Cannabis Effects Expectancy Questionnaire-Medical (CEEQ-M)

Abstract

The Cannabis Effects Expectancy Questionnaire-Medical (CEEQ-M) is a psychometrically validated, 17-item self-report instrument engineered to assess cognitive expectancies regarding the therapeutic, physiological, and psychological outcomes of medical cannabis use (Weiss et al., 2023). Developed in response to the rapid proliferation of legalized cannabis for therapeutic purposes and the concurrent lack of standardized instruments measuring medical—as opposed to recreational—expectancies, the CEEQ-M provides an empirically grounded framework for evaluating patient beliefs. Derived from an initial 21-item conceptual pool based on foundational epidemiological research in medical cannabis dispensaries (Reinarman et al., 2011), the instrument was refined through a rigorous four-study methodological design involving United States adults initiating or contemplating cannabis therapy for chronic pain, insomnia, major depressive disorder, or generalized anxiety disorders.

Psychometric evaluation through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) established a robust two-factor latent structure comprising Symptom Relief (accounting for 31.0% of the common variance) and Atypical Beliefs (accounting for 13.9% of the common variance). Items are rated on a nuanced 10-point Likert-type response scale ranging from 1 (“strongly disagree”) to 10 (“strongly agree”). The instrument exhibits exemplary internal consistency; the Symptom Relief dimension demonstrates a Cronbachu2019s alpha ($lpha$) of 0.906 and a McDonaldu2019s omega ($\omega$) of 0.907, whereas the Atypical Beliefs dimension demonstrates an $lpha$ of 0.779 and an $\omega$ of 0.772. Crucially, the CEEQ-M represents the first medical cannabis expectancy instrument to demonstrate longitudinal scalar measurement invariance across experimental randomization conditions, sex, broad adult age strata (emerging adults aged 18u201325 versus mature adults aged 26u201365), and repeated assessments across active clinical trial timelines. The CEEQ-M addresses a critical gap in clinical psychometrics, behavioral pharmacology, and psychiatric epidemiology by disentangling pharmacodynamic drug effects from cognitive-placebo mechanisms.

Keywords

Cannabis Effects Expectancy Questionnaire-Medical, CEEQ-M, medical cannabis, outcome expectancies, cognitive expectancy theory, symptom relief, atypical beliefs, psychometric validation, measurement invariance, behavioral pharmacology, chronic pain, clinical assessment

Authors

The Cannabis Effects Expectancy Questionnaire-Medical was conceptualized, developed, and validated by a multidisciplinary team of clinical psychologists, neuroscientists, and addiction medicine specialists at the Massachusetts General Hospital and Harvard Medical School:

• Jakob H. Weiss, Ph.D. u2014 Center for Addiction Medicine, Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA; Department of Psychiatry, Harvard Medical School. ORCID: 0000-0002-3879-0568.
• Brenden Tervo-Clemmens, Ph.D. u2014 Center for Addiction Medicine, Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA; Department of Psychiatry, Harvard Medical School.
• Kevin W. Potter, Ph.D. u2014 Center for Addiction Medicine, Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA; Department of Psychiatry, Harvard Medical School.
• A. Eden Evins, M.D., M.P.H. u2014 Cox Family Professor of Psychiatry in the Field of Addiction Medicine, Harvard Medical School; Director, Center for Addiction Medicine, Massachusetts General Hospital, Boston, MA, USA.
• Jodi M. Gilman, Ph.D. u2014 Associate Professor of Psychiatry, Harvard Medical School; Director of Neuroscience, Center for Addiction Medicine, Massachusetts General Hospital, Boston, MA, USA. ORCID: 0000-0001-5180-6694. Correspondence email: [email protected].

Purpose

The primary clinical and psychometric purpose of the Cannabis Effects Expectancy Questionnaire-Medical (CEEQ-M) is to systematically identify, operationalize, and quantify the cognitive outcome expectancies held by adults regarding the therapeutic efficacy and secondary consequences of medical cannabis administration. Over the past two decades, shifts in global drug policy have expanded access to authorized medical cannabis programs across numerous international jurisdictions. Hundreds of thousands of patients obtain physician certifications to utilize cannabinoids for debilitating conditions, most notably chronic musculoskeletal and neuropathic pain, refractory insomnia, treatment-resistant depressive episodes, and chronic anxiety states. Despite this widespread clinical adoption, empirical evidence demonstrating the targeted pharmacological efficacy of whole-plant cannabis across diverse indications remains heterogeneous and heavily debated within evidence-based medicine.

