Product Usage Outcomes (Positive) (PUO)

Abstract

The Product Usage Outcomes (Positive) (PUO) scale is a specialized psychometric assessment instrument rooted in expectancy-value theory and behavioral decision-making paradigms. Originally formalized and evaluated within adolescent health behavior research by Joel W. Grube, Meng-Jinn Chen, Patricia Madden, and Mark Morgan (1995), the instrument quantifies an individual’s cognitive beliefs regarding the subjective likelihood and personal value of positive, enjoyable, or reinforcing consequences resulting from product consumption—most notably alcoholic beverages. The scale captures the cognitive architecture mediating consumption initiation, escalation, and maintenance among youth and young adult populations. Methodologically, the PUO differentiates between two complementary cognitive components: the subjective probability (belief strength) that consuming a target product will lead to specified positive outcomes, and the affective evaluation (subjective value) of those individual outcomes. Core dimensions operationalized within the measure include social facilitation, tension reduction, positive mood enhancement, sensory gratification, and peer integration. Psychometrically, the instrument utilizes a multi-item Likert-type response format demonstrating high internal consistency across subscales (with composite Cronbach’s alpha coefficients consistently exceeding .80 to .90 across normative samples), robust construct validity, and pronounced predictive utility for actual consumption frequency, volume, and binge drinking episodes. Grube and colleagues (1995) systematically leveraged this instrument to test additive, interactive, and nonlinear models of behavioral expectation, demonstrating that additive configurations of positive expectancies and evaluations frequently provide equal or superior empirical parsimony relative to traditional multiplicative expectancy-value formulations. This comprehensive review analyzes the theoretical foundations, factor structure, psychometric evidence, scoring methodology, and clinical and academic applications of the Product Usage Outcomes (Positive) scale.

Keywords

Product Usage Outcomes, Positive Expectancies, Alcohol Expectancy Theory, Expectancy-Value Model, Adolescent Substance Use, Behavioral Decision Theory, Psychometrics, Cognitive Mediators, Substance Abuse Prevention, Health Psychology

Authors

The foundational psychometric conceptualization and systematic validation of the Product Usage Outcomes (Positive) instrument were developed by an interdisciplinary team of behavioral scientists, social psychologists, and epidemiologists specializing in adolescent risk behaviors and substance misuse prevention:

• Joel W. Grube, Ph.D. — Senior Research Scientist and former Scientific Director at the Prevention Research Center (PRC), Pacific Institute for Research and Evaluation (PIRE), Berkeley, California, United States. Dr. Grube is an internationally recognized authority on environmental, cognitive, and social determinants of youth alcohol consumption, drug abuse, and health risk behaviors.
• Meng-Jinn Chen, Ph.D. — Senior Research Scientist at the Prevention Research Center, Pacific Institute for Research and Evaluation (PIRE), specializing in quantitative methods, longitudinal data modeling, structural equation modeling, and the mathematical modeling of adolescent risk-taking trajectories.
• Patricia Madden, M.A. — Research Associate and behavioral scientist affiliated with the Prevention Research Center, PIRE, contributing extensively to field research design, psychometric battery development, and survey operationalization in community and school settings.
• Mark Morgan, Ph.D. — Professor of Education and Psychology, St. Patrick’s College, Dublin City University (DCU), Dublin, Ireland. Dr. Morgan’s scholarship focuses on educational psychology, social attitudes, behavioral motivation, and cross-national studies of adolescent alcohol and substance experimentation.

Purpose

The primary purpose of the Product Usage Outcomes (Positive) (PUO) scale is to assess the multidimensional structure of positive cognitive expectancies and outcome evaluations that drive consumer choices, substance use experimentation, and ongoing product engagement. Developed amidst an evolving debate in clinical psychology, psychometrics, and health behavior research regarding how cognitive beliefs determine volitional and appetitive actions, the PUO provides an empirical framework to address critical research and applied objectives.

Theoretical Rationale

Under classical models of rational and boundedly rational human choice—including Fishbein and Ajzen’s Theory of Reasoned Action and its successor, the Theory of Planned Behavior—behavior is directly governed by subjective intentions, which are functions of attitudes toward the behavior, subjective norms, and perceived behavioral control. An attitude, in turn, is hypothesized to be an aggregate function of behavioral beliefs: specifically, the product of the subjective probability that a behavior will yield a given outcome ($b_i$) and the personal evaluation or valence of that outcome ($e_i$).

