: A statistical perspective

    Alasmari,A.A.,Alruwaili,R.F.,Alotaibi,R.F.,Youssef,I.K.,Asklany,S.A.,2025. Demographicinfluencesontrustinartificialintelligenceacross cognitive domains: A statistical perspective. PLoS ONE 20, e0331003. doi:10.1371/journal.pone.0331003. Araujo, T., Brosius, A., Goldberg, A.C., Möller, J., de Vreese, C.,

    work page doi:10.1371/journal.pone.0331003 2025

  2. [2]

    Transport Policy 181, 104062

    Balancing innovation and safety: Public preferences for autonomous vehicle testing in Australia. Transport Policy 181, 104062. doi:10.1016/j.tranpol.2026.104062. Bakk, Z., Vermunt, J.K.,

    work page doi:10.1016/j.tranpol.2026.104062 2026

  3. [3]

    Structural Equation Modeling: A Multidisciplinary Journal , author =

    Robustness of stepwise latent class modeling with continuous distal outcomes. Structural Equation Modeling: A Multidisciplinary Journal 23, 20–31. doi:10.1080/10705511.2014.955104. Bansal, P., Kockelman, K.M., Singh, A.,

    work page doi:10.1080/10705511.2014.955104 2014

  4. [4]

    Transportation Research Part C: Emerging Technologies 67, 1–14

    Assessing public opinions of and interest in new vehicle technologies: An Austin perspective. Transportation Research Part C: Emerging Technologies 67, 1–14. doi:10.1016/j.trc.2016.01.019. Bao, L., Calice, M.N., Brossard, D.,

    work page doi:10.1016/j.trc.2016.01.019 2016

  5. [5]

    Journal of Risk Research doi:10.1080/13669877.2025.2553077

    AI scientists’ and lay publics’ views of AI’s social impacts: Segmentation analyses on risk and benefit perceptions. Journal of Risk Research doi:10.1080/13669877.2025.2553077. Bao, L., Krause, N.M., Calice, M.N., Scheufele, D.A., Wirz, C.D., Brossard, D., Newman, T.P., Xenos, M.A.,

    work page doi:10.1080/13669877.2025.2553077 2025

  6. [6]

    Computers in Human Behavior 130, 107182

    Whose AI? How different publics think about AI and its social impacts. Computers in Human Behavior 130, 107182. doi:10.1016/j.chb.2022.107182. Bartolucci,A.,Aquilino,M.C.,Bril,L.,Duncan,J.,VanSteen,T.,2023.Effectivenessofaudiencesegmentationininstructionalriskcommunication: A systematic literature review. International Journal of Disaster Risk Reduction 9...

    work page doi:10.1016/j.chb.2022.107182 2022

  7. [7]

    Journal of Medical Internet Research 25, e40337

    Surveying public perceptions of artificial intelligence in health care in the United States: Systematic review. Journal of Medical Internet Research 25, e40337. doi:10.2196/40337. Bergsma, W.,

    work page doi:10.2196/40337

  8. [8]

    Journal of the Korean Statistical Society 42, 323–328

    A bias-correction for Cramér’s𝑉and Tschuprow’s𝑇. Journal of the Korean Statistical Society 42, 323–328. doi:10.1016/ j.jkss.2012.10.002. Bolck, A., Croon, M., Hagenaars, J.,

    2012

  9. [9]

    Political Analysis 12, 3–27

    Estimating latent structure models with categorical variables: One-step versus three-step estimators. Political Analysis 12, 3–27. doi:10.1093/pan/mph001. Brauner, P., Glawe, F., Liehner, G.L., Vervier, L., Ziefle, M.,

    work page doi:10.1093/pan/mph001

  10. [10]

    Technological Forecasting and Social Change 220, 124304

    Mapping public perception of artificial intelligence: Expectations, risk–benefit tradeoffs, and value as determinants for societal acceptance. Technological Forecasting and Social Change 220, 124304. doi:10.1016/j.tech fore.2025.124304. Collins, L.M., Lanza, S.T.,

    work page doi:10.1016/j.tech 2025

  11. [11]

