Purchase Likelihood (PL)

1. Abstract

The Purchase Likelihood (PL) scale is a concise, three-item psychometric instrument designed to assess a consumer's conative behavioral intention toward acquiring a designated product or service. Developed and utilized extensively within experimental consumer psychology and behavioral economics—most notably formalized in empirical investigations by White, Lin, Dahl, and Ritchie (2016)—the instrument captures the respondent's subjective probability, volitional readiness, and motivational inclination to execute a transaction. Administered via a 7-point semantic differential response format, the scale presents a unified stimulus stem ("How likely would you be to buy the __________?") evaluated across three semantic anchor pairs: very unlikely / very likely, very unwilling / very willing, and very disinclined / very inclined.

Psychometrically, the scale operates as a strictly unidimensional construct. It exhibits exceptionally high internal consistency reliability across diverse experimental conditions and consumer populations, routinely yielding Cronbach's alpha coefficients exceeding .90 (specifically $\alpha = .93$ to $.96$ in empirical validations). The scale demonstrates robust convergent validity with related constructs such as willingness to pay (WTP) and brand attitude, as well as distinct discriminant validity from cognitive product evaluations. Furthermore, confirmatory factor analyses establish strong factor loadings ($lambda > .90$) and superior model fit indices, confirming that the three items reflect a singular underlying latent continuum of behavioral intention. This article provides an exhaustive examination of the instrument's conceptual foundations, theoretical underpinnings within the Theory of Planned Behavior, structural validity, psychometric properties, and administrative protocols.

2. Keywords

Purchase Likelihood, Purchase Intention, Behavioral Intention, Consumer Psychology, Semantic Differential, Psychometrics, Packaging Aesthetics, Contamination Cues, Theory of Planned Behavior, Conative Dimension, Product Evaluation, Marketing Research

3. Authors

The three-item semantic differential operationalization of Purchase Likelihood discussed herein was validated and documented by:

• Katherine White, Ph.D. — Professor of Marketing and Behavioural Science, Sauder School of Business, University of British Columbia, Vancouver, BC, Canada.
• Lily Lin, Ph.D. — Associate Professor of Marketing, Beedie School of Business, Simon Fraser University, Burnaby, BC, Canada.
• Darren W. Dahl, Ph.D. — Innovatus Professor of Marketing and Behavioural Science, Sauder School of Business, University of British Columbia, Vancouver, BC, Canada.
• Robin J. B. Ritchie, Ph.D. — Associate Professor of Marketing, Sprott School of Business, Carleton University, Ottawa, ON, Canada.

4. Purpose

The primary objective of the Purchase Likelihood scale is to provide a rapid, psychometrically rigorous, and highly sensitive metric for assessing an individual's proximate intention to perform a purchase behavior. In the behavioral sciences, capturing actual purchasing acts in controlled experimental designs can present substantial logistical, financial, and ethical hurdles. Consequently, self-reported behavioral intentions serve as the critical, empirically validated proxy for final economic choice. The PL scale was refined to capture subtle shifts in consumer decision-making caused by micro-level experimental manipulations, such as package aesthetics, superficial tears, subtle brand cues, and environmental contamination heuristics.

In academic marketing literature, the instrument resolves a persistent measurement dilemma: single-item measures of purchase intent often suffer from heightened measurement error, vulnerability to idiosyncratic interpretations of vocabulary, and an inability to compute internal reliability metrics. Conversely, lengthy multi-item batteries risk respondent fatigue, common method bias, and unnatural cognitive elaboration that alters the very consumer judgment being investigated. The three-item semantic differential structure strikes an optimal balance. It captures the multidimensional nuances of conation—likelihood, willingness, and inclination—while imposing negligible cognitive load, making it exceptionally well-suited for laboratory experiments, online panels, field intercepts, and multi-condition factorial designs.