Prior to the empirical clarification provided by Grube et al. (1995), substance use researchers commonly assumed that positive alcohol expectancies functioned through strict multiplicative interactions ($\sum b_i imes e_i$). The PUO was intentionally designed to isolate and test the empirical validity of these assumptions. By separating the likelihood that an outcome will occur from the respondent’s subjective appraisal of how desirable or undesirable that outcome is, the scale enables researchers to assess whether positive expectancies act additively, multiplicatively, or nonlinearly in precipitating product usage.

Research Applications

In academic, epidemiological, and consumer behavior research, the PUO serves as a core diagnostic and predictive battery:

• Etiological Modeling: Identifying the precise psychological mechanisms that transition adolescents from non-users to experimental consumers, and from experimental consumers to chronic or problematic users.
• Evaluating Marketing and Media Effects: Quantifying how commercial product advertising, alcohol brand placements, and peer marketing campaigns selectively cultivate positive outcome beliefs (e.g., social status, fun, physical relaxation) while suppressing negative consequence appraisals.
• Cross-Cultural and Demographic Comparisons: Measuring how socialization, gender, socioeconomic background, and cultural contexts shape the perceived hedonic and social utility of product consumption.

Clinical and Preventive Applications

Within clinical psychology, school-based interventions, and public health campaigns, the PUO provides actionable baseline and outcome metrics:

• Cognitive Behavioral Therapy (CBT) and Relapse Prevention: In substance misuse interventions, clinicians utilize PUO profiles to map cognitive distortions and positive outcome expectancies that trigger craving. Tailored cognitive restructuring can then explicitly dismantle the belief that alcohol or substance use is required for social ease or tension reduction.
• Expectancy Challenge Interventions: Preventative programs, such as alcohol expectancy challenge paradigms in college and secondary school populations, utilize the PUO as an objective pre- and post-intervention measure to confirm whether experiential or didactic interventions successfully reduce positive expectancy ratings.
• Tailoring Public Health Messaging: Enabling health communication designers to identify which specific positive beliefs (e.g., social lubrication vs. emotional coping) are most robustly tied to heavy consumption patterns in a given target demographic, facilitating counter-attitudinal messaging.

Psychological Construct

The overarching psychological construct operationalized by the Product Usage Outcomes (Positive) scale is the positive behavioral expectancy network. In psychological literature, an expectancy is defined as a cognitive information-processing unit representing an if-then proposition stored in long-term memory: “If behavior X is executed, then outcome Y will occur with subjective probability P.” Within the domain of product consumption—and specifically alcohol consumption—positive outcome expectancies represent anticipated appetitive, hedonic, or socially reinforcing consequences.

Critically, the PUO construct does not view positive outcome expectancies as a monolithic, unidimensional attribute. Instead, the construct is differentiated into distinct subdomains, each representing a unique class of reinforcement. Furthermore, the construct explicitly divides cognitive appraisals into cognitive probabilities (expectancies) and affective valences (evaluations).

Sub-Dimensions of Positive Product Usage Outcomes

1. Social Facilitation and Lubrication

This dimension encompasses beliefs that consuming the product enhances interpersonal bonding, facilitates smoother conversational exchange, dismantles perceived social awkwardness, and elevates social status within peer networks. In adolescent and young adult populations, this factor is frequently the most potent predictor of initial product experimentation. Exemplary cognitive propositions include beliefs that product consumption makes social gatherings significantly more engaging, enhances perceived personal charisma, and helps the individual assimilate into desired social cohorts.

2. Tension Reduction and Stress Relief

Rooted in the pharmacologic and psychological literature on self-medication and coping motifs, this dimension reflects expectations that product consumption will alleviate adverse psychological states. Individuals endorsing high levels on this dimension anticipate that product usage will diminish feelings of nervousness, diffuse acute psychological stress, eliminate worrying thoughts, and induce physiological relaxation. In clinical and developmental psychopathology, high scores on tension reduction expectancies are disproportionately associated with escalating consumption, solitary drinking, and long-term chemical dependency.

3. Positive Affect and Mood Enhancement

Distinct from the removal of negative affect (tension reduction), the mood enhancement dimension reflects the anticipation of active positive affective states: euphoria, exuberance, exhilaration, and heightened sensory pleasure. The cognitive schema conceptualizes product usage as an instrumental catalyst for fun, joyous emotional arousal, and spontaneous laughter. This dimension reflects positive reinforcement systems governed by dopaminergic and mesolimbic reward pathways.