    Wiley, Hoboken, NJ

    Latent Class and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences. Wiley, Hoboken, NJ. doi:10.1002/9780470567333. Dake, K.,

    work page doi:10.1002/9780470567333

  12. [12]

    volume 22, pp

    Orienting dispositions in the perception of risk: An analysis of contemporary worldviews and cultural biases, in: Journal of Cross-Cultural Psychology. volume 22, pp. 61–82. doi:10.1177/0022022191221006. De-Santis, A., Torres-Toukoumidis, Á., Morejón-Vallejo, R.,

    work page doi:10.1177/0022022191221006

  13. [13]

    SocialSciences&HumanitiesOpen13,102562

    Artificial intelligence in political communication and citizens’ perceptions of disinformationanddemocraticlegitimacyinEcuador. SocialSciences&HumanitiesOpen13,102562. doi:10.1016/j.ssaho.2026.102562. Dempster, A.P., Laird, N.M., Rubin, D.B.,

    work page doi:10.1016/j.ssaho.2026.102562 2026

  14. [14]

    P., LAIRD, N

    Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society: Series B (Methodological) 39, 1–38. doi:10.1111/j.2517-6161.1977.tb01600.x. Douglas, M., Wildavsky, A.,

    work page doi:10.1111/j.2517-6161.1977.tb01600.x 1977

  15. [15]

    University of California Press, Berkeley, CA

    Risk and Culture: An Essay on the Selection of Technological and Environmental Dangers. University of California Press, Berkeley, CA. Dreksler,N.,Law,H.,Ahn,C.,Schiff,D.,Schiff,K.J.,2025. WhatdoesthepublicthinkaboutAI?anoverviewofthepublic’sattitudestowardsAI and a resource for future research. SSRN Electronic Journal doi:10.2139/ssrn.5108572. Dziak, J.J....

    work page doi:10.2139/ssrn.5108572 2025

  16. [16]

    Methodology 12, 107–116

    Comparing the performance of improved classify-analyze approaches for distal outcomes in latent profile analysis. Methodology 12, 107–116. doi:10.1027/1614-2241/a000114. Elliott,M.R.,Jacobsohn,L.,Winston,F.K.,Ginsburg,K.R.,2012. Determiningsubgroupsofteensfortargeteddrivinginjurypreventionstrategies: A latent class analysis approach. Traffic Injury Preven...

    work page doi:10.1027/1614-2241/a000114 2012

  17. [17]

    Transportation Research Part A: Policy and Practice 77, 167–181

    Preparing a nation for autonomous vehicles: Opportunities, barriers and policy recommendations. Transportation Research Part A: Policy and Practice 77, 167–181. doi:10.1016/j.tra.2015.04.003. Hausman, J., McFadden, D.,

    work page doi:10.1016/j.tra.2015.04.003 2015

  18. [18]

    Econometrica 52, 1219–1240

    Specification tests for the multinomial logit model. Econometrica 52, 1219–1240. doi:10.2307/1910997. Hilgarter, K., Granig, P.,

    work page doi:10.2307/1910997

  19. [19]

    Transportation Research Part F: Traffic Psychology and Behaviour 72, 226–243

    Public perception of autonomous vehicles: A qualitative study based on interviews after riding an autonomous shuttle. Transportation Research Part F: Traffic Psychology and Behaviour 72, 226–243. doi:10.1016/j.trf.2020.05.012. Ho, S.S., Goh, T.J.,

    work page doi:10.1016/j.trf.2020.05.012 2020

  20. [20]