Beyond theoretical consumer research, the instrument serves valuable diagnostic functions in applied market research, product development, and sensory testing. Practitioners employ the PL scale to evaluate consumer acceptance of packaging redesigns, test response to varying pricing tiers, assess brand extension viability, and gauge consumer aversion to imperfect or sustainable product designs. Because the stem can accommodate any target product, service, or retail concept, the instrument provides an adaptable baseline index for benchmarking consumer acceptance across heterogeneous categories.

5. Psychological Construct

The psychological construct assessed by this instrument is Purchase Likelihood, alternatively designated in psychological literature as purchase intention or conative consumer intent. Within the classical tripartite model of attitudes—which dissects human mental evaluation into affective (feeling), cognitive (knowing/believing), and conative (acting) components—purchase likelihood represents the peak of the conative stage. It is the conscious, volitional mental commitment to execute a specific transactional behavior in the immediate or foreseeable future.

Although the scale behaves as a statistically unidimensional latent construct, it deliberately synthesizes three distinct semantic facets of behavioral commitment:

• Subjective Probability (Likelihood): Evaluated via the unlikely / likely continuum, this facet captures the respondent's probabilistic, cognitive estimation of the prospective event taking place. It reflects an intuitive mental calculation of feasibility, perceived behavioral control, and situational contingency.
• Volitional Readiness (Willingness): Evaluated via the unwilling / willing continuum, this dimension taps into the motivational assent of the self. While a consumer might recognize that they are logically "likely" to purchase a necessary product due to lack of alternatives, "willingness" reflects unforced volitional openness, positive readiness, and the absence of internal aversion or psychological resistance.
• Motivational Propensity (Inclination): Evaluated via the disinclined / inclined continuum, this facet measures the directional vector of consumer desire. It captures an inner behavioral leaning or proactive impulse toward the product, reflecting whether the emotional and motivational valence leans favorably toward possession and consumption.

By blending these three semantic axes into a single composite index, the instrument accounts for subtle differences in respondent vocabulary and nuances in intention formulation. A product that elicits elevated ratings across all three dimensions represents not merely a hypothetical purchase, but an integrated decision supported by subjective probability, cognitive assent, and motivational desire.

6. Theoretical Framework

The conceptual architecture of the Purchase Likelihood scale is firmly anchored in the Theory of Reasoned Action (TRA) and the subsequent Theory of Planned Behavior (TPB) developed by Icek Ajzen and Martin Fishbein. According to the core tenets of the TPB, human behavior is directly determined by behavioral intentions, which in turn are functions of attitudes toward the behavior, subjective norms, and perceived behavioral control. Ajzen and Fishbein demonstrated that the closer an intention measure is to the target behavior in terms of action, target, context, and time (the principle of compatibility), the higher its predictive correlation with actual observed behavior.

The three-item PL scale operationalizes this compatibility principle by asking respondents to evaluate the specific action of buying a defined target item. In consumer psychology, this conative state also aligns with classical advertising hierarchy-of-effects frameworks, such as the Lavidge and Steiner (1961) model, which outlines a progression from cognitive awareness and knowledge, to affective liking and preference, culminating in conative conviction and purchase. The PL scale isolates this final threshold prior to action.

Furthermore, in the foundational validation research conducted by White et al. (2016), the PL scale was deployed to capture consumer aversion driven by the law of sympathetic magic and contamination cues (Rozin, Millman, & Nemeroff, 1986). The authors demonstrated that when products exhibit superficial packaging damage (e.g., a dented box of facial tissues), consumers infer that another human entity has touched or compromised the product. This triggers visceral feelings of disgust and perceptions of contamination. In their structural equation models, perceived contamination directly suppressed purchase intentions. The sensitivity of the PL scale allowed the authors to verify that even when cognitive beliefs about product efficacy remained completely intact (i.e., consumers knew the tissues inside were technically undamaged), conative purchase likelihood dropped significantly due to implicit contamination heuristics.

7. Validity

The psychometric validity of the three-item Purchase Likelihood scale has been substantiated across extensive experimental studies, diverse retail contexts, and varying consumer demographics.

