Perceived Ad Personalization Scale (PAPS)

1. Abstract

The Perceived Ad Personalization Scale (PAPS) is a psychometric instrument designed to evaluate the extent to which an individual subjectively identifies and acknowledges that a commercial message, digital advertisement, or persuasive communication has been deliberately adapted, tailored, or customized to their idiosyncratic characteristics, preferences, or personal background. Rooted in early computer-tailored health communication paradigms pioneered by Arie Dijkstra (2005) and subsequently adapted into online consumer psychology and marketing by Aguirre et al. (2015), the PAPS addresses a foundational distinction between objective ad personalization algorithms and subjective cognitive appraisal. Comprising a parsimonious single-factor structure with three core items, the PAPS assesses explicit recognition using a standard 7-point Likert response scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”). Across diverse empirical investigations in computational advertising, digital marketing, electronic commerce, and behavioral medicine, the scale consistently exhibits exceptional psychometric properties, including high internal consistency reliability (typically Cronbach’s α > .90), strong factor determinacy, robust convergent validity with constructs such as perceived relevance and self-referencing, and sharp discriminant validity against affective constructs like ad liking, privacy concern, and psychological reactance. By capturing the subjective threshold where algorithmic customization transitions into cognitive awareness, the scale functions as an indispensable manipulation check, mediator, and independent variable in advanced empirical research evaluating consumer responses, personalization paradoxes, and digital persuasion dynamics.

2. Keywords

Perceived Ad Personalization Scale, PAPS, advertising personalization, tailored communication, algorithmic customization, subjective perception, consumer cognition, digital marketing, privacy paradox, psychological reactance, self-referencing, psychometric validation

3. Authors

The conceptual foundation and original measurement operationalization were established by Arie Dijkstra, Professor of the Psychology of Health and Illness at the University of Groningen, Department of Social and Organizational Psychology, Groningen, The Netherlands. His primary research focus encompasses computer-tailored behavioral interventions, health education, and cognitive processing of personalized persuasive communications.

The scale was adapted, refined, and validated for digital marketing, consumer behavior, and online advertising contexts by a team of prominent marketing scholars:

• Elizabeth Aguirre – Department of Marketing and Supply Chain Management, School of Business and Economics, Maastricht University, Maastricht, The Netherlands.
• Dominik Mahr – Department of Marketing and Supply Chain Management, School of Business and Economics, Maastricht University, Maastricht, The Netherlands.
• Dhruv Grewal – Toyota Chair in Commerce and Electronic Business and Professor of Marketing, Babson College, Babson Park, Massachusetts, USA.
• Kees de Ruyter – Professor of Marketing, Cass Business School, City University London, London, UK, and Maastricht University.
• Martin Wetzels – Department of Marketing and Supply Chain Management, School of Business and Economics, Maastricht University, Maastricht, The Netherlands.

4. Purpose

In contemporary programmatic media environments, machine learning algorithms and real-time bidding architectures deliver hyper-targeted content based on browsing history, geolocation, demographic segmentation, and psychographic profiling. However, a persistent theoretical and empirical issue in consumer psychology is the misalignment between objective personalization (the algorithmic manipulation executed by the advertiser) and subjective perceived personalization (the conscious recognition by the recipient that the message reflects their personal identity or preferences). The primary purpose of the Perceived Ad Personalization Scale is to capture this subjective appraisal accurately, independently of affective evaluations, brand attitudes, or behavioral intentions.

