Cannabis is one of the most widely used psychoactive substances in the world. The United Nations Office on Drugs and Crime estimates that roughly 200 million people, about 4.4 percent of the global population aged 15 to 64, used it in 2022. In Europe the picture is even more striking among the young: 18.6 percent of adults aged 15 to 24 reported use in the past year, and around 2 percent of people aged 15 to 34 reported daily or near-daily consumption. Frequent use carries documented costs, including poorer mental health, cognitive impairment, elevated risk of motor vehicle accidents, and a higher likelihood of developing cannabis use disorder. Yet predicting who will use cannabis, and how often, has remained a stubbornly difficult problem for behavioral scientists.
A new longitudinal study published in the International Journal of Mental Health and Addiction offers one of the most detailed answers to date. A team of Spanish researchers led by Alberto Parrado-González, José Carmona-Márquez, Belén del Valle Vera, and Fermín Fernández-Calderón set out to test whether the dominant psychological model of health behavior, the theory of planned behavior, could explain cannabis use among young adults, and whether adding measures of habit and past behavior would substantially improve that explanation. Their findings suggest that for regular users, the automatic pull of habit may matter as much as, or more than, conscious deliberation.
The theory of planned behavior, formulated by Icek Ajzen in 1991, holds that the strongest immediate predictor of any behavior is the intention to perform it. Intention, in turn, is shaped by three belief-based constructs: attitudes, which reflect evaluations of the consequences of the behavior; subjective norms, which capture perceived approval from significant others; and perceived behavioral control, which encompasses both perceived control over external circumstances and self-efficacy, the confidence in one’s ability to carry out the required actions. In the substance use context, self-efficacy is typically operationalized as confidence in the ability to refrain from using in high-risk or tempting situations. Meta-analyses indicate that the model explains roughly 40 to 50 percent of the variance in intentions and 25 to 40 percent of the variance in health-related behaviors, and prior cannabis studies have found similar patterns.
But the theory rests on an assumption that has come under increasing scrutiny: that behavior flows from rational, deliberate decision-making. A growing body of research shows that behavior is also shaped by automatic, unconscious processes, particularly when actions are performed repeatedly under stable conditions in response to consistent environmental cues. In such cases, cognitive deliberation becomes a weaker predictor and behavior is enacted with minimal conscious processing. Habit, defined as behavioral patterns automatically triggered by environmental cues consistently associated with the behavior, is the key construct representing these automatic processes. Early research often used past behavior frequency as a proxy for habit, but this approach has limitations, since past behavior also shapes beliefs and intentions through social-cognitive channels, and habit itself extends beyond mere repetition to include automaticity, lack of awareness, and reduced volitional control.
To disentangle these pathways, the researchers recruited 606 Spanish-speaking young adults aged 18 to 25 from public venues in Huelva and Seville, Spain, using non-probability targeted sampling designed to maximize diversity in age, gender, college status, and use frequency. All participants reported cannabis use in the past month, with a mean of 18.1 days of use. At baseline, they completed surveys measuring attitudes, subjective norms, perceived control, refusal self-efficacy, intention to use cannabis over the next 90 days, and habit strength, assessed with a brief two-item measure capturing automaticity and the customary nature of use. Three months later, 501 participants, or 82.6 percent, completed a follow-up assessment of use frequency and habit strength. The study was approved by the Bioethics Research Committee of Andalusia and complied with the Declaration of Helsinki.
The results were striking. The original theory of planned behavior model showed good fit to the data and explained 39.2 percent of the variance in intention and 56.2 percent of the variance in cannabis use frequency at follow-up. All deliberative constructs were significantly associated with intention, with subjective norms showing the strongest association, a pattern the authors attribute to the illicit and stigmatized status of cannabis, which tends to concentrate use within small, close-knit social networks where the expectations of friends and family carry unusual weight. Intention was prospectively associated with use, while perceived control and refusal self-efficacy showed negative relationships, meaning that participants who felt less able to resist used more frequently.