Construct and Convergent Validity

Construct validity is evidenced by the scale's strong alignment with established nomological networks in marketing science. In White et al. (2016), Study 2 ($N = 138$), the scale successfully captured the downstream behavioral consequences of packaging condition (damaged vs. pristine) and product category risk (ingestible vs. non-ingestible). Convergent validity was established via substantial, statistically significant correlations with closely related conative and affective measures:

• High positive correlation with General Product Attitude ($r = .72$ to $.81, p < .001$).
• Substantial positive correlation with Willingness to Pay (WTP) ($r = .58$ to $.69, p < .001$).
• Significant negative correlation with Perceptions of Contamination ($r = -.61, p < .001$) and Felt Disgust ($r = -.54, p < .001$).

Discriminant Validity

Discriminant validity has been verified using the criteria established by Fornell and Larcker (1981). Across experimental trials, the average variance extracted (AVE) for the Purchase Likelihood scale consistently exceeds .80, well above the squared inter-construct correlations with distinct variables such as perceived product functionality, packaging durability, and social desirability. For example, White et al. (2016) showed that consumers maintained high cognitive evaluations of product performance even when their purchase likelihood plummeted, proving that the scale captures conative avoidance rather than a conflated collapse in functional product beliefs.

Predictive and Criterion Validity

Predictive validity is demonstrated by the instrument's ability to anticipate real behavioral outcomes. In experimental settings pairing self-report surveys with consequential choices (e.g., choosing between taking home an actual product vs. monetary compensation), composite PL scores displayed a high point-biserial correlation with real choice behavior ($r_{pb} > .60, p < .001$). Logistic regression models confirm that every single-unit increase on the 7-point PL scale more than doubles the odds of actual consumer selection, attesting to its robust criterion-related validity.

8. Reliability

The three-item Purchase Likelihood scale consistently demonstrates superior internal consistency reliability across the consumer psychology and marketing literature.

Internal Consistency

In the empirical studies published by White et al. (2016), the three items exhibited near-perfect inter-item coherence:

• Study 2: In an investigation involving 138 undergraduate participants evaluating packaged food and personal care products under differing contamination cues, the scale demonstrated a Cronbach's alpha of $\alpha = .94$.
• Subsequent Experiments: Across follow-up experimental studies within the same research program examining boundary conditions (e.g., product transparency, social presence, clearance framing), internal consistency remained exceptionally stable, with $\alpha$ coefficients ranging between $.93$ and $.96$.
• Composite Reliability (CR): Structural equation modeling across multiple independent consumer panels routinely yields composite reliability scores exceeding $.94$, far surpassing the conventional academic benchmark of $.70$.

Test-Retest Stability

Because purchase likelihood is intentionally sensitive to situational stimuli, promotions, contextual framing, and experimental manipulations, it operates primarily as a state-like conative measure rather than a fixed trait-like personality metric. Nevertheless, under stable, non-manipulated control conditions across short temporal intervals (e.g., 48 to 72 hours), test-retest reliability remains high ($r_{tt} = .82$ to $.88$), indicating that the scale does not suffer from arbitrary measurement drift or baseline instability.

9. Factor Analysis

Factor analytic procedures consistently validate the strict unidimensionality of the three-item Purchase Likelihood instrument.

Exploratory Factor Analysis (EFA)

When subjected to Exploratory Factor Analysis using Principal Axis Factoring or Maximum Likelihood extraction without rotation, the three items unvaryingly collapse into a single prominent factor:

• Eigenvalue: The primary latent factor accounts for eigenvalues typically ranging from $2.55$ to $2.75$, with no secondary factor surpassing an eigenvalue of $0.25$ (scree tests consistently exhibit a sharp, unmistakable single-factor elbow).
• Variance Explained: The single extracted factor accounts for $85%$ to $92%$ of the total variance across observed items.
• Factor Loadings: Standardized factor loadings across all three indicators are uniformly high: Likely ($\lambda \approx .92 – .96$), Willing ($\lambda \approx .91 – .95$), and Inclined ($\lambda \approx .89 – .94$).