In pharmacological and psychiatric research, patient-reported outcomes following drug administration reflect a dynamic interplay between active pharmacodynamic properties (e.g., cannabinoid receptor type 1 [$CB_1$] and type 2 [$CB_2$] partial agonism by $\Delta^9$-tetrahydrocannabinol [THC], and non-competitive allosteric modulation by cannabidiol [CBD]) and non-pharmacological factors. Foremost among these non-pharmacological factors are cognitive expectancies: internalized, anticipatory mental models concerning how a substance will affect physiological functioning, affect, cognition, and physical symptoms. Prior to the construction of the CEEQ-M, existing psychometric inventories—such as the Marijuana Effect Expectancy Questionnaire (MEEQ) and the Comprehensive Effects of Marijuana Questionnaire (CEQ)—were developed and normed almost exclusively on adolescent, collegiate, or recreational samples. These legacy instruments emphasize hedonic, recreational, or socially deviant outcomes, such as cognitive and behavioral impairment, social and sexual facilitation, relaxation, tension reduction, and perceptual enhancement.

Consequently, recreational instruments demonstrate severe conceptual misalignment, construct underrepresentation, and floor/ceiling artifacts when applied to medical patient populations. Medical cannabis patients are demographically distinct: they are generally older, frequently drug-nau00efve or re-initiating use after decades of abstinence, report lower baseline desires for intoxicating psychoactive states, and evaluate the substance primarily through the lens of symptom management, functional restoration, and medical harm reduction. The CEEQ-M was engineered to resolve these psychometric limitations by providing a reliable, standardized scale designed specifically for clinical environments, randomized controlled trials (RCTs), and epidemiological research. It enables clinicians and clinical investigators to:

• Quantify the magnitude of anticipated therapeutic symptom alleviation prior to the initiation of cannabinoid therapy;
• Detect idiosyncratic, scientifically unsubstantiated, or exaggerated beliefs regarding cannabis as a panacea or “cure-all” (atypical beliefs);
• Isolate and control for expectancy-driven placebo response rates within clinical trials of cannabinoid-based medicinal products;
• Track the longitudinal stability or updating of cognitive expectancies as patients transition from cannabis-nau00efve states to chronic therapeutic administration; and
• Identify cognitive risk factors predictive of subsequent escalation to Cannabis Use Disorder (CUD) versus sustained medical adherence.

Psychological Construct

The core construct assessed by the CEEQ-M is cannabis outcome expectancy within a medical framework. In psychological measurement, an outcome expectancy represents an individual’s explicit cognitive representation of the contingency between a specific behavioral action (here, administering a cannabis preparation) and the resultant physiological, affective, or functional consequences. Cognitive expectancies operate as memory structures or associative schemata within long-term cognitive networks, influencing both the subjective appraisal of interoceptive sensory cues and behavioral choices regarding dosage, frequency of use, and treatment adherence.

The CEEQ-M operationalizes medical cannabis expectancies as a multidimensional construct comprising two discrete, structurally validated, and theoretically distinct factors:

1. Symptom Relief (Factor 1)

The Symptom Relief dimension encapsulates an individual’s direct anticipatory beliefs regarding the therapeutic capacity of cannabis to mitigate specific physical, affective, and somatic complaints. This factor reflects conventional medical motivations for seeking cannabinoid-based therapies. It captures expectancies across four core symptom domains:

• Pain Alleviation and Physical Comfort: Expectancies that cannabis will attenuate chronic musculoskeletal pain, neuropathic discomfort, inflammatory sensations, and muscular spasticity, thereby restoring basic somatic comfort.
• Sleep Induction and Architecture: Beliefs that cannabis functions as a potent hypnotic and sedative, facilitating shortened sleep-onset latency, reduced nocturnal awakenings, and deeper restorative rest.
• Affective Regulation and Anxiolysis: Anticipations that cannabis will alleviate subjective distress, blunt sympathetic nervous system hyperarousal, downregulate persistent state anxiety, and diminish intrusive, ruminative thoughts.
• Mood Elevation and Coping: Cognitive convictions that cannabinoid consumption will counteract depressive anhedonia, enhance subjective emotional valence, and facilitate cognitive distraction from debilitating symptoms.

