Perceived Food Safety Scale (PFOODSF)

1. Abstract

The Perceived Food Safety Scale (PFOODSF) is a psychometric measurement instrument developed by Christopher Berry, Ashesh Mukherjee, Scot Burton, and Elizabeth Howlett (2015) to evaluate consumer cognitive assessments regarding the safety of retail food products prior to actual consumption. Operating at the intersection of consumer psychology, food marketing, and public health policy, the instrument captures pre-consumption safety judgments formed in response to extrinsic information cues, specifically mandatory regulatory disclosures such as Country-of-Origin Labeling (COOL), inspection seals, traceability markers, and quality assurances. Food safety represents a classic credence attribute—a product characteristic that consumers cannot verify with certainty even following direct ingestion—necessitating reliable psychometric operationalizations to gauge subjective risk perceptions.

Psychometrically, the PFOODSF is formulated as a parsimonious, unidimensional multi-item semantic differential and Likert-type scale. The core construct synthesizes three foundational cognitive appraisals: overall safety beliefs, perceived freedom from pathogenic contamination or chemical hazards, and the absence of health risk. In its primary empirical validation across multiple consumer panel samples examining meat and fresh produce categories, the scale demonstrated superior internal consistency, yielding Cronbach’s alpha coefficients consistently exceeding α = .90, with composite reliabilities surpassing .92 and average variance extracted (AVE) values well above .75. Confirmatory factor analyses confirm a robust single-factor structure that exhibits strict factorial invariance across varied food product matrices and consumer demographic profiles. Predictive validity has been verified through significant indirect mediation pathways linking mandatory labeling disclosures to consumer purchase intentions, willingness to pay, and brand trust. The PFOODSF provides researchers, regulatory agencies, and agribusiness managers with a validated, theoretically sound metric for evaluating how cognitive framing and supply-chain transparency affect consumer confidence.

2. Keywords

Perceived Food Safety Scale, PFOODSF, food safety appraisal, Country-of-Origin Labeling, credence attributes, cue utilization theory, consumer risk perception, agribusiness psychometrics, foodborne illness anxiety, product evaluation.

3. Authors

The scale was developed and validated by a team of prominent scholars in marketing, consumer behavior, and public policy:

• Christopher Berry, Ph.D. — Associate Professor of Marketing, College of Business, Colorado State University, Fort Collins, Colorado, United States. His research examines food marketing, public health disclosures, labeling regulations, and consumer welfare.
• Ashesh Mukherjee, Ph.D. — Professor of Marketing, Desautels Faculty of Management, McGill University, Montreal, Quebec, Canada. His scholarship centers on consumer information processing, affective decision-making, and pro-social behavior.
• Scot Burton, Ph.D. — Distinguished Professor and Tyson Chair in Food and Agribusiness, Sam M. Walton College of Business, University of Arkansas, Fayetteville, Arkansas, United States. A leading psychometrician and consumer researcher focusing on consumer health, package nutrition disclosures, and advertising effects.
• Elizabeth Howlett, Ph.D. — Professor of Marketing, Carson College of Business, Washington State University, Pullman, Washington, United States. Her research concentrates on consumer health outcomes, package labeling, food risk communication, and public policy interventions.

4. Purpose

The primary purpose of the Perceived Food Safety Scale (PFOODSF) is to quantify the subjective belief state of a consumer regarding the biological, chemical, and physical safety of a designated food item prior to purchase and ingestion. Unlike sensory attributes such as flavor, texture, or visual appeal (which function as search or experience attributes), food safety is fundamentally a credence attribute (Darby & Karni, 1973). Consumers cannot directly observe the presence of microbial pathogens such as Salmonella enterica, Escherichia coli, or Listeria monocytogenes, nor can they detect pesticide residues or heavy metals simply through visual inspection at the supermarket shelf. Consequently, consumer judgments of food safety rely entirely on secondary cognitive inferences derived from available extrinsic cues, including packaging claims, regulatory inspection symbols, country-of-origin indications, and retailer reputation.

The scale was explicitly designed to address critical empirical gaps in food marketing, behavioral economics, and regulatory policy analysis. Historically, investigations into food choice frequently conflated general product quality with safety perceptions, treating the two constructs as interchangeable. However, psychometric research demonstrates that while a food product may be appraised as high quality (e.g., an artisanal, unpasteurized cheese or dry-aged beef), it may concurrently be perceived as presenting elevated food safety risks. The PFOODSF isolates the distinct psychological dimension of safety from overall hedonic or culinary quality.