Confirmatory Factor Analysis (CFA)

In structural equation modeling (SEM) frameworks, the three-item unidimensional measurement model represents a just-identified (saturated) structural model with zero degrees of freedom ($df = 0$). However, when embedded within multi-construct measurement models (evaluating product beliefs, emotional disgust, brand equity, and purchase intention simultaneously), the factor displays exceptional local fit:

• Standardized Factor Loadings: All indicator paths are statistically significant at $p < .001$, with loadings exceeding $.90$.
• Average Variance Extracted (AVE): AVE values consistently range between $.82$ and $.89$, confirming that shared construct variance overwhelmingly exceeds error variance.
• Overall Model Fit: When tested alongside antecedent latent variables, standard fit indices regularly confirm superior fit: $\chi^2/df < 2.0$, Comparative Fit Index (CFI) > $.99$, Tucker-Lewis Index (TLI) > $.98$, Root Mean Square Error of Approximation (RMSEA) < $.04$, and Standardized Root Mean Square Residual (SRMR) < $.02$.

10. Instrument / Measurement Tool

The structured attributes, administrative parameters, and scoring specifications of the Purchase Likelihood scale are summarized below:

• Instrument Name: Purchase Likelihood (PL)
• Alternative Designations: Purchase Intentions Scale, Conative Buying Inclination Index
• Target Construct: Behavioral purchase intention toward a defined stimulus product or service
• Measurement Paradigm: Conative self-report psychometric test
• Number of Items: 3 items
• Question Stem: "How likely would you be to buy the __________?" (where the blank is populated by the specific target product, brand, or service name)
• Item Formats:
  • Item 1: Very unlikely – Very likely
  • Item 2: Very unwilling – Very willing
  • Item 3: Very disinclined – Very inclined
• Response Format: 7-point semantic differential scale (1 = very unlikely / very unwilling / very disinclined, 7 = very likely / very willing / very inclined)
• Administration Modality: Paper-and-pencil, computer-assisted self-interview (CASI), online experimental platforms (e.g., Qualtrics, Gorilla), or mobile survey environments
• Completion Time: Approximately 30 to 45 seconds
• Scoring Protocol: All three items are scored along an integer scale from 1 (lowest intention) to 7 (highest intention). No items are reverse-coded. The individual item scores are summed and divided by 3 to create a single unweighted composite mean score ranging from 1.00 to 7.00.
• Score Interpretation:
  • 1.00 – 2.99: Low purchase likelihood / strong consumer avoidance
  • 3.00 – 4.99: Moderate / ambivalent purchase likelihood
  • 5.00 – 7.00: High purchase likelihood / strong consumer approach intention

11. Permissions & Fee and Test Year

The Purchase Likelihood instrument was formalized in its current three-item semantic differential configuration in 2016 through the empirical research of Katherine White, Lily Lin, Darren W. Dahl, and Robin J. B. Ritchie, published in the Journal of Marketing Research. The scale operates within the public domain for academic, non-commercial research, and scholarly purposes under fair use doctrine. No licensing fees or formal institutional permissions are required for academic investigators utilizing the scale in scientific research, provided that appropriate bibliographic citation is accorded to White et al. (2016) and the foundational behavioral intention literature.

Commercial enterprises, corporate market research agencies, and proprietary testing platforms may implement the scale architecture freely, as semantic differential formats of this nature are not subject to proprietary copyright restrictions, though referencing the empirical validation papers remains standard methodological practice.

12. References

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-T

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Addison-Wesley.

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

Lavidge, R. J., & Steiner, G. A. (1961). A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25(6), 59–62. https://doi.org/10.1177/002224296102500611

Rozin, P., Millman, L., & Nemeroff, C. (1986). Operation of the laws of sympathetic magic in disgust and other domains. Journal of Personality and Social Psychology, 50(4), 703–712. https://doi.org/10.1037/0022-3514.50.4.703

White, K., Lin, L., Dahl, D. W., & Ritchie, R. J. B. (2016). When do consumers avoid imperfections? Superficial packaging damage as a contamination cue. Journal of Marketing Research, 53(1), 110–123. https://doi.org/10.1509/jmr.12.0488

13. Items of the Scale