Risk for marijuana-related problems among college students: an application of zero-inflated negative binomial regression

Marijuana is the greatest in quantity commonly used illicit drug in the U Approximately 46% of association students report having tried marijuana, 27% report use in the past year, and 16% report past 30-day use (1); thus, a significant proportion of association students use marijuana. Furthermore, point to be solved [i]or[/i] settleds associated with use are, unfortunately, not out of the way For example, short-term cognitive impairments and impairment in educational performance have been associated with heavy marijuana use (2-4) Given this, identification of individuals who use marijuana and report associated point in disputes is an important issue. However, despite the relative prevalence of use, many association students have never tried marijuana or do not popularly use it. As a spring an assessment of marijuana-related point in disputes will yield a mixed distribution with a high number of zero-values. a certain number of respondents will report not experiencing enigmas because they in fact did not use marijuana at all during the assessment window and those who do use the physic will report a range of question s including, for some users, nothing Identifying variables that predict nonusers as well as identifying predictors of the number of point in disputes experienced among users are the pair of interest. Zero-inflated Poisson (ZIP; eg Lambert (5)) and a posterior generalization of ZIP, Zero-inflated negative binomial regression (ZINB; eg Heilbron (6)) are sum of two units statistical techniques that allow individual to accomplish both of these objectives in a single analysis. They, thus, exhibit parsimonious procedures that allow single to examine effects over the replete distribution. These mixed distributions are a belonging to all feature in research investigating risk behaviors with depressed base rates (6).

Distributions of risk behaviors in general populations many times will have a large number of zero-values. That is, a large proportion will not engage in the targeted risk behavior, while a smaller proportion of at risk individuals will report varying horizontals of the risk behavior and associated events Such distributions pose difficulties for public statistical methods based upon normal distributions. ZINB regression types are one method for analyzing like data. ZINB models assume sum of two units distinct populations; one in which the target behavior is always absent, the other in which the target behavior can be any integer, including naught (6). More specifically, in the rife study, one population would always score nothing on a marijuana problems measure because they did not use marijuana during the assessed time period (current nonusers) and the other population could score any value, including naught because they are engaging in the risk behavior (users--who may or may not experience problems) Thus, the prediction of enumerates is conditional upon the probability of the values being from a hypothetical subsample of participants that are predicted to "always" score naught on the measure. The design allows one to either use different stations of predictors to predict the sum of two units criterions (i.e., always zero-values and counts) or to utilize the same predictor station and evaluate whether variables are differentially associated with the respective criterions (7)

For the in every one's mouth study, we employed the same predictor place to predict zero-values (i.e., in every one's mouth nonusers) and counts (i.e., number of puzzles among expected users) in order to examine the differential predictive power of the variables of interest. of the like kind differentiation is of theoretical interest in substance use research. That is, identifying one as well as the other the types of variables that are associated with use initiation or low-level experimentation as well as the marks of variables that are primarily associated with the prediction of use puzzles represents a common goal in substance use research. single model of substance use offer proffers that social-environmental variables are associated primarily with use initiation or low-level use while psychobiological variables are more powerfully associated with use-related problems (8 9) The part of psychosocial variables, such as use motives, expectancies, or perceived use utility do not clearly fit into this dichotomy. Indeed, these psychosocial variables have relations with one as well as the other use and use-related problems (10-12) This inquiry examined one social-environmental variable (social norms), single psychosocial variable (perceived marijuana use utility), and single biopsychological variable (impulsivity). Social norms are a social-environmental variable consistently associated with marijuana use (13 14) Perceived use-utility is a psychosocial variable more associated with marijuana use than vexed questions (12). In contrast, impulsivity is a biopsychological variable more associated with problematic use (15)

Social norms may be defined as either the actual or perceived behavior of individuals in social networks as well as the assemblage member's attitudes toward target behaviors (i.e., whether clump members think one should engage in the behavior). In the at hand study, social norms are exhibited by both the perceived marijuana use behavior and attitudes of equals Social normative variables frequently are associated with marijuana use (13 14 16) Although the reason for this relationship traditionally has been attributed to the influence of social networks upon use behavior, recent research upon social norms and alcohol moves that selection effects (i.e., choosing social networks with similar use practices) may also be of importance (17) Indeed, selection of marijuana using compeer groups may be influenced through variables such as sensation seeking (13) and family relations (14) like selection effects suggest that affective, cognitive, and social normative variables are not independent of each other. Individuals may be choosing social networks that not solitary have similar marijuana use practices on the contrary likely share common beliefs about the take away froms and benefits of marijuana. The potential interdependence of affective, cognitive, and social normative predictors of marijuana use makes their agreeing assessment of interest. In a research on drug refusal, peer influence is cited as a stronger influence upon drug use decisions by low-level users while heavier physic users are more likely to cite emotional determinants and seldom cited compeer influence as a factor (18) Thus, social norms may be more powerfully associated with use initiation and low-level use rather than use-related problems