    Transportation Research Part F: Traffic Psychology and Behaviour 109, 1293–1312

    Explicating multiple dimensions of public’s and technology developers’ mental models of risks and benefits of artificial intelligence in autonomous vehicles in Singapore. Transportation Research Part F: Traffic Psychology and Behaviour 109, 1293–1312. doi:10.1016/j.trf.2025.01.016. Rafe et al.: Page 19 of 21 AI Risk Perception and Driving Safety Concerns ...

    work page doi:10.1016/j.trf.2025.01.016 2025

  21. [21]

    Scandinavian Journal of Statistics 6, 65–70

    A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics 6, 65–70. Horowitz,M.C.,Kahn,L.,2021. Whatinfluencesattitudesaboutartificialintelligenceadoption:EvidencefromU.S.localofficials. PLoSONE16, e0257732. doi:10.1371/journal.pone.0257732. Hulse, L.M., Xie, H., Galea, E.R.,

    work page doi:10.1371/journal.pone.0257732 2021

  22. [22]

    Safety Science 102, 1–13

    Perceptions of autonomous vehicles: Relationships with road users, risk, gender and age. Safety Science 102, 1–13. doi:10.1016/j.ssci.2017.10.001. Kalambay, P., Novat, N., Kidando, E., Kutela, B., Kitali, A.,

    work page doi:10.1016/j.ssci.2017.10.001 2017

  23. [23]

    doi:10.1016/j.samod.2025.1 00049

    How do socio-demographic and political factors affect public acceptance of autonomous passenger vehicle technology for large trucks? Sustainability Analytics and Modeling 6, 100049. doi:10.1016/j.samod.2025.1 00049. Kish, L.,

    work page doi:10.1016/j.samod.2025.1 2025

  24. [24]

    AI & Society 39, 2445–2456

    Application of artificial intelligence: Risk perception and trust in the work context with different impact levels and task types. AI & Society 39, 2445–2456. doi:10.1007/s00146-023-01699-w. Krügel,S.,Uhl,M.,2025. Aninternationalsurveyonriskdistributionpreferencesforautonomousvehicles. TransportationResearchPartA:Policy and Practice 201, 104695. doi:10.10...

    work page doi:10.1007/s00146-023-01699-w 2025

  25. [25]

    Transportation Research Part F: Traffic Psychology and Behaviour 32, 127–140

    Public opinion on automated driving: Results of an international questionnaire among 5000 respondents. Transportation Research Part F: Traffic Psychology and Behaviour 32, 127–140. doi:10.1016/j.trf.2015.04.014. Labbo, M.S., Qu, L., Xu, C., Bai, W., Ayele Atumo, E., Jiang, X.,

    work page doi:10.1016/j.trf.2015.04.014 2015

  26. [26]

    Nicolás, The bar derived category of a curved dg algebra, Journal of Pure and Applied Algebra 212 (2008) 2633–2659

    Understanding risky driving behaviors among young novice drivers in Nigeria:A latentclassanalysiscoupled withassociationrule miningapproach. AccidentAnalysis&Prevention 200,107557. doi:10.1016/j. aap.2024.107557. Lee,D.,Hess,D.J.,2022. Publicconcernsandconnectedandautomatedvehicles:Safety,privacy,anddatasecurity. HumanitiesandSocialSciences Communications 9,

    work page doi:10.1016/j 2024

  27. [27]

    Leiserowitz, A., Maibach, E.W., Roser-Renouf, C., Smith, N.,

    doi:10.1057/s41599-022-01110-x. Leiserowitz, A., Maibach, E.W., Roser-Renouf, C., Smith, N.,

    work page doi:10.1057/s41599-022-01110-x

  28. [28]

    Biometrika 88, 767–778

    Testing the number of components in a normal mixture. Biometrika 88, 767–778. doi:10.1093/biom et/88.3.767. Lorenzo-Seva, U., ten Berge, J.M.F.,

    work page doi:10.1093/biom

  29. [29]