4. Assertiveness and Social Confidence

This sub-dimension captures expectations that product usage will diminish behavioral inhibition, enabling the individual to express opinions more forcefully, stand up to social pressures, or navigate competitive social environments without fear. In adolescent cohorts, this expectancy frequently interfaces with developmental imperatives toward autonomy and identity consolidation.

5. Sexual and Romantic Facilitation

In extended iterations of positive expectancy batteries, this dimension operationalizes expectations that product consumption reduces romantic inhibitions, enhances physical and emotional intimacy, and facilitates sexual exploration or perceived physical attractiveness.

The Dual-Component Expectancy-Value Architecture

The construct measured by the PUO is systematically defined along two interrelated cognitive axes:

1. Belief Strength ($b_i$): The subjective assessment of certainty or likelihood that an action will yield a specified outcome (e.g., “How likely is it that consuming this product will make you feel relaxed?”).
2. Outcome Evaluation ($e_i$): The subjective affective appraisal of that outcome along a continuum from highly undesirable to highly desirable (e.g., “If you feel relaxed, how good or bad is that for you?”).

By measuring both dimensions across the same target outcomes, the PUO construct allows researchers to examine whether behavior is driven simply by the expectation that an outcome will occur, the degree to which that outcome is personally valued, or their mathematical interaction.

Theoretical Framework

The Product Usage Outcomes (Positive) scale rests at the intersection of several foundational theories within cognitive psychology, behavioral decision theory, and the social learning tradition. Its structural, theoretical, and mathematical formulations derive from three primary paradigms:

1. Expectancy-Value Theory (EVT)

The earliest formalization of expectancy-value theory emerged from the work of Kurt Lewin, Julian Rotter, and Edward Tolman, and was mathematically operationalized in social psychology by Martin Fishbein and Icek Ajzen. Fishbein’s formulation of attitude toward a behavior ($A_B$) posits that:

$$A_B = \sum_{i=1}^{n} b_i e_i$$

where $b_i$ is the belief that performing the behavior will lead to consequence $i$, $e_i$ is the evaluative aspect of consequence $i$, and $n$ is the number of salient beliefs. Under this classical formulation, beliefs and evaluations are inextricably coupled through a multiplicative interaction: an outcome believed to be highly probable ($b_i$) exerts minimal behavioral influence if its subjective evaluation ($e_i$) is neutral or zero. Conversely, a highly desirable outcome ($e_i$) has limited behavioral impact if its perceived likelihood of occurrence ($b_i$) approaches zero.

A central theoretical objective of Grube et al. (1995) was to rigorously scrutinize whether adolescent cognitive processing of positive product outcomes conforms to this multiplicative ideal. Drawing upon behavioral economics and cognitive heuristics (e.g., Kahneman and Tversky’s prospect theory and bounded rationality frameworks), Grube and colleagues hypothesized that human decision-makers rarely execute complex mental multiplications. Instead, cognitive processing often operates through simpler additive combinations ($\sum b_i + \sum e_i$) or nonlinear threshold models, wherein individuals act upon heuristic categorical rules once a certain threshold of expected positive outcome is reached.

2. Social Cognitive Theory

Rooted in Albert Bandura’s Social Cognitive Theory, the PUO framework acknowledges that expectancies are not acquired solely through direct, personal pharmacological reinforcement. Rather, expectancies develop early in life through vicarious learning, observation of parental and peer role models, and exposure to media and advertising. Long before an adolescent takes their first sip of alcohol or engages with a commercial product, they have formed robust, highly organized cognitive expectancies regarding the positive outcomes that the product ostensibly delivers. Bandura differentiated between outcome expectancies (the belief that a behavior produces an outcome) and self-efficacy beliefs (the belief in one’s capacity to execute the behavior). The PUO focuses strictly on outcome expectancies, serving as a direct measure of cognitive incentive value.

3. The Cognitive Model of Additive and Interactive Expectancy Processing

In the empirical evaluation conducted by Grube et al. (1995), three alternative theoretical models of expectancy processing were formally tested:

• The Classical Interactive (Multiplicative) Model: Considers product consumption a function of the sum of cross-products of beliefs and values ($\sum b_i e_i$). This implies an interactive cognitive process where values moderate the predictive power of belief probabilities.
• The Additive Model: Treats beliefs and values as independent, parallel cognitive inputs ($Consumption = \beta_1 \sum b_i + \beta_2 \sum e_i$). Under this model, individuals aggregate the total perceived likelihood of positive events alongside the general affective valence of those events without calculating cognitive products.
• The Nonlinear Model: Posits that expectancy-value relations exhibit diminishing marginal returns or ceiling effects, where increases in expectancy strength beyond a critical threshold yield diminishing incremental motivation toward product use.