When the researchers extended the model to include past behavior and habit strength at both time points, explanatory power jumped dramatically. The extended model accounted for 65.7 percent of the variance in intention and 67.6 percent of the variance in cannabis use frequency. More revealing was what happened to the deliberative constructs: their associations with intention were attenuated or nullified entirely. Only subjective norms and attitudes remained significantly associated with intention, and these associations were weaker in magnitude. Perceived control and self-efficacy no longer showed significant associations with either intention or behavior. Notably, baseline habit strength was not significantly associated with intention, a finding the authors interpret as consistent with the view that intention and habit represent partly independent pathways to behavior. Automatic responses do not necessarily generate conscious intention; individuals can act habitually without actively intending to do so in the moment.
Instead, past behavior and habit strength at follow-up emerged as the strongest predictors of actual cannabis use. Indirect-effect analyses revealed three complementary pathways through which past behavior shaped later use: through deliberative constructs such as attitudes and subjective norms, consistent with Ajzen’s proposal that repeated performance shapes cognitive evaluations; through habit strength at both time points, with baseline habit showing a significant indirect association with follow-up behavior via follow-up habit; and through a residual direct association that habit alone did not capture, suggesting that additional non-deliberative factors such as implicit attitudes, implicit identity, and impulsivity may also be at work. The strong association between baseline and follow-up habit underscores the persistence and resistance to change of established habits.
Crucially, the extended model proved statistically invariant across gender and college status. Multigroup analyses showed that constraining the model’s coefficients to be equal across men and women, and across college students and non-students, did not significantly worsen fit. This suggests that the same psychological architecture underlies cannabis use across these subgroups, and that interventions derived from the model could be applied broadly. The finding is notable because prior evidence on gender differences in theory of planned behavior relationships has been mixed, and most cannabis studies have focused on college populations, raising concerns about generalizability to young adults outside academic settings.
The practical implications are substantial. If cannabis use is driven partly by automatic responses to environmental cues, then interventions that target only beliefs and attitudes may be insufficient for established users. The authors suggest that prevention and reduction programs could help participants identify the situations in which they commonly use cannabis, such as spending time with particular friends, visiting certain places, or experiencing boredom or stress, and then modify routines, reduce access, or substitute alternative responses such as contacting a supportive person or leaving the setting. Self-regulation strategies including goal setting, self-monitoring, and contingency management may support change by increasing awareness of use patterns. At the same time, because intention remained significantly associated with use even after controlling for habit and past behavior, conscious decision-making still matters. Techniques such as mental simulation and implementation intentions, for example forming if-then plans like deciding in advance to take a walk before acting on an urge, may help individuals plan alternative responses in cue-rich situations. The study’s authors caution that the non-probability sampling design, reliance on self-report, and brief two-item habit measure limit generalizability, and that the three-month follow-up may have introduced temporal variability. Still, the core message is clear: understanding cannabis use requires accounting for both the reasoned and the automatic, and the most effective interventions will likely need to speak to both.
Subject of Research: The role of habit strength and past behavior as non-conscious predictors of cannabis use frequency within an extended theory of planned behavior model among young adults.
Article Title: The Role of Habit and Past Behavior in Cannabis Use: An Extended Theory of Planned Behavior Model
Article References: Parrado-González, A., Carmona-Márquez, J., Vera, B. D. V., & Fernández-Calderón, F. (2026). The Role of Habit and Past Behavior in Cannabis Use: An Extended Theory of Planned Behavior Model. International Journal of Mental Health and Addiction. https://doi.org/10.1007/s11469-026-01746-9
Image Credits: AI Generated
DOI: 10.1007/s11469-026-01746-9
Keywords: cannabis use, theory of planned behavior, habit strength, past behavior, dual-process models, young adults, substance use, self-efficacy, subjective norms, longitudinal study, behavior change, prevention
News Source: Ophelia Keating. (October 16, 2026). Habit Outweighs Deliberate Choice in Predicting Cannabis Use Among Young Adults. Scienmag.