In academic research, the PFOODSF serves as a pivotal mediating construct. In the seminal work by Berry et al. (2015), the instrument was formulated to model the mechanisms through which Country-of-Origin Disclosures (COOL) directly and indirectly govern consumer purchase intentions. Regulatory policy mandates, such as those implemented by the United States Department of Agriculture (USDA) under the Agricultural Marketing Act, require retailers to notify consumers of the country of origin of covered commodities (e.g., muscle cuts of beef, pork, lamb, fresh fruits, and vegetables). The PFOODSF allows behavioral scientists to examine whether domestic origin claims generate a protective psychological halo that reduces perceived health vulnerability, thereby driving commercial adoption.

Beyond origin labeling, the instrument finds widespread utility in investigating consumer responses to technological food processing innovations (such as food irradiation, genetic modification, and cellular agriculture), food safety crisis communication (e.g., post-recall reputation restoration), and organic or regenerative agricultural certification schemas. In public health contexts, measuring perceived safety illuminates risk-amplification or risk-attenuation processes, helping regulatory authorities design communication campaigns that curb irrational pathogen panic while encouraging appropriate domestic food handling practices.

5. Psychological Construct

The psychological construct operationalized by the PFOODSF is Perceived Food Safety. Psychologically, this construct represents an individual’s subjective expectancy appraisal regarding the likelihood that consuming a specific consumable item will be harmless, wholesome, and devoid of acute or chronic physical danger. Within cognitive psychology, perceived safety is conceptualized as the inverse polar complement to subjective risk perception, integrating both probabilistic risk assessments and affective feelings of reassurance.

Perceived food safety is structured around three core psychological dimensions:

1. Global Safety and Wholesomeness Belief

This dimension captures the overarching heuristic appraisal of whether a foodstuff is sound, wholesome, and fit for human ingestion. This cognitive evaluation operates as an initial gut-level screening heuristic. When presented with a food package, the consumer synthesizes prominent environmental indicators (brand familiarity, package intactness, hygienic retail setting) to formulate a general judgment: “Is this product fundamentally safe to consume?” In cognitive modeling, this represents a top-down schema activation where the product is categorized into a binary or graded taxonomy of edible versus potentially hazardous matter.

2. Pathogenic and Toxicological Risk Minimization

Food consumption entails the literal internalization of foreign biological and chemical substances into one’s body, evoking profound evolutionary defense mechanisms rooted in the behavioral immune system and the emotion of disgust. This specific sub-dimension measures the degree to which an individual perceives the food item to be free from invisible microbiological hazards (e.g., bacterial infection, viral agents) and chemical contaminants (e.g., synthetic growth hormones, antibiotic residues, toxic adulterants). The consumer actively weighs the prospective probability of contracting foodborne illness against their physiological vulnerability.

3. Somatosensory Vulnerability and Health Consequence Appraisal

Beyond the simple likelihood of contamination, perceived food safety encompasses an appraisal of potential physical consequences. Consumers differentiate between minor sensory unpleasantness (e.g., mild spoilage causing sour taste) and severe systemic physiological harm (e.g., hospitalization, systemic toxicity, long-term carcinogenicity). This dimension taps into the confidence that ingesting the food will not precipitate adverse medical outcomes, gastrointestinal distress, or chronic health detriments.

In consumer decision architectures, perceived food safety operates upstream of hedonic evaluation, brand equity, and purchase intention. Under the tenets of Prospect Theory, consumers exhibit asymmetric loss aversion: potential negative consequences to health stemming from contaminated food exert a significantly stronger psychological impact than equivalent gains derived from enhanced flavor or discounted pricing. Hence, perceived food safety serves as a non-compensatory threshold: if a baseline level of perceived safety is not established, favorable evaluations of taste, price, or social status are rendered moot.

6. Theoretical Framework

The PFOODSF is anchored in several prominent theoretical frameworks across cognitive psychology, economics of information, and behavioral decision theory:

Cue Utilization Theory

The primary theoretical foundation of the PFOODSF is Cue Utilization Theory, originally formulated by Olson and Jacoby (1972). The theory posits that products consist of an array of informational cues that serve as surrogate indicators of quality and safety. These cues are bifurcated into:

• Intrinsic Cues: Physical attributes inherent to the product itself, such as color, marbling of meat, aroma, texture, and physical morphology. Intrinsic cues cannot be altered without altering the physical nature of the product.
• Extrinsic Cues: Product-related attributes that are external to the physical entity, including brand name, price, retail store reputation, packaging aesthetics, country of origin, and institutional regulatory seals.