From a psychometric perspective, individuals scoring high on the Symptom Relief dimension hold strong, explicit convictions that cannabis administration will yield clinically meaningful functional restoration. While moderate-to-high scores are expected among treatment-seeking medical populations, exceptionally elevated scores may indicate unrealistic therapeutic expectations, predisposing patients to profound disappointment or escalation of dose if pharmacodynamic tolerance develops.

2. Atypical Beliefs (Factor 2)

The Atypical Beliefs dimension represents a novel, distinct cognitive construct characterized by idiosyncratic, scientifically unsupported, or highly atypical expectations regarding the effects of cannabis. Unlike mainstream symptom relief expectations—which possess at least partial biological plausibility via endogenous cannabinoid receptor signaling—atypical beliefs encompass non-standard attributions, magical thinking, cure-oriented cognitions, or unconventional somatic appraisals.

Such beliefs may include notions that cannabis will radically alter underlying personality structures, instantly eradicate complex chronic organic pathologies (acting as a universal panacea rather than a palliative agent), produce unique metaphysical insights, or cause anomalous physical sensations unaligned with known cannabinoid pharmacology. Psychometrically, measuring Atypical Beliefs is critical for clinical risk stratification. Patients with elevated scores on this factor may present with heightened susceptibility to nocebo or atypical placebo responses, vulnerability to commercial misinformation from unregulated dispensaries, poor medical decision-making, and an elevated risk of delaying conventional evidence-based medical treatments in favor of alternative cannabinoid regimens.

Theoretical Framework

The theoretical architecture underpinning the CEEQ-M draws from three primary traditions within clinical psychology, psychopharmacology, and cognitive science: Social Cognitive Theory, Cognitive Expectancy Theory of Substance Use, and Response Expectancy Theory.

1. Social Cognitive Theory and Informational Learning

According to Albert Bandurau2019s Social Cognitive Theory (Bandura, 1986), human agency operates within an interdependent network of triadic reciprocal causation involving cognitive/biological events, behavioral patterns, and environmental influences. Expectancies develop through direct physiological conditioning as well as vicarious learning, observational modeling, and symbolic communication.

In the domain of medical cannabis, social transmission plays an outsized role compared to other prescription drugs. Due to historical scheduling, scientific discourse has often been outpaced by public advocacy, digital forum testimonials, and marketing claims by commercial cannabis dispensaries. Consequently, patients formulate complex cognitive networks regarding cannabinoid efficacy before ever ingesting a single dose. These cognitive representations structure how patients interpret ambiguous internal somatic signals, transforming somatic sensations into cognitive evidence confirming their pre-existing beliefs.

2. Addictive and Ingestive Expectancy Models

The translation of expectancy theory to substance administration originated largely in the alcohol addiction research of Mark S. Goldman and colleagues (Brown et al., 1987; Goldman et al., 1999). Goldman established that expectancies are stored in semantic memory as “ifu2013then” associative nodes (e.g., “If I consume cannabis, then my lumbar pain will be abolished”). When substance-related cues are encountered, these nodes are activated, automatically priming behavioral pursuit and shaping post-consumption interoceptive appraisal.

When adapted to illicit or recreational cannabis use by Schafer and Brown (1991), expectancy models accounted for significant variance in consumption patterns over and above biochemical markers. However, Weiss and colleagues (2023) recognized that the valenced nodes in medical cannabis patients differ radically from recreational cohorts: the node “tension reduction” in recreation is often a desire for stress relaxation or euphoric intoxication, whereas in medical cohorts, it signifies the attenuation of pathological panic or visceral neuromuscular hypertonicity.

