PROJECT SUMMARY/ABSTRACT With expanding cannabis legalization and concurrent significant increases in cannabis-related traffic fatalities, driving under the influence of cannabis (DUIC) has become a major public health concern. Despite the impairment in driving skills and increased crash risk associated with cannabis intoxication, little is known about how individuals make the decision to drive (or not) after using cannabis. The proposed project advances our team’s work on cannabis and DUIC risk factors to identify person-level (e.g., DUIC cognitions), within-person processes (e.g., subjective intoxication, affective state, cannabinoid concentration, quantity, route of administration), and contextual influences (e.g., social context, environmental conditions) that predict DUIC behaviors in daily life. The proposed multi-method project will provide unique qualitative and quantitative information about individual-level DUIC behavior using geographically explicit ecological momentary assessment (EMA), which combines event-level data on cannabis use with spatio-temporal data on vehicle movement. Our preliminary data has demonstrated the feasibility of these methods by integrating passively collected continuous location data using a vehicle-based GPS tracking device with EMA data on driving and cannabis use. Frequent and less frequent cannabis users (N = 260) will complete smartphone measures of cannabis use, concurrent substance use, affect, momentary impulsivity, perceived DUIC dangerousness, driving intentions, driving motives and destinations/location, and context during a 4-week EMA period. DUIC (i.e., driving within a pharmacologically-relevant timeframe) will be identified by integrating geospatial data (i.e., latitude/longitude/time and vehicle’s movement) passively and continuously collected in the field with EMA data on cannabis use. Weekly testing of participant cannabis for Δ⁹-tetrahydrocannabinol (THC) and cannabidiol (CBD) concentration will be done using a near-infrared spectroscopy device. This will be the first study to prospectively examine the influence of within-person and contextual predictors on DUIC likelihood and distance traveled (Aim 1). We will explore driving-related cognitions, indices of working memory capacity, and user characteristics in relation to cannabis use and driving experience as potential moderators of the effects of event-level predictors on DUIC (Aim 2). To provide insights on contextual factors and decision-making processes related to DUIC, participants will complete a narrative qualitative interview focused on annotating maps of representative DUIC and non-DUIC trips generated by GPS data during one of the four weeks of the EMA period (Aim 3). Exploratory machine learning analyses will be used to characterize and distinguish DUIC episodes from non-DUIC driving episodes (Aim 4). The current project will be the first full-scale study to integrate such real-time individual-level exposure data with DUIC outcome...