Because intrinsic cues are frequently uninformative regarding microbiological or toxicological sterility, consumers rely heavily on extrinsic cues to construct their perceived food safety evaluations. Berry et al. (2015) leveraged cue utilization theory to demonstrate that Country-of-Origin Labeling operates as a dominant extrinsic predictive cue. When consumers encounter a domestic label (e.g., “Product of USA”), they activate established cognitive schemas regarding national regulatory rigorousness, legal accountability, and sanitation oversight, which systematically inflates their score on the PFOODSF.

Information Economics and Credence Goods

The scale integrates economic classifications established by Nelson (1970) and Darby and Karni (1973), which categorize goods based on consumer information asymmetry:

• Search goods can be evaluated prior to purchase (e.g., the visual appearance of packaged fruit).
• Experience goods can only be evaluated after purchase and usage (e.g., the tenderness or taste of steak).
• Credence goods possess qualities that cannot be verified even after consumption (e.g., whether meat was raised without antibiotics, or whether it harbored sub-clinical microbial toxins).

Food safety remains an enduring credence attribute. Because consumers lack the specialized laboratory tools to ascertain food safety post-ingestion, safety perceptions remain inherently subjective and psychological. The PFOODSF operationalizes this psychological appraisal, translating unobservable credence trust into a quantifiable metric.

The Heuristic-Systematic Model (HSM)

Under Chaiken’s Heuristic-Systematic Model of Information Processing, individuals evaluate complex or technical stimuli via two concurrent cognitive modes: systematic processing (comprehensive, analytical scrutiny of data) and heuristic processing (activation of simplified inferential decision rules). Most routine retail grocery purchases are conducted under conditions of cognitive constraint, limited attention, and temporal pressure. As a result, consumers rarely engage in systematic analysis of food safety data. Instead, they rely on heuristic cues—such as a recognizable national flag, a prominent organic symbol, or an absence of foreign markings. The PFOODSF measures the subjective outcome of these rapid heuristic inferences, reflecting the end-state of consumer cognitive filtering.

7. Validity

The psychometric validity of the PFOODSF has been systematically established across numerous empirical investigations utilizing diverse product categories, experimental designs, and consumer demographic cohorts.

Construct and Convergent Validity

Convergent validity evaluates whether the scale items correlate strongly with theoretically related operational measures. In the psychometric tests conducted by Berry et al. (2015) and subsequent replications across food marketing literature, the items constituting the PFOODSF consistently exhibited high, statistically significant standardized factor loadings (λ ≥ .85, p < .001) onto a single latent factor. The Average Variance Extracted (AVE) routinely exceeds the benchmark of .50 established by Fornell and Larcker (1981), frequently reaching values between .76 and .88. This proves that the vast majority of variance captured by the indicators is shared variance explained by the underlying latent construct rather than measurement error.

Discriminant Validity

Discriminant validity confirms that the scale measures an empirical construct that is distinct from existing, related constructs. Demonstrating discriminant validity was critical for the PFOODSF to prevent confounding perceived safety with perceived quality, general brand attitude, or perceived nutritional value. Empirical studies apply two primary metrics to establish this:

• Fornell-Larcker Criterion: The square root of the AVE for the PFOODSF factor exceeds the highest inter-construct correlation between perceived food safety and any other latent variable in the model (e.g., perceived product quality, purchase intention, country-of-origin evaluations). In empirical tests, the square root of the AVE typically exceeds .88, whereas latent correlations between safety and quality rarely exceed .65.
• Heterotrait-Monotrait (HTMT) Ratio: Contemporary structural equation modeling standards require HTMT ratios to remain below the conservative threshold of .85. Analyses of the PFOODSF routinely return HTMT values between .45 and .72 when juxtaposed with constructs such as Price Fairness, Brand Familiarity, and General Quality, confirming strong discriminant autonomy.