    Methodology 2, 57–64

    Tucker’s congruence coefficient as a meaningful index of factor similarity. Methodology 2, 57–64. doi:10.1027/1614-2241.2.2.57. Mack, E.A., Miller, S.R., Chang, C.H., Van Fossen, J.A., Cotten, S.R., Savolainen, P.T., Mann, J.,

    work page doi:10.1027/1614-2241.2.2.57

  30. [30]

    Telematics and Informatics 61, 101604

    The politics of new driving technologies: Political ideology and autonomous vehicle adoption. Telematics and Informatics 61, 101604. doi:10.1016/j.tele.2021.101604. Magidson, J., Vermunt, J.K.,

    work page doi:10.1016/j.tele.2021.101604 2021

  31. [31]

    PLoS ONE 6, e17571

    Identifying like-minded audiences for global warming public engagement campaigns: An audience segmentation analysis and tool development. PLoS ONE 6, e17571. doi:10.1371/journal.pone.0017571. Mannering, F.L., Bhat, C.R.,

    work page doi:10.1371/journal.pone.0017571

  32. [32]

    Analytic Methods in Accident Research 1, 1–22

    Analytic methods in accident research: Methodological frontier and future directions. Analytic Methods in Accident Research 1, 1–22. doi:10.1016/j.amar.2013.09.001. Martínez-Buelvas,L.,Rakotonirainy,A.,Grant-Smith,D.,Haque,M.M.,Oviedo-Trespalacios,O.,2025. Publicperceptionsandsupportforpolicies enhancing safety and justice in connected and automated vehic...

    work page doi:10.1016/j.amar.2013.09.001 2013

  33. [33]

    Safety Science 127, 104763

    Public perceptions of autonomous vehicle safety: An international comparison. Safety Science 127, 104763. doi:10.1016/j.ssci.2020.104763. Naiseh,M.,Clark,J.,Akarsu,T.,Hanoch,Y.,Brito,M.,2025. Trust,riskperception,andintentiontouseautonomousvehicles:Aninterdisciplinary bibliometric review. AI & Society 40, 1053–1073. doi:10.1007/s00146-024-01895-2. Nationa...

    work page doi:10.1016/j.ssci.2020.104763 2020

  34. [34]

    Accessed: 2025-03-29

    https://crashstats.nhtsa.dot.gov/A pi/Public/ViewPublication/813561. Accessed: 2025-03-29. National Highway Traffic Safety Administration,

    2025

  35. [35]

    Accessed: 2026-04-03

    https: //crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/813800. Accessed: 2026-04-03. Nazari, F., Noruzoliaee, M., Mohammadian, A.,

    2026

  36. [36]

    Multimodal Transportation 5, 100252

    Autonomous vehicle adoption behavior and safety concern: A study of public perception. Multimodal Transportation 5, 100252. doi:10.1016/j.multra.2025.100252. Nylund, K.L., Asparouhov, T., Muthén, B.O.,

    work page doi:10.1016/j.multra.2025.100252 2025

  37. [37]

    Structural Equation Modeling: A Multidisciplinary Journal 14, 535–569

    Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal 14, 535–569. doi:10.1080/10705510701575396. Othman, K.,

    work page doi:10.1080/10705510701575396

  38. [38]

    International Journal of Sustainable Transportation 18, 62–83

    Segmenting the target audience for transportation demand management programs: An investigation between mode shift and individual characteristics. International Journal of Sustainable Transportation 18, 62–83. doi:10.1080/15568318.2023.2201941. Peng, Y.,

    work page doi:10.1080/15568318.2023.2201941 2023

  39. [39]

    Public Understanding of Science 29, 436–451

    The ideological divide in public perceptions of self-driving cars. Public Understanding of Science 29, 436–451. doi:10.1177/09 63662520917339. Peterson, C.M., Nelson, T.F., Pereira, M.A.,

    work page doi:10.1177/09

  40. [40]

