Antecedents of Risk Perception and Compliance With Public Health Guidance During COVID-19–Model Questionnaire

Abstract

The Antecedents of Risk Perception and Compliance With Public Health Guidance During COVID-19—Model Questionnaire (Kim, Lee, & Dai, 2023) is a comprehensive, 28-item psychometric instrument developed to evaluate the structural mechanisms through which misinformation exposure influences public trust, cognitive risk appraisals, and behavioral adherence to non-pharmaceutical interventions (NPIs) during global health emergencies. Grounded in risk communication theory, the Health Belief Model, and institutional trust literature, the instrument captures five distinct latent constructs: COVID-19 Misinformation Exposure (4 items), Trust in Public Health Experts (6 items), Trust in Government (6 items), Perceived Severity (3 items), and Compliance with Public Health Guidance (9 items). Administered to a nationally representative sample of United States adults (aged 18 and older), the measurement model demonstrates exceptional psychometric properties. Confirmatory factor analysis (CFA) established a robust structural fit (χ² = 1533.373, df = 336, Comparative Fit Index [CFI] = 0.967, Tucker-Lewis Index [TLI] = 0.963, Root Mean Square Error of Approximation [RMSEA] = 0.050, and Standardized Root Mean Square Residual [SRMR] = 0.045). Convergent validity is evidenced by Average Variance Extracted (AVE) values consistently meeting or exceeding the 0.50 threshold across all constructs, alongside high composite reliability coefficients ranging between 0.80 and 0.99. Discriminant validity was rigorously verified via the Fornell–Larcker criterion, where each latent variable’s AVE surpassed its squared correlations with all other model constructs. Utilizing mixed 5-point and 7-point Likert scales tailored to each dimension, this questionnaire provides epidemiologists, behavioral scientists, and public health authorities with an empirically validated diagnostic tool to identify institutional vulnerabilities, model risk trajectories, and formulate targeted behavioral interventions against health infodemics.

Keywords

COVID-19 Misinformation Exposure, Trust in Public Health Experts, Trust in Government, Perceived Severity, Compliance with Public Health Guidance, Risk Perception, Public Health Attitudes, Structural Equation Modeling, Health Communication, Protective Behaviors, Confirmatory Factor Analysis

Authors

The instrument was designed, psychometrically evaluated, and validated by a multidisciplinary team of communication scientists and risk researchers:

• Ji Won Kim, Ph.D. (Corresponding Author)
  ORCID: 0000-0001-6680-2258
  Affiliation: Department of Media and Communication, City University of Hong Kong, Kowloon Tong, Hong Kong, China.
  Email: [email protected]
• Jiyoung Lee, Ph.D.
  ORCID: 0000-0002-0800-9355
  Affiliation: Department of Media and Communication, Sungkyunkwan University, Seoul, Republic of Korea.
• Yue Dai, M.A. / Ph.D. Candidate
  Affiliation: Department of Media and Communication, City University of Hong Kong, Kowloon Tong, Hong Kong, China.

Purpose

The outbreak of the COVID-19 pandemic precipitated an unprecedented global crisis characterized not only by severe epidemiological threats but also by a widespread “infodemic”—a virulent overabundance of accurate and inaccurate information circulating across digital and interpersonal communication ecosystems. In response to this public health challenge, Kim, Lee, and Dai (2023) developed the Antecedents of Risk Perception and Compliance With Public Health Guidance During COVID-19—Model Questionnaire to disentangle the complex socio-cognitive pathways connecting false information exposure to behavioral outcomes.

The primary clinical, epidemiological, and theoretical purpose of this measurement tool is to evaluate how the consumption of prevalent medical and institutional falsehoods alters citizen appraisal of systemic trust, subjective threat perceptions, and voluntary adherence to disease mitigation protocols. Prior public health surveillance systems routinely documented compliance rates or misinformation prevalence in isolation; however, they frequently lacked psychometrically robust, integrated frameworks capable of modeling the causal chain from external informational inputs to internal affective-cognitive states and terminal health actions.