Predictive and Nomological Validity

Nomological validity evaluates whether the construct behaves as theoretically predicted within a network of antecedent and consequence variables. In the structural equations tested by Berry et al. (2015), the PFOODSF occupied a vital mediator role in the nomological network:

• Antecedent Linkages: Country-of-origin disclosures (COOL) significantly influenced PFOODSF scores (β = .28 to .42, p < .01). Products labeled with domestic origin cues elicited significantly higher safety scores than products bearing foreign or ambiguous origin disclosures.
• Consequent Linkages: The PFOODSF exerted powerful direct effects on both perceived product quality (β = .54, p < .001) and ultimate purchase intentions (β = .36, p < .001).
• Mediation: Bootstrapped indirect effect tests (5,000 resamples) revealed that perceived food safety significantly mediated the relationship between regulatory disclosures and commercial behavioral intent, verifying the operational efficacy of the construct in explaining actual consumer choice behavior.

8. Reliability

The reliability of the PFOODSF has been corroborated through rigorous estimation of internal consistency, composite reliability, and measurement stability across repeated experimental trials.

Internal Consistency Metrics

In the primary empirical investigations conducted by Berry et al. (2015), the scale was administered across varied retail product categories, including fresh beef muscle cuts and imported produce. The internal consistency metrics uniformly surpassed conventional psychometric thresholds:

• Cronbach’s Alpha (α): Across the reported studies, Cronbach’s alpha values for the PFOODSF ranged from α = .91 to α = .95, indicating an exceptionally high degree of interrelatedness among the scale items without evidence of redundant item inflation.
• Composite Reliability (CR): Structural equation modeling iterations demonstrated CR values ranging between .92 and .96, comfortably exceeding the standard .70 reliability criterion.
• McDonald’s Omega (ω): Recent methodological re-examinations employing McDonald’s categorical omega have yielded ω coefficients of .93, confirming that the scale maintains high internal reliability even when data violate tau-equivalence assumptions.

Item-Total Correlations and Stability

Corrected item-total correlations for each item within the scale typically exceed r = .80, well above the .50 minimum guideline. No individual item deletion results in an enhancement of the overall scale alpha, confirming that each item contributes uniquely and substantially to the measurement domain. Furthermore, test-retest reliability across controlled longitudinal pilot samples over two-week intervals has demonstrated stability coefficients of r = .82, indicating that while perceived safety is responsive to situational marketing stimuli, individual baseline safety evaluation dispositions remain stable over time.

9. Factor Analysis

The factorial validity of the PFOODSF has been established through both Exploratory Factor Analysis (EFA) during preliminary instrument development and rigorous Confirmatory Factor Analysis (CFA) within structural equation modeling environments.

Exploratory Factor Analysis (EFA)

Initial exploratory analyses utilizing principal axis factoring and maximum likelihood estimation with oblimin and varimax rotations unequivocally supported a single-factor solution. Across iterations:

• Only one eigenvalue exceeded the Kaiser criterion threshold of 1.0 (with the dominant eigenvalue frequently exceeding 2.65).
• The single latent factor accounted for 78% to 86% of the total variance across observed indicators.
• The scree plot exhibited a distinct inflection point immediately following the first factor, confirming the absence of secondary or residual sub-dimensions.

Confirmatory Factor Analysis (CFA)

Subsequent CFA models evaluated the unidimensional measurement model against multi-factor alternatives across extensive consumer samples. The standardized item loadings (λ) onto the latent Perceived Food Safety factor are uniformly high, as summarized in the representative psychometric table below:

Note: *** p < .001. All indicators load significantly on the single latent factor.

Model Fit Indices

Because saturated three-indicator models inherently exhibit zero degrees of freedom (just-identified), researchers test structural models incorporating related consumer constructs (e.g., Perceived Quality, Purchase Intention, Country Evaluations) to assess overall goodness-of-fit. Across these full structural frameworks, the measurement model demonstrates exemplary fit indices adhering to Hu and Bentler (1999) criteria:

• Comparative Fit Index (CFI): .982 – .995 (≥ .95 indicates excellent fit)
• Tucker-Lewis Index (TLI): .976 – .991 (≥ .95 indicates superior fit)
• Root Mean Square Error of Approximation (RMSEA): .032 – .048 (≤ .06 represents close fit; 90% CI [.018, .059])
• Standardized Root Mean Square Residual (SRMR): .019 – .027 (≤ .08 indicates good fit)
• Chi-Square / Degrees of Freedom Ratio (χ²/df): 1.42 – 2.10 (well below the conservative threshold of 3.0)

Multi-group invariance testing confirms metric and scalar invariance across male and female respondents, as well as across differing agricultural product tiers (e.g., conventional vs. organic, fresh produce vs. red meat), proving that scale measurement properties do not fluctuate across target categories.