From an applied research perspective, the PAPS serves multiple critical roles:

• Experimental Manipulation Check: In laboratory and field experiments evaluating personalized versus non-personalized promotional stimuli, researchers utilize the PAPS to confirm whether experimental treatments successfully triggered conscious perception of tailoring across experimental conditions.
• Investigation of the Personalization Paradox: The scale enables scholars to examine the paradoxical consequences of targeted advertising, wherein moderate perceived personalization enhances perceived relevance and click-through intentions, whereas overt or excessive perceived personalization triggers privacy concerns, perceived surveillance, and psychological reactance.
• Mediation and Moderation Modeling: Within structural equation modeling frameworks, the PAPS operates as a crucial cognitive mediator between objective customization parameters (e.g., overt vs. covert data collection, first-party vs. third-party tracking) and downstream outcomes, including brand equity, cognitive resistance, and purchase conversion.
• Public Health and Persuasion Studies: Beyond commercial marketing, the instrument assists public health communicators in validating digital interventions designed for smoking cessation, vaccination promotion, and chronic disease management, verifying whether recipients perceive health messaging as targeted to their unique stage of change.

By measuring cognitive acknowledgment without confounding it with positive sentiment (e.g., ad liking) or negative sentiment (e.g., ad intrusion), the PAPS preserves high theoretical fidelity and provides an unclouded metric of consumer awareness.

5. Psychological Construct

The psychological construct assessed by the PAPS is perceived ad personalization. Conceptually, this construct represents a higher-order cognitive categorization process wherein an individual categorizes an external persuasive stimulus as having been intentionally configured to match their unique self-schema, behavioral history, or individual profile.

Cognitive Recognition vs. Evaluative Appraisal

A crucial theoretical hallmark of this construct is its non-evaluative, cognitive nature. In psychometric literature, scales often suffer from construct contamination when cognitive recognition is blended with affective response (e.g., “I enjoyed this ad because it felt made for me”). The PAPS deliberately isolates cognitive recognition. An individual may score at the maximum of the scale (agreeing completely that an advertisement is personalized) while simultaneously experiencing intense discomfort or anger due to invasive behavioral tracking. Conversely, a consumer might enjoy an advertisement purely because of entertaining creative elements while recognizing that the advertisement is entirely generic and mass-marketed.

Unidimensionality and Semantic Redundancy

The construct is operationalized as unidimensional. It relies on three linguistic representations that reflect the semantic core of individualized communication:

• Personalized (Item 1): Denotes that the message possesses personal relevance and incorporates individualized identifiers or data points.
• Tailored (Item 2): Originating from sartorial metaphors and adapted into behavioral medicine, tailoring emphasizes deliberate adaptation to specific requirements, psychological readiness, or structural characteristics of the individual recipient.
• Customized (Item 3): Refers to deliberate modification from a standard or default format into an individualized structure tailored to user-specific inputs or observed behaviors.

These three facets capture the common core of perceived communicative adaptation, minimizing measurement error while maximizing structural parsimony.

6. Theoretical Framework

The Perceived Ad Personalization Scale is grounded in established models of cognitive psychology, communication theory, and consumer information processing.

The Elaboration Likelihood Model and Self-Referencing

According to the Elaboration Likelihood Model (ELM) developed by Petty and Cacioppo, persuasion occurs via either the central or peripheral route. Central route processing requires both the motivation and ability to scrutinize message arguments. Perceived personalization serves as a primary motivator of central route processing through the psychological mechanism of self-referencing — the cognitive tendency to relate incoming environmental information to one’s existing self-concept. When an individual recognizes an ad as personalized (high PAPS score), self-referencing triggers heightened cognitive elaboration, prompting deeper memory encoding and thorough message evaluation.

Dijkstra’s Working Mechanisms of Computer Tailoring

Dijkstra’s (2005) foundational framework in health education posits that computer-tailored communications achieve their effects via three sequential cognitive processes: attention capture, perceived personal relevance, and message-induced cognitive processing. Perceived personalization constitutes the initial necessary gateway: recipients must first perceive the information as uniquely configured for them before higher-order cognitive restructuring and behavioral adaptation can take place.

The Personalization–Privacy Paradox and Reactance Theory

In digital marketing, the theoretical underpinnings of the PAPS are intertwined with Brehm’s Psychological Reactance Theory and Communication Privacy Management (CPM) Theory. When consumers recognize that an advertisement has been customized based on covert behavioral surveillance (demonstrated by high PAPS scores in the absence of explicit consumer consent), the ad violates perceived privacy boundaries. This violation can induce reactance — an aversive motivational state directed toward restoring threatened behavioral freedom — which drives ad avoidance, ad-blocker adoption, and negative brand attitudes (Aguirre et al., 2015).