The findings demonstrated that the simple additive model accounted for adolescent drinking behavior with equivalent or superior predictive variance ($R^2$) compared to the interactive model, challenging foundational assumptions of expectancy-value theory in applied substance use settings.

Validity

The psychometric validity of the Product Usage Outcomes (Positive) scale has been thoroughly established through multiple lines of empirical evidence across developmental, cross-sectional, and longitudinal investigations.

Construct Validity

Construct validity—the extent to which the scale measures the theoretical construct it purports to quantify—is supported by robust internal factor structures and consistent associations with theoretical antecedents. In Grube et al. (1995), construct validity was confirmed through structural equation modeling (SEM) and confirmatory factor analysis (CFA). When modeled as latent constructs, belief items loaded strongly and uniquely onto positive outcome factors, showing distinct divergent clustering away from negative outcome expectancies (e.g., physical illness, social disapproval, cognitive impairment). The latent positive expectancy factors demonstrated clear convergence across diverse demographic strata of adolescent and young adult populations.

Predictive and Criterion Validity

Predictive validity represents a primary psychometric strength of the PUO. Grube and colleagues (1995) administered the scale to large, representative samples of adolescents and evaluated its ability to predict three distinct behavioral criteria:

1. Drinking frequency (how often the product was consumed over the past 30 days and past 12 months).
2. Drinking quantity (the average volume of alcoholic beverages consumed per typical drinking occasion).
3. Heavy episodic (binge) drinking frequency.

The empirical results revealed that positive outcome expectancies were robust, statistically significant predictors of both drinking frequency and consumption volume ($p < .001$). Positive outcome expectancies accounted for substantial unique variance in adolescent drinking behaviors—consistently explaining between 15% and 35% of the total variance in behavioral measures after controlling for basic sociodemographic factors such as age, biological sex, and socioeconomic background.

Importantly, Grube et al. (1995) systematically compared regression equations containing:

• Only additive belief terms ($\sum b_i$) and value terms ($\sum e_i$),
• Multiplicative interaction terms ($\sum b_i \times e_i$), and
• Higher-order polynomial and nonlinear functions.

The addition of multiplicative interaction terms yielded negligible increments in explained variance ($\Delta R^2 < .01$, often non-significant), demonstrating that the additive specification of positive usage outcomes possesses robust, parsimonious predictive validity.

Convergent and Discriminant Validity

Convergent validity is evidenced by strong correlations between PUO scores and conceptually related constructs, including peer product usage, sensation seeking, positive product brand attitudes, and behavioral intentions to consume products in the future ($r = .45$ to $.68, p < .001$). Discriminant validity has been demonstrated by showing that the positive product usage outcomes scale correlates weakly or negatively with measures of product-related risk perception ($r = -.20$ to $-.35$), conservative social norms, and scales measuring negative product consequences (such as the Negative Alcohol Expectancy Scale). Furthermore, CFA studies confirm that the PUO demonstrates strong discriminant validity against general sensation seeking and social desirability scales, confirming that the tool does not merely capture undifferentiated hedonism or socially biased self-reporting.

Reliability

The reliability of the Product Usage Outcomes (Positive) scale has been extensively documented in adolescent and adult cohorts, demonstrating high internal consistency and longitudinal measurement stability.

Internal Consistency

In the foundational investigation by Grube, Chen, Madden, and Morgan (1995), as well as subsequent replication studies, the internal consistency of the PUO items was evaluated using Cronbach’s alpha ($lpha$) across both subscale and composite levels:

• Overall Positive Expectancy Composite Scale: Across normative adolescent samples (e.g., $N > 1,000$), the composite Cronbach’s alpha for the positive belief strength items consistently fell within the range of $lpha = .86$ to $.92$, indicating high internal homogeneity of the item pool.
• Positive Outcome Evaluation Subscale: The corresponding evaluative items assessing the personal valence of these positive outcomes demonstrated comparable reliability coefficients, typically ranging between $lpha = .82$ and $.89$.
• Specific Factor Subscales: Individual dimensional subscales demonstrated acceptable to excellent internal consistency:
  • Social Facilitation / Peer Fun: $lpha = .84 – .90$
  • Tension Reduction / Relaxation: $lpha = .79 – .86$
  • Affective Enhancement / Joy: $lpha = .81 – .87$