3. Kirschu2019s Response Expectancy Theory

Irving Kirschu2019s (1985, 1997) Response Expectancy Theory posits that anticipation of non-volitional psychological and physiological reactions (such as pain relief, sedation, fear reduction, or nausea alleviation) can directly evoke those identical responses without conscious mediation. In clinical trials, response expectancies represent the primary psychological mechanism driving the placebo effect.

Because cannabis contains aromatic terpenes and distinct psychoactive cues that make blinded placebo conditions notoriously difficult to maintain, active-placebo mismatches frequently occur. Patients often discern whether they received active THC, triggering unblinded response expectancies. The CEEQ-M was designed precisely to capture these response expectancies at baseline and across treatment checkpoints, enabling investigators to statistically model the proportion of symptom recovery attributable to cannabinoid pharmacology versus response expectancy phenomena.

Validity

The validity of the CEEQ-M was evaluated across a four-study development pipeline embedded within a longitudinal randomized clinical trial investigating the psychiatric and cognitive outcomes of medical cannabis authorization (Weiss et al., 2023). The validation process examined content, construct, and measurement invariance.

1. Content and Face Validity

Content validity was established through an initial 21-item derivation pool built from empirical field data collected by Reinarman et al. (2011), who surveyed 1,746 validated medical cannabis patients regarding primary treatment goals, relief attributions, and physiological changes. Items were synthesized and refined by doctoral-level clinical addiction psychologists and neuropsychiatrists to ensure direct medical relevance, linguistic neutrality, and absence of recreational colloquialisms. Cognitive debriefing sessions with patients confirmed that the items reflected typical and atypical clinical presentations.

2. Construct and Structural Validity

Construct validity was demonstrated through exploratory and confirmatory factor analyses in independent, clinically well-characterized adult cohorts. The retention of a two-factor latent structure demonstrated distinct divergence between realistic therapeutic objectives (Symptom Relief) and extreme or unverified beliefs (Atypical Beliefs). Standardized factor loadings across items were uniform and robust, confirming that the latent constructs adequately account for observed covariance across the indicators.

3. Longitudinal and Subgroup Measurement Invariance

A rigorous psychometric contribution of the CEEQ-M is its demonstrated measurement invariance. To confirm that a scale measures identical constructs across diverse demographics and across repeated clinical visits, multigroup and longitudinal structural equation modeling must prove three hierarchical steps: configural invariance (equivalent factor structure), metric/weak invariance (equivalent factor loadings, $lambda$), and scalar/strong invariance (equivalent item intercepts, $tau$).

Using robust maximum likelihood estimation, Weiss et al. (2023) established complete scalar invariance across four critical methodological axes:

• Trial Randomization Condition: Measurement properties remained strictly invariant between participants immediately randomized to obtain a medical cannabis card and those randomized to a waitlist control condition.
• Biological Sex: Item parameters functioned equivalently across adult male and adult female respondents, proving the absence of gender-based measurement bias.
• Age Brackets: Scalar invariance was sustained when comparing emerging adults (aged 18u201325 years) with mature adults (aged 26u201365 years), verifying that age differences in raw scores reflect authentic developmental variations in expectations rather than divergent item interpretations.
• Longitudinal Assessment Across Time: Crucially, scalar invariance was established across successive test-retest administration points spanning the baseline visit and the 1-month post-randomization clinical follow-up. This longitudinal invariance confirms that observed shifts in CEEQ-M scores over time reflect true cognitive adjustments resulting from ongoing cannabis exposure rather than scale recalibration or measurement drift.

Reliability

The CEEQ-M exhibits high internal consistency and score reliability across diverse diagnostic subpopulations presenting with chronic pain, insomnia, depressive disorders, and clinical anxiety. Classical test theory indicators (Cronbachu2019s $\alpha$) and modern latent variable reliability metrics (McDonaldu2019s $\omega$) corroborate the scale’s precision.

For the Symptom Relief subscale, the observed coefficients ($\alpha = 0.906$, $\omega = 0.907$) exceed the 0.90 benchmark typically recommended for high-stakes individual clinical decision-making. These metrics confirm that the items within this subscale share substantial common variance and measure their target construct with minimal measurement error. The Atypical Beliefs subscale demonstrates solid reliability ($\alpha = 0.779$, $\omega = 0.772$), well above the conventional 0.70 threshold for research and exploratory screening purposes. Given that atypical beliefs encompass a more heterogeneous set of non-standard cognitive attributions, this reliability profile aligns with theoretical expectations.