7. Validity

The Perceived Ad Personalization Scale has been rigorously evaluated across multiple empirical investigations, demonstrating exceptional psychometric validity across diverse experimental and observational contexts.

Construct and Convergent Validity

Convergent validity is established when an instrument correlates strongly with other theoretical constructs with which it should conceptually align. In the seminal work by Aguirre et al. (2015), the PAPS correlated positively and significantly with perceived relevance ($r = .68, p < .001$) and message involvement ($r = .54, p < .001$). Furthermore, Dijkstra (2005) established that participants exposed to objectively computer-tailored feedback reported significantly higher scores on perceived tailoring items compared to participants exposed to standard, non-tailored materials ($F(1, 412) = 84.32, p < .001$), confirming strong construct validity.

Discriminant Validity

Discriminant validity confirms that the scale measures a distinct construct rather than confounding neighboring phenomena. Using the Fornell-Larcker criterion, the average variance extracted (AVE) of the PAPS consistently exceeds .80, which is substantially higher than the squared correlations between the PAPS and conceptually related constructs, including:

• Perceived Intrusion / Vulnerability: Inter-construct correlations typically range between $r = .31$ and $r = .45$, demonstrating that recognizing personalization is distinct from feeling violated.
• General Ad Attitude ($A_{ad}$): Correlations fluctuate widely (from $r = -.15$ to $r = .52$) depending on whether data collection was overt or covert, proving that the scale does not conflate personalization with positive affect.
• Brand Familiarity: Correlations are modest ($r < .25$), demonstrating that prior brand exposure does not drive scores on perceived ad personalization.

Predictive and Nomological Validity

Nomological validity is demonstrated through structural equation modeling wherein the PAPS acts as an intermediary variable. Studies show that PAPS scores significantly predict click-through intentions, willingness to disclose information, and purchase intentions when mediated by perceived utility, while concurrently predicting negative reactance when moderated by covert tracking contexts (Aguirre et al., 2015; Van Doorn & Hoekstra, 2013).

8. Reliability

Across empirical marketing, health communication, and human-computer interaction literature, the Perceived Ad Personalization Scale displays outstanding internal consistency and reliability metrics.

Internal Consistency

Standard psychometric guidelines dictate that an instrument should achieve a Cronbach’s alpha (α) and composite reliability (CR) of at least .70 for exploratory research and .80 for established measures. The PAPS routinely exceeds these benchmarks:

• Aguirre et al. (2015): Reported a Cronbach’s α of .94 across online behavioral targeting experiments.
• Dijkstra (2005): The original tailored health intervention assessment demonstrated high internal consistency with Cronbach’s α exceeding .88 across longitudinal waves.
• Subsequent Replications: Multiple contemporary studies in leading marketing journals (e.g., Journal of Interactive Marketing, International Journal of Advertising) have reported Cronbach’s α values ranging between .91 and .96, and McDonald’s omega (ω) values exceeding .92.

Test-Retest and Stability Metrics

In laboratory test-retest assessments where stimulus exposure is controlled across brief intervals (e.g., 48-hour re-exposure without altering ad configuration), the scale exhibits high stability coefficients ($r_{tt} > .82$). Because ad exposure in digital media is inherently situational, internal consistency (Cronbach’s α and composite reliability) remains the primary benchmark for reliability evaluation.