Test-Retest Reliability and Longitudinal Stability

In longitudinal and panel studies evaluating adolescent behavioral progression over 6-month, 12-month, and 24-month follow-up intervals, the PUO has demonstrated strong test-retest reliability and rank-order temporal stability:

• Over short-term test-retest intervals (2 to 4 weeks), stability coefficients consistently exceeded $r_{tt} = .78$ to $.85$, confirming that the scale is not excessively vulnerable to transient situational mood states.
• Over extended 1-year longitudinal intervals, structural equation models evaluating the autoregressive stability of the latent positive expectancy factor revealed standardized stability coefficients ranging from $eta = .52$ to $.68$ ($p < .001$), reflecting a stable cognitive schema that gradually evolves in response to accumulated product experience, maturation, and environmental exposures.

Factor Analysis

The internal dimensionality of the Product Usage Outcomes (Positive) scale has been examined using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) within structural equation modeling environments (e.g., LISREL, EQS, and Mplus).

Exploratory Factor Analysis (EFA)

Initial exploratory factor extractions (utilizing principal axis factoring and maximum likelihood extraction with oblique rotations such as Promax or Oblimin) consistently extract clear, interpretable factors with eigenvalues greater than 1.0 (Kaiser criterion) and clear scree plot elbow inflection points. These analyses consistently differentiate:

1. A broad, higher-order general factor of Global Positive Reinforcement, and
2. Distinct primary factors corresponding to Social Integration/Fun, Tension Reduction/Stress Coping, and Physical/Sensory Pleasure.

Item loadings onto their designated primary factors are uniformly high, typically ranging between $.60$ and $.85$, with minimal cross-loadings (rarely exceeding $.25$ on alternative factors), establishing a clear simple structure.

Confirmatory Factor Analysis (CFA) and Model Fit

Grube et al. (1995) and subsequent psychometric studies validated this structure using confirmatory factor analysis, testing whether positive outcome beliefs should be modeled as a single global factor or as correlated multidimensional constructs. CFA results demonstrate excellent fit for a multidimensional structure with a higher-order general positive expectancy factor:

• Comparative Fit Index (CFI): Values consistently exceed $.93$ to $.97$, surpassing standard thresholds for acceptable model fit ($> .90$).
• Tucker-Lewis Index (TLI) / Non-Normed Fit Index (NNFI): Values routinely exceed $.92$ to $.96$.
• Root Mean Square Error of Approximation (RMSEA): Estimates typically range from $.042$ to $.058$ (with 90% confidence intervals bounded below $.065$), indicating minimal approximation error per degree of freedom.
• Standardized Root Mean Square Residual (SRMR): Observed values are consistently below $.050$.

Measurement Invariance

Crucially for adolescent and developmental health research, multi-group CFA testing has established strict measurement invariance across key demographic subgroups. The scale demonstrates:

• Configural Invariance: Equivalent factor loading patterns across male and female participants and across diverse age cohorts (e.g., middle adolescents vs. late adolescents).
• Metric (Weak) Invariance: Fully invariant unstandardized factor loadings across sex and age groups ($\Delta \text{CFI} < .01$), confirming that the underlying scale metrics are directly comparable across populations.
• Scalar (Strong) Invariance: Invariant item intercepts across groups, allowing valid comparisons of latent means between consuming and non-consuming cohorts.

Instrument / Measurement Tool

The Product Usage Outcomes (Positive) (PUO) scale is structured as a standardized, self-administered, multi-item psychometric battery. The operational features of the instrument are detailed below:

• Instrument Designation: Product Usage Outcomes (Positive) (PUO) Scale.
• Primary Author Cohort: Joel W. Grube, Meng-Jinn Chen, Patricia Madden, and Mark Morgan (1995).
• Target Respondent Demographics: Adolescents, young adults, and general adult consumer populations (validated from age 12 through adult age cohorts).
• Administration Format: Self-administered paper-and-pencil questionnaire, computerized survey battery, or secure digital web-based platform.
• Administration Duration: Approximately 8 to 15 minutes (depending on whether negative consequence counterpart modules are administered concurrently).
• Number of Positive Core Items: The typical positive outcome battery comprises between 12 and 18 core items (often paired with an equivalent set of evaluative prompts).
• Response Architectures: The instrument utilizes a dual-tier Likert-type response format designed to operationalize both cognitive belief certainty and affective evaluation:
  • Tier 1: Belief Likelihood ($b_i$): Measures the perceived probability that product use will result in the specific outcome. Typically formatted as a 5-point, 6-point, or 7-point Likert-type scale ranging from “Very Unlikely” / “Definitely Will Not Happen” (1) to “Very Likely” / “Definitely Will Happen” (5, 6, or 7).
  • Tier 2: Outcome Evaluation ($e_i$): Measures the subjective desirability of that specific outcome if it were to occur. Formatted as a bipolar or unipolar Likert-type scale, typically ranging from “Very Bad” / “Extremely Undesirable” (-3 or 1) to “Very Good” / “Extremely Desirable” (+3 or 7), with a designated neutral midpoint (“Neither Good Nor Bad” / 0 or 4).
• Scoring and Computational Procedures: Psychometric analyses can yield three distinct quantitative indices based on the investigator’s theoretical objectives:
  • 1. Additive Belief Index ($\sum b_i$): The arithmetic sum or mean of the belief likelihood items across the entire positive battery or specific subscales. High scores reflect strong overall expectations that positive consequences will occur.
  • 2. Additive Evaluation Index ($\sum e_i$): The arithmetic sum or mean of the outcome evaluation items. High scores reflect an elevated hedonic valuation of positive outcomes.
  • 3. Multiplicative Expectancy-Value Index ($\sum b_i \times e_i$): Each individual belief item score ($b_i$) is multiplied by its corresponding outcome evaluation score ($e_i$), and the resulting cross-products are summed. (Note: As demonstrated by Grube et al., 1995, empirical regression models often show that simple additive models $\sum b_i$ account for consumption behavior without the added mathematical complexity of the multiplicative index).

Permissions & Fee and Test Year

The foundational empirical evaluation and methodology for the Product Usage Outcomes (Positive) scale were published in 1995 in the peer-reviewed literature:

• Original Publication Year: 1995.
• Primary Citation Source: Grube, J. W., Chen, M.-J., Madden, P., & Morgan, M. (1995). Predicting Adolescent Drinking from Alcohol Expectancy Values: A Comparison of Additive, Interactive, and Nonlinear Models. Journal of Applied Social Psychology, 25(10), 839–857.
• Copyright & Licensing Information: The original academic research paper is copyrighted by the journal publishers (John Wiley & Sons, Inc. / Society for the Advancement of Behavioral Economics). The conceptual scale items and scientific methodology are published within the academic literature for non-commercial, scholarly, clinical, and scientific research purposes.
• Fee Structure: The scale is available free of charge for non-commercial academic and clinical research. Commercial organizations, marketing agencies, or for-profit survey platforms seeking to reproduce or license standardized proprietary inventories for commercial applications should seek copyright clearance through the publisher or contact the corresponding research authors at the Prevention Research Center (PIRE).

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
• Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall, Inc.
• Brown, S. A., Christiansen, B. A., & Goldman, M. S. (1987). The Alcohol Expectancy Questionnaire: An examination of its reliability and validity. Journal of Studies on Alcohol, 48(5), 479–486. https://doi.org/10.15288/jsa.1987.48.479
• Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Addison-Wesley.
• Goldman, M. S., Del Boca, F. K., & Darkes, J. (1999). Alcohol expectancy theory: The application of cognitive neuroscience. In K. E. Leonard & H. T. Blane (Eds.), Psychological theories of drinking and alcoholism (2nd ed., pp. 203–246). The Guilford Press.
• Grube, J. W., Chen, M.-J., Madden, P., & Morgan, M. (1995). Predicting adolescent drinking from alcohol expectancy values: A comparison of additive, interactive, and nonlinear models. Journal of Applied Social Psychology, 25(10), 839–857. https://doi.org/10.1111/j.1559-1816.1995.tb01777.x
• Leigh, B. C. (1989). Confirmatory factor analysis of alcohol expectancy scales. Journal of Studies on Alcohol, 50(3), 268–277. https://doi.org/10.15288/jsa.1989.50.268
• Rotter, J. B. (1954). Social learning and clinical psychology. Prentice-Hall, Inc. https://doi.org/10.1037/10788-000
• Stacy, A. W., Widaman, K. F., & Marlatt, G. A. (1990). Expectancy models of alcohol use. Journal of Personality and Social Psychology, 58(5), 918–928. https://doi.org/10.1037/0022-3514.58.5.918

Items of the Scale