|Home | About | Journals | Submit | Contact Us | Français|
Marijuana is the most commonly used illicit drug by college students. Prior studies have established an association between marijuana use and poor academic performance in college, but research on the frequency of marijuana use over the entire college career is limited. The study objective was to examine the association of marijuana use trajectories on academic outcomes, including senior year enrollment, plans to graduate on time, and GPA.
Data were collected from a cohort of 3,146 students