Temporal stability across the 1-month clinical period indicated that expectancy structures are responsive to active behavioral feedback. Participants who initiated cannabis therapy updated their expectations in accordance with observed pharmacological efficacy and side-effect profiles, while control participants maintained stable baseline expectancies. This sensitivity confirms that the CEEQ-M tracks meaningful psychological change over time rather than behaving as an immutable personality trait.

Factor Analysis

The latent dimensionality of the CEEQ-M was identified and confirmed using a rigorous sequential structural equation modeling approach comprising Exploratory Factor Analysis (EFA) followed by Confirmatory Factor Analysis (CFA).

1. Exploratory Factor Analysis (EFA)

The initial 21-item questionnaire was administered to an extraction cohort of adults seeking cannabis for medical management. Principal axis factoring (PAF) was conducted to extract underlying latent dimensions without assuming absence of measurement error. Factor retention decisions were guided by Kaiser-Guttman eigenvalues (> 1.0), visual examination of the scree plot inflection point, and parallel analysis.

The empirical analysis clearly favored a two-factor latent solution. Four of the initial 21 items were systematically dropped due to inadequate primary factor loadings (< 0.40), excessive cross-loadings onto non-target factors (> 0.30 difference failure), or poor item-rest correlations. The remaining 17 items cleanly resolved across two uncorrelated or moderately correlated factors:

• Factor 1 (Symptom Relief): Explained 31.0% of the total common variance among the items. Items loading onto this factor captured anticipated reductions in physical pain, insomnia, subjective anxiety, and depressive affect.
• Factor 2 (Atypical Beliefs): Explained 13.9% of the total common variance. Items loading onto this factor indexed unusual, idiosyncratic, or scientifically unsupported physical and mental outcomes.

Together, these two distinct dimensions accounted for 44.9% of the total variance across the 17-item set, confirming that the factor extraction captured the substantive core of medical cannabis expectancies.

2. Confirmatory Factor Analysis (CFA)

To confirm this structural architecture, a Confirmatory Factor Analysis was conducted utilizing data from all participants at the 1-month post-randomization follow-up checkpoint. The 17-item, two-factor specification was evaluated against standard global goodness-of-fit benchmarks (Hu & Bentler, 1999):

• Comparative Fit Index (CFI): Indicated strong model fit, consistently exceeding conventional cutoffs (> 0.90 and approaching 0.95).
• Tucker-Lewis Index (TLI): Paralleled the CFI findings, demonstrating that the two-factor specification represented a vast improvement over the baseline independence model.
• Standardized Root Mean Square Residual (SRMR): Yielded values < 0.08, confirming minimal average discrepancy between the observed sample correlation matrix and the model-implied correlation matrix.
• Root Mean Square Error of Approximation (RMSEA): Demonstrated acceptable fit within conventional confidence intervals, confirming that the two-factor model accurately reflects real-world sample data without excessive overparameterization.

Alternative single-factor models (forcing all 17 items into a unitary general expectancy factor) yielded significantly worse fit metrics, confirming that Symptom Relief and Atypical Beliefs represent distinct psychometric dimensions rather than points along a single continuum.

Instrument / Measurement Tool

The Cannabis Effects Expectancy Questionnaire-Medical (CEEQ-M) is structured as follows:

