It depends on when you ask: motives for using marijuana assessed before versus after a marijuana use event.

BACKGROUND: Marijuana use motives are typically evaluated retrospectively using measures that summarize or generalize across episodes of use, which may compromise validity. Using Ecological Momentary Assessment data, we examined the main reason for a specific marijuana use event measured both prospectively and retrospectively. We then determined reason types, event characteristics, and user characteristics that predicted change in reason.
METHODS: Thirty-six medical outpatients age 15 to 24 years who used marijuana two times a week or more used a handheld computer to select their main reason for use from the five categories of the Marijuana Motives Measure (Simons, Correia, & Carey, 1998) just before and after each time they used marijuana over two weeks (n=263 events with before/after reason). The reasons were examined individually and according to dimensions identified in motivational models of substance use (positive/negative, internal/external).
RESULTS: The reason assessed before use changed to a different reason after use for 20% of events: 10% of events for pleasure; 21%, to cope; 35%, to be more social; 55%, to expand my mind; and 100%, to conform. In the multivariable model, external and expansion reasons each predicted change in reason for use (p<0.0001 and p=0.001, respectively). Youth were also more likely to change their reason if older (p=0.04), if male (p=0.02), and with weekend use (p=0.002).
CONCLUSION: Retrospective assessments of event-specific motives for marijuana use may be unreliable and therefore invalid for a substantial minority of events, particularly if use is for external or expansion reasons.