9. Factor Analysis

The structural dimensionality of the Perceived Ad Personalization Scale has been validated utilizing both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

When subjected to EFA using principal axis factoring or maximum likelihood extraction with oblimin rotation, the three items consistently load onto a single dominant factor with an eigenvalue exceeding 2.5, explaining over 85% of the total variance. Factor loadings across studies are uniformly high:

• Item 1 (“The ad was personalized”): Standardized factor loading λ ≥ .88
• Item 2 (“The ad was tailored to me”): Standardized factor loading λ ≥ .92
• Item 3 (“The ad was customized for me”): Standardized factor loading λ ≥ .89

Confirmatory Factor Analysis (CFA)

When evaluated within structural equation modeling software (e.g., AMOS, Mplus, lavaan in R), a single-factor CFA model demonstrates strong fit to empirical data. Because a three-item single-factor model has zero degrees of freedom ($df = 0$), it is mathematically just-identified (saturated). However, when evaluated concurrently within multi-construct measurement models containing related constructs, the measurement properties remain robust:

• Standardized Root Mean Square Residual (SRMR): ≤ .024
• Comparative Fit Index (CFI): ≥ .992
• Tucker-Lewis Index (TLI): ≥ .987
• Root Mean Square Error of Approximation (RMSEA): ≤ .045 (90% CI [.000, .068])
• Average Variance Extracted (AVE): Consistently > .80, indicating that over 80% of the variance in the indicators is captured by the underlying latent construct rather than error.

10. Instrument / Measurement Tool

The operational specifications of the Perceived Ad Personalization Scale are detailed below:

• Instrument Name: Perceived Ad Personalization Scale (PAPS)
• Instrument Type: Self-report psychometric questionnaire / Likert-type cognitive scale
• Number of Items: 3 items
• Target Population: Consumers, internet users, and recipients of targeted digital or print communications
• Administration Format: Digital online survey (Qualtrics, SurveyMonkey, MTurk, Prolific) or paper-and-pencil laboratory questionnaire
• Administration Time: Under 1 minute
• Response Scale: 7-point Likert scale (1 = strongly disagree, 7 = strongly agree)
• Scoring Protocol: There are no reverse-coded items. The overall score of perceived ad personalization is computed by calculating the arithmetic mean of all three items:$$\text{PAPS Score} = \frac{\text{Item}_1 + \text{Item}_2 + \text{Item}_3}{3}$$Higher scores represent a higher perceived degree of intentional ad customization and tailoring directed toward the respondent.

11. Permissions & Fee and Test Year

The conceptual framework of computer-tailoring perception was published by Arie Dijkstra in 2005, and its adapted consumer advertising operationalization was published by Elizabeth Aguirre, Dominik Mahr, Dhruv Grewal, Kees de Ruyter, and Martin Wetzels in 2015 in the Journal of Marketing.

The scale is considered an open-access academic instrument. It may be used free of charge for non-commercial academic research, scientific inquiry, doctoral dissertations, and educational evaluations without explicit written permission from the authors, provided appropriate scholarly attribution is cited. Commercial entities, market research platforms, or organizations seeking to embed the scale into proprietary commercial software suites should consult the copyright policies of the American Marketing Association (AMA) or Oxford University Press regarding original publication permissions.

12. References

• Aguirre, E., Mahr, D., Grewal, D., de Ruyter, K., & Wetzels, M. (2015). Unraveling the Personalization Paradox: The Effect of Information Collection and Delivery on Online Advertising Effectiveness. Journal of Marketing, 79(5), 34–49. https://doi.org/10.1509/jm.13.0320
• Brehm, J. W. (1966). A Theory of Psychological Reactance. Academic Press.
• Dijkstra, A. (2005). Working mechanisms of computer-tailored health education: Evidence from smoking cessation. Health Education Research, 20(5), 527–539. https://doi.org/10.1093/her/cyh012
• 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
• Petty, R. E., & Cacioppo, J. T. (1986). The Elaboration Likelihood Model of Persuasion. Advances in Experimental Social Psychology, 19, 123–205. https://doi.org/10.1016/S0065-2601(08)60214-2
• Van Doorn, J., & Hoekstra, J. C. (2013). Customization of online advertising: The role of privacy concerns and trust. Journal of Interactive Marketing, 27(4), 264–274. https://doi.org/10.1016/j.intmar.2013.09.006

13. Items of the Scale