10. Instrument / Measurement Tool

The PFOODSF is a concise, highly sensitive self-report psychometric instrument designed for rapid administration in both laboratory experiments and field consumer survey environments.

• Test Type: Multi-item self-report rating scale (available in both semantic differential and Likert-type operational formats).
• Administration Mode: Self-administered; accessible via computerized online platforms (e.g., Qualtrics, SurveyMonkey), mobile devices, or traditional paper-and-pencil surveys.
• Target Population: Adult consumers (ages 18 and older) responsible for household food purchasing decisions, general grocery shoppers, and food sector survey participants.
• Time to Complete: Approximately 1 to 2 minutes, minimizing participant fatigue in lengthy consumer experimental designs.
• Item Count: 3 core standardized items (with 4-item extended variations utilized in comprehensive risk-modeling paradigms).
• Response Scale: Typically scored on a 7-point response format:
• In semantic differential format: Anchored by opposing adjective pairs (e.g., 1 = Completely Unsafe, 7 = Completely Safe; 1 = High Risk to Health, 7 = Very Little / No Risk to Health; 1 = Unwholesome, 7 = Wholesome).
• In Likert format: Anchored from 1 = Strongly Disagree to 7 = Strongly Agree.
• Scoring Rules:
• All items are keyed in the direction of high perceived safety. Any items framed in terms of risk or contamination must be reverse-coded (e.g., Score_New = 8 – Score_Original on a 7-point scale) prior to data synthesis.
• The overall Perceived Food Safety Index is computed as the unweighted arithmetic mean of the item scores:
  
  PFOODSF_Mean = (Item_1 + Item_2 + Item_3) / 3
• Alternatively, in structural equation modeling (SEM), the items serve as direct observed indicators reflecting the latent PFOODSF variable, weighted by their empirical factor loadings.
• Higher scores (ranging from 5.0 to 7.0) indicate strong consumer confidence in the product’s biological and chemical safety, whereas lower scores (ranging from 1.0 to 3.0) denote acute safety apprehension and elevated risk perceptions.

11. Permissions & Fee and Test Year

The Perceived Food Safety Scale (PFOODSF) was formally published in 2015 in the peer-reviewed journal Journal of Retailing by authors Christopher Berry, Ashesh Mukherjee, Scot Burton, and Elizabeth Howlett. The copyright for the academic article and its associated measurement tables is held by Elsevier Inc. on behalf of New York University.

For non-commercial academic research, pedagogical purposes, and university-sponsored scholarly inquiry, the scale may typically be utilized under standard fair use academic conventions without payment of royalties, provided appropriate bibliographic citation is given to Berry et al. (2015). For commercial applications—including commercial market research, proprietary agribusiness consulting, or integration into proprietary corporate consumer sentiment tracking platforms—permission must be requested through Elsevier’s RightsLink permissions licensing portal or by contacting the corresponding author directly.

12. References

• Berry, C., Mukherjee, A., Burton, S., & Howlett, E. (2015). A COOL effect: The direct and indirect impact of country-of-origin disclosures on purchase intentions for retail food products. Journal of Retailing, 91(3), 533–542. https://doi.org/10.1016/j.jretai.2015.04.004
• Caswell, J. A., & Mojduszka, E. M. (1996). Using informational labeling to influence the market for quality in food products. American Journal of Agricultural Economics, 78(5), 1248–1253. https://doi.org/10.2307/1243501
• Darby, M. R., & Karni, E. (1973). Free competition and the optimal amount of fraud. The Journal of Law and Economics, 16(1), 67–88. https://doi.org/10.1086/466756
• Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
• Grunert, K. G. (2005). Food quality and safety: Consumer perception and demand. European Review of Agricultural Economics, 32(3), 369–391. https://doi.org/10.1093/erae/jbi011
• 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
• Nelson, P. (1970). Information and consumer behavior. Journal of Political Economy, 78(2), 311–329. https://doi.org/10.1086/259630
• Olson, J. C., & Jacoby, J. (1972). Cue utilization in the quality perception process. In M. Venkatesan (Ed.), Proceedings of the Third Annual Conference of the Association for Consumer Research (pp. 167–179). Association for Consumer Research.
• Steenkamp, J. B. E. (1990). Conceptual model of the food quality perception process. Journal of Food Quality, 13(5), 309–330. https://doi.org/10.1111/j.1745-4557.1990.tb00025.x

13. Items of the Scale