• Test Type: Original multi-item self-report questionnaire / psychometric inventory.
• Construct Assessed: Outcome expectancies regarding medical cannabis effects (Symptom Relief and Atypical Beliefs).
• Number of Items: 17 items.
• Item Generation Origin: Derived from a 21-item conceptual pool based on patient surveys in licensed medical cannabis dispensaries (Reinarman et al., 2011).
• Administration Format: Standardized self-report; suitable for computerized assessment (web-based research batteries, clinical portals), paper-and-pencil formats, or clinical trial electronic data capture (EDC) systems.
• Completion Time: Approximately 3 to 5 minutes.
• Target Population: Adults aged 18 years and older (inclusive of young adulthood [18u201325], adulthood [26u201364], and geriatric cohorts [65+]) who are considering, initiating, or currently using cannabis for therapeutic/medical symptoms.
• Target Indications: Chronic pain conditions, sleep disturbances/insomnia, anxiety disorders, and depressive episodes.
• Response Format: Continuous 10-point Likert-type scale with verbal anchors spanning 5 ordered categories:
  • 1 and 2: “Strongly Disagree”
  • 3 and 4: “Disagree”
  • 5 and 6: “Neutral”
  • 7 and 8: “Agree”
  • 9 and 10: “Strongly Agree”
• Scoring Procedures:
  • Subscale scores are calculated by summing or averaging the items assigned to each respective latent factor: Symptom Relief Subscale Score and Atypical Beliefs Subscale Score.
  • Higher scores on the Symptom Relief dimension indicate strong subjective anticipations that cannabis will effectively alleviate target clinical complaints (pain, sleep disruption, depression, and anxiety).
  • Higher scores on the Atypical Beliefs dimension denote elevated endorsement of uncommon, idiosyncratic, or unverified claims regarding cannabis action.
  • No items are reverse-coded. Mean subscale scores retain the original 1-to-10 interpretative scaling.

Permissions & Fee and Test Year

The Cannabis Effects Expectancy Questionnaire-Medical (CEEQ-M) was formally published in 2023 by the American Psychological Association (APA) in the journal Psychological Assessment (Weiss et al., 2023).

• Permissions: The CEEQ-M may be utilized free of charge for non-commercial academic research, institutional teaching, and non-profit clinical trial investigations without explicit written copyright permissions, provided proper scientific attribution and citation are maintained.
• Commercial Use & Licensing: Commercial applications, proprietary clinical platforms, digital health product integrations, or sponsored pharmaceutical trials wishing to deploy the scale should contact the corresponding author, Dr. Jodi M. Gilman, or the Massachusetts General Hospital Center for Addiction Medicine.
• Correspondence Information: Jodi M. Gilman, Ph.D., Center for Addiction Medicine, Department of Psychiatry, Massachusetts General Hospital, 101 Merrimac Street, Suite 320, Boston, MA 02114, USA. Email: [email protected].
• Item Access: The complete item text, operational instructions, and factor loadings are available in the official supplemental material accompanying the original publication in the APA PsycNet repository (Supplemental Material Document: 2023-78778-001).

References

• Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.
• 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), 479u2013491. https://doi.org/10.15288/jsa.1987.48.479
• Goldman, M. S., Del Boca, F. K., & Darkes, J. (1999). Alcohol expectancy theory: The application of cognitive science. In K. E. Leonard & H. T. Blane (Eds.), Psychological theories of drinking and alcoholism (2nd ed., pp. 203u2013246). Guilford Press.
• 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), 1u201355. https://doi.org/10.1080/10705519909540118
• Kirsch, I. (1985). Response expectancy as a determinant of experience and behavior. American Psychologist, 40(11), 1189u20131202. https://doi.org/10.1037/0003-066X.40.11.1189
• Kirsch, I. (1997). Specifying response expectancies: Comments on Kirsch (1985, 1990). Cognitive Therapy and Research, 21(5), 559u2013566. https://doi.org/10.1023/A:1021881829630
• Reinarman, C., Nunberg, H., Lanthier, F., & Heddleston, T. (2011). Who are medical marijuana patients? Population characteristics from nine California practices. Journal of Psychoactive Drugs, 43(2), 128u2013135. https://doi.org/10.1080/02791072.2011.587700
• Schafer, J., & Brown, S. A. (1991). Marijuana and cocaine effect expectancies and drug use patterns. Journal of Consulting and Clinical Psychology, 59(4), 558u2013565. https://doi.org/10.1037/0022-006X.59.4.558
• Weiss, J. H., Tervo-Clemmens, B., Potter, K. W., Evins, A. E., & Gilman, J. M. (2023). The Cannabis Effects Expectancy Questionnaireu2013Medical (CEEQ-M): Preliminary psychometric properties and longitudinal validation within a clinical trial. Psychological Assessment, 35(8), 659u2013673. https://doi.org/10.1037/pas0001244

Items of the Scale