Specifically, the instrument addresses several vital objectives in public health research and policy design:

• Quantifying Misinformation Burden: It operationalizes the self-reported frequency of encountering dominant, high-impact conspiracy theories and medical untruths (e.g., financial manipulation of mortality counts, underestimation of viral severity, pseudoscientific pharmacological cures, and anti-mask falsehoods).
• Disaggregating Institutional Trust: The scale distinguishes between epistemic trust in specialized medical and epidemiological authorities (Public Health Experts) and political trust in administrative governing bodies (Government), recognizing that these two institutional dimensions often function disparately during polarized socio-political crises.
• Assessing Cognitive Risk Appraisals: It captures perceived disease severity—a critical psychological catalyst within protective behavior frameworks—allowing researchers to test whether misinformation directly blunts threat perception or acts indirectly by eroding institutional credibility.
• Tracking Adherence to Non-Pharmaceutical Interventions: The instrument inventories a comprehensive battery of preventative behaviors recommended by global bodies such as the World Health Organization (WHO) and the Centers for Disease Control and Prevention (CDC), including physical distancing, surface sanitation, personal hygiene, and facial mask usage.

By capturing these multi-tiered dynamics within a unified structural architecture, the questionnaire functions as an empirical assessment battery for academic investigators modeling crisis communication dynamics, health agencies formulating counter-misinformation campaigns, and policy analysts evaluating why specific demographic cohorts resist scientifically supported prophylactic directives.

Psychological Construct

The instrument operationalizes five interrelated psychological constructs, reflecting cognitive, affective, evaluative, and behavioral facets of crisis adaptation:

1. COVID-19 Misinformation Exposure

Misinformation exposure within this psychometric battery is conceptualized as the conscious awareness and frequency of encountering empirically disproven, misleading, or conspiratorial narratives regarding the etiology, severity, management, and treatment of COVID-19. Drawing upon cognitive information-processing paradigms, this construct does not measure initial belief endorsement directly, but rather the ambient informational salience of pseudoscientific claims circulating in media and interpersonal environments. The scale samples four high-frequency misinformation themes identified in national fact-checking databases: institutional corruption (hospitals inflating death metrics for financial gain), risk trivialization (claims that 99% of infections are completely harmless), physical harm narratives regarding protective gear (allegations of hypoxia or hypercapnia from mask-wearing), and unverified pharmacotherapies (claims that hydroxychloroquine is a proven cure).

2. Trust in Public Health Experts

Epistemic institutional trust reflects public confidence in the technical competence, procedural transparency, benevolence, and communicative integrity of professional health authorities (such as epidemiologists, medical advisory boards, and public health agencies). Adapted from validated organizational trust paradigms (e.g., Lang & Hallman, 2005), this 6-item construct measures six core structural facets: perceived domain competence, administrative efficacy, information disclosure, dedication to the collective welfare, responsiveness to public anxieties, and truthfulness. This construct isolates the expert medical authority from broader political governance, enabling granular assessment of whether scientific legitimacy remains intact when political polarization deepens.

3. Trust in Government

Socio-political trust represents citizen appraisals of state and federal administrative entities responsible for crisis management and public policy enactment. Using a symmetrical 6-item structure parallel to the expert trust scale, this dimension evaluates perceived state competence, policy performance, transparency of institutional communications, adherence to public interest, public responsiveness, and overall veracity. In risk analysis literature, government trust serves as a political filter: when citizens view the state as benevolent and truthful, official behavioral mandates are perceived as legitimate public safety measures rather than authoritarian overreach.

4. Perceived Severity

Perceived severity constitutes a central cognitive evaluation within classic threat-coping theories, reflecting an individual’s subjective appraisal of the magnitude, lethality, and clinical consequence of contracting the pathogen. Grounded in the conceptualizations of Witte (1996) and the Health Belief Model, this 3-item construct measures beliefs regarding whether COVID-19 represents a fatal, permanently damaging, or intrinsically lethal biological agent. Highly perceived severity fosters sufficient cognitive tension to motivate self-protective actions, whereas attenuated severity appraisals create psychological inertia and non-compliance.

5. Compliance with Public Health Guidance

The behavioral outcome dimension captures self-reported execution frequency across nine specific non-pharmaceutical interventions officially recommended during the pandemic. Rather than assessing abstract health intentions, the 9-item scale evaluates behavioral adherence across distinct behavioral domains: spatial restriction (staying home, avoiding crowds, avoiding social gatherings), physical distancing (maintaining a 6-foot perimeter), personal barrier protection (wearing face masks), and hygiene protocols (handwashing for 20 seconds, surface disinfection, avoiding mucosal touch, and hygienic respiratory etiquette). This granular operationalization allows researchers to evaluate whether misinformation impairs individual protective practices universally or alters specific behaviors selectively.

Theoretical Framework

The conceptual architecture of the model synthesizes principles from cognitive psychology, health communication, risk sociology, and political science. Primarily, it integrates three foundational frameworks: the Health Belief Model (HBM) (Rosenstock, 1974; Becker, 1974), the Extended Parallel Process Model (EPPM) (Witte, 1992, 1996), and the Protective Action Decision Model (PADM) (Lindell & Perry, 2012).