    Driver speeding typologies by roadway behaviours and beliefs: A latent class analysis with a multistate sample of U.S. adults. Transportation Research Part F: Traffic Psychology and Behaviour 81, 373–383. doi:10.1016/j.trf.2021 .06.016. Pew Research Center,

    work page doi:10.1016/j.trf.2021 2021

  41. [41]

    doi:10.1016/j.trip.2022.100594

    Can a racial justice frame help overcome opposition to automated traffic enforcement? Transportation Research Interdisciplinary Perspectives 14, 100594. doi:10.1016/j.trip.2022.100594. Ren, Q., Xu, M.,

    work page doi:10.1016/j.trip.2022.100594 2022

  42. [42]

    Hashemi, M

    Heterogeneity in crash patterns of autonomous vehicles: The latent class analysis coupled with multinomial logit model. Accident Analysis & Prevention 209, 107827. doi:10.1016/j.aap.2024.107827. SAE International,

    work page doi:10.1016/j.aap.2024.107827 2024

  43. [43]

    Technical Report J3016_202104

    Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles. Technical Report J3016_202104. SAE International. Schandl,F.,Fischer,P.,Hudecek,M.F.C.,2025. Predictingacceptanceofautonomousshuttlebusesbypersonalityprofiles:Alatentprofileanalysis. Transportation 52, 1–32. doi:10.1007/s11116-023-10447-4. Schwarz, G.,

    work page doi:10.1007/s11116-023-10447-4 2025

  44. [44]

    The Annals of Statistics7(1), 1–26 (1979) https://doi.org/10.1214/aos/1176344552

    Estimating the dimension of a model. Annals of Statistics 6, 461–464. doi:10.1214/aos/1176344136. Slovic, P.,

    work page doi:10.1214/aos/1176344136

  45. [45]

    Science 236, 280–285

    Perception of risk. Science 236, 280–285. doi:10.1126/science.3563507. Tennant, C., Stilgoe, J., Vucevic, S., Stares, S.,

    work page doi:10.1126/science.3563507

  46. [46]

    Mobilities 20, 292–309

    Public anticipations of self-driving vehicles in the UK and US. Mobilities 20, 292–309. doi:10.1080/17450101.2024.2325386. Tyson, A., Kikuchi, E.,

    work page doi:10.1080/17450101.2024.2325386 2024

  47. [47]

    Technical Report

    Growing public concern about the role of artificial intelligence in daily life. Technical Report. Pew Research Center. URL: https://www.pewresearch.org/science/2023/08/28/growing-public-concern-about-the-role-of-artificial-intelligence-in-daily-life/. Vermunt, J.K.,

    2023

  48. [48]

    International Journal of Human–Computer Interaction, 1–29 (2025) https://doi

    Understanding risk preference and risk perception when adopting high-risk and low-risk AI technologies. International Journal of Human–Computer Interaction 41, 15295–15310. doi:10.1080/10447318.2025.2495844. World Health Organization,

    work page doi:10.1080/10447318.2025.2495844 2025

  49. [49]

    Journalism & Mass Communication Quarterly 102, 382–406

    In AI we trust: The interplay of media use, political ideology, and trust in shaping emerging AI attitudes. Journalism & Mass Communication Quarterly 102, 382–406. doi:10.1177/10776990231190868. Zhang, B., Dafoe, A.,

    work page doi:10.1177/10776990231190868

  50. [50]

    SSRN Electronic Journal doi:10.2139/ssrn.3312874

    Artificial intelligence: American attitudes and trends. SSRN Electronic Journal doi:10.2139/ssrn.3312874. Zhang, Q., Wallbridge, C.D., Jones, D.M.,

    work page doi:10.2139/ssrn.3312874

  51. [51]

    Transportation Research Part A: Policy and Practice 179, 103921

    Public perception of autonomous vehicle capability determines judgment of blame and trust in road traffic accidents. Transportation Research Part A: Policy and Practice 179, 103921. doi:10.1016/j.tra.2023.103921. Rafe et al.: Page 21 of 21

    work page doi:10.1016/j.tra.2023.103921 2023