Cognitive Threat and Behavioral Motivation

According to the Health Belief Model and EPPM, protective health actions are fundamentally driven by cognitive appraisals composed of perceived susceptibility and perceived severity. When individuals perceive a health threat as severe and fatal, an internal drive state emerges, prompting protective behavioral responses—provided that perceived barriers are low and response efficacy is established. In Kim et al.’s (2023) structural framework, perceived severity acts as the direct psychological antecedent that translates distal institutional evaluations into proximal physical compliance.

The Dual-Pathway Model of Institutional Trust

Institutional trust literature (Luhmann, 1979; Mayer, Davis, & Schoorman, 1995) posits that trust functions as a vital cognitive shortcut (heuristic) to reduce social complexity during crises of acute scientific uncertainty. When ordinary citizens lack the specialized technical knowledge required to evaluate novel virological data, they rely on institutional cues. The model formalizes two separate institutional trust pathways: epistemic trust (public health experts) and political-administrative trust (government). Under normative risk communication paradigms, elevated trust in both entities fosters heightened threat comprehension and compliance with behavioral recommendations. However, during health crises, political polarization and infodemics create institutional asymmetries where citizens may align trust along partisan or ideological boundaries.

Misinformation as Cognitive Disruption

The unifying theoretical postulate of the model is that misinformation operates as a systemic cognitive disruptor. Exposure to counter-attitudinal or pseudoscientific narratives diminishes compliance through two primary mechanisms:

1. Direct Threat Mitigation: False narratives often minimize the biological lethality of the virus (e.g., claiming 99% of cases are harmless), directly diminishing perceived severity and thereby eroding the baseline motivation to act defensively.
2. Institutional Trust Erosion: Misinformation frequently weaponizes conspiracy theories (e.g., hospitals inflating fatality metrics for financial kickbacks) that specifically undermine the perceived benevolence, integrity, and competence of public health institutions and governing authorities.

As institutional trust collapses, official warnings are discounted as manipulative, rendering the resulting behavioral guidelines suspect or actively rejected. Thus, the instrument models misinformation exposure not merely as an isolated knowledge deficit, but as an upstream antecedent that cascades through institutional credibility networks to influence risk appraisal and individual health actions.

Validity

The psychometric validity of the instrument was established through rigorous analytical procedures conducted on a nationwide adult sample in the United States, adhering to established structural equation modeling standards (Hair et al., 2010; Kline, 2015).

Convergent Validity

Convergent validity evaluates the extent to which observed indicators of a specific construct share a high proportion of common variance. In Kim et al.’s (2023) validation study, convergent validity was confirmed using two rigorous psychometric benchmarks:

• Average Variance Extracted (AVE): The AVE reflects the overall amount of variance captured by the construct relative to measurement error. For every latent factor in the model (Misinformation Exposure, Expert Trust, Government Trust, Perceived Severity, and Behavioral Compliance), the empirical AVE value was found to meet or exceed the standard threshold of 0.50, demonstrating that more than 50% of the variance in the items is explained by the underlying construct.
• Standardized Factor Loadings: Individual indicator loadings across all subscales demonstrated statistically significant (p < .001) and substantial standardized coefficients, corroborating that the items reliably manifest their designated latent domains.

Discriminant Validity

Discriminant validity confirms that latent constructs within the model are empirically distinct from one another. This was evaluated via the classic Fornell–Larcker criterion. According to this standard, discriminant validity is demonstrated if the square root of the AVE for each construct is substantially greater than the correlation between that construct and any other construct in the model (or equivalently, if each construct’s AVE exceeds its squared inter-construct correlations). In the validation sample, all latent constructs fulfilled this condition without exception:

• The AVE of Trust in Public Health Experts exceeded its shared variance with Trust in Government, proving that participants cognitively differentiated medical-scientific authority from political-administrative leadership.
• The AVE of COVID-19 Misinformation Exposure demonstrated distinct independence from cognitive appraisals of Perceived Severity, verifying that misinformation tracking does not conflate with personal threat evaluations.

Construct and Predictive Validity

Structural equation modeling demonstrated robust predictive and criterion-related validity. As theoretically anticipated, exposure to COVID-19 misinformation exerted significant negative direct and indirect effects on public health compliance, operating through the mediation of institutional trust and risk perception. Trust in public health experts emerged as a powerful positive predictor of perceived severity and subsequent compliance, whereas trust in government exhibited nuanced pathways influenced by ideological framing, confirming the instrument’s sensitivity to real-world socio-political dynamics.

Reliability

The scale’s reliability was determined through rigorous evaluation of internal consistency across each individual dimension within the structural measurement model.

Rather than relying solely on Cronbach’s alpha—which assumes equal tau-equivalence and often underestimates reliability in multidimensional settings—the authors computed Composite Reliability (CR) indices for all five latent constructs. In psychometric theory (Hair et al., 2010), a CR value of 0.70 represents the standard benchmark for acceptable internal consistency, with values exceeding 0.80 denoting superior reliability.

The empirical findings revealed exceptional reliability parameters across the scale:

• Composite Reliability Range: The CR coefficients for all latent factors ranged from 0.80 to 0.99, reflecting strong internal consistency and indicating that measurement error within individual items was minimal.
• Dimensional Performance: Both institutional trust dimensions (Expert Trust and Government Trust) demonstrated high internal consistency (CR > 0.90), reflecting the systematic structural coherence of the adapted 6-item evaluation batteries. Similarly, the 3-item Perceived Severity subscale and the 9-item Compliance inventory demonstrated strong homogeneity, ensuring that composite scores derived from these subscales can be employed in complex predictive and multivariate modeling without attenuation from measurement error.

Factor Analysis

To assess the structural integrity, dimensional uniqueness, and overall measurement architecture of the 28-item battery, Kim, Lee, and Dai (2023) conducted Confirmatory Factor Analysis (CFA) within a covariance-based Structural Equation Modeling (SEM) framework.

Model Fit Parameters

The initial hypothesized five-factor measurement model was tested against the empirical covariance matrix collected from the national sample of U.S. adults. Goodness-of-fit was determined using widely recognized cut-off conventions established by Hu and Bentler (1999) and Kline (2015) (e.g., CFI and TLI ≥ 0.95; RMSEA ≤ 0.06; SRMR ≤ 0.08). The confirmatory factor analysis yielded the following fit indices:

• Chi-Square Statistic (χ²): 1533.373
• Degrees of Freedom (df): 336
• Comparative Fit Index (CFI): 0.967
• Tucker-Lewis Index (TLI): 0.963
• Root Mean Square Error of Approximation (RMSEA): 0.050
• Standardized Root Mean Square Residual (SRMR): 0.045

The CFI (0.967) and TLI (0.963) values both exceed the rigorous 0.95 criterion, indicating good incremental fit relative to a baseline null model. Concurrently, the absolute fit indices confirmed model parsimony: the RMSEA of 0.050 falls precisely within the threshold indicative of close model fit, while the SRMR value of 0.045 demonstrates that the average standardized discrepancy between observed and model-implied covariances is small.

Factor Architecture and Structural Stability

The five-factor solution demonstrated stability without exhibiting anomalous parameter estimates, such as negative error variances (Heywood cases) or unstandardized loadings exceeding unity. The CFA definitively demonstrated that:

1. The 4 misinformation items coalesce into a coherent standalone indicator of ambient false-narrative exposure.
2. The 12 trust items cleanly separate into two distinct 6-item factors (Experts vs. Government), corroborating the theoretical postulate that citizens maintain distinct evaluations for technocratic versus political institutions.
3. The 3 severity indicators load onto a discrete threat-appraisal dimension.
4. The 9 preventive actions represent an integrated behavioral adherence construct spanning multiple non-pharmaceutical intervention classes.

Instrument / Measurement Tool

• Full Instrument Name: Antecedents of Risk Perception and Compliance With Public Health Guidance During COVID-19—Model Questionnaire
• Primary Reference Citation: Kim, J. W., Lee, J., & Dai, Y. (2023). Misinformation and the paradox of trust during the COVID-19 pandemic in the U.S.: Pathways to risk perception and compliance behaviors. Journal of Risk Research, 26(5), 469–484. https://doi.org/10.1080/13669877.2023.2176910
• Test Type: Original Multi-Construct Survey Inventory / Psychometric Measurement Model
• Target Population: Human adults (18 years and older, male and female); originally validated among United States community-dwelling adults.
• Administration Format: Self-administered paper-and-pencil or computer-assisted self-interview (CASI) online survey.
• Total Number of Items: 28 items across 5 latent constructs.
• Dimensional Breakdown:
  • Construct 1: COVID-19 Misinformation Exposure (4 items)
  • Construct 2: Trust in Public Health Experts (6 items)
  • Construct 3: Trust in Government (6 items)
  • Construct 4: Perceived Severity (3 items)
  • Construct 5: Compliance with Public Health Guidance (9 items)
• Response Formats:
  • Misinformation Exposure Subscale: 5-point Likert scale ranging from 1 (“Not at all”) to 5 (“A lot”).
  • Trust Subscales (Experts & Government) and Perceived Severity Subscale: 7-point Likert scale ranging from 1 (“Strongly disagree”) to 7 (“Strongly agree”).
  • Compliance Subscale: 7-point behavioral frequency scale ranging from 1 (“Not at all”) to 7 (“All of the time”).
• Scoring and Computational Procedures:
  • Individual subscale scores are typically computed as arithmetic means across items within each respective domain, yielding continuous subscale scores reflecting: (a) Misinformation Exposure Intensity (range 1–5), (b) Epistemic Expert Trust (range 1–7), (c) Political Government Trust (range 1–7), (d) Disease Severity Appraisal (range 1–7), and (e) Non-Pharmaceutical Compliance Level (range 1–7).
  • No reverse-coded items are present in the published 28-item model; all statements are worded consistently within each construct domain. Higher scores represent higher levels of the measured construct.
  • For advanced structural equation modeling, indicators are modeled as continuous or ordinal observed variables linked to their respective latent factors, allowing measurement error correction.

Permissions & Fee and Test Year

The Antecedents of Risk Perception and Compliance With Public Health Guidance During COVID-19—Model Questionnaire was published in 2023. The intellectual property and formal copyright for the original article and measurement tables reside with the authors (Ji Won Kim, Jiyoung Lee, and Yue Dai) and the publisher, Taylor & Francis Group (in the Journal of Risk Research).

Fee Structure: The scale is available free of charge for non-commercial academic research, pedagogical purposes, and epidemiological public health surveillance, provided appropriate formal citation of the original source publication (Kim, Lee, & Dai, 2023) is maintained.

Permissions and Licensing: Researchers intending to adapt the items, translate the scale into other languages, or incorporate the instrument into funded commercial interventions or proprietary software platforms should contact the corresponding author, Dr. Ji Won Kim ([email protected]), or request permissions through the Copyright Clearance Center / Taylor & Francis RightsLink service.

References

• Becker, M. H. (1974). The Health Belief Model and personal health behavior. Health Education Monographs, 2(4), 324–473. https://doi.org/10.1177/109019817400200407
• Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis: A global perspective (7th ed.). Pearson Prentice Hall.
• Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/10705519909540118
• Kim, J. W., Lee, J., & Dai, Y. (2023). Misinformation and the paradox of trust during the COVID-19 pandemic in the U.S.: Pathways to risk perception and compliance behaviors. Journal of Risk Research, 26(5), 469–484. https://doi.org/10.1080/13669877.2023.2176910
• Kline, R. B. (2015). Principles and practice of structural equation modeling (4th ed.). Guilford Press.
• Lang, J. T., & Hallman, W. K. (2005). Who does the public trust? The case of genetically modified food in the United States. Risk Analysis, 25(5), 1241–1252. https://doi.org/10.1111/j.1539-6924.2005.00668.x
• Lindell, M. K., & Perry, R. W. (2012). The Protective Action Decision Model: Theoretical modifications and additional evidence. Risk Analysis, 32(4), 616–632. https://doi.org/10.1111/j.1539-6924.2011.01647.x
• Liu, Z. F., & Yang, J. Z. (2020). In the wake of spirals: Testing the effects of exposure to misinformation on social media and the role of social presence in shaping risk perception. Telematics and Informatics, 54, 101460. https://doi.org/10.1016/j.tele.2020.101460
• Luhmann, N. (1979). Trust and power. John Wiley & Sons.
• Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709–734. https://doi.org/10.2307/258792
• Rosenstock, I. M. (1974). Historical origins of the Health Belief Model. Health Education Monographs, 2(4), 328–335. https://doi.org/10.1177/109019817400200403
• Witte, K. (1992). Putting the fear back into fear appeals: The Extended Parallel Process Model. Communication Monographs, 59(4), 329–349. https://doi.org/10.1080/03637759209376276
• Witte, K., Cameron, K. A., McKeon, J. M., & Berkowitz, J. M. (1996). Predicting risk behaviors: Development and validation of a diagnostic scale. Journal of Health Communication, 1(4), 317–341. https://doi.org/10.1080/108107396127988

13. Items of the Scale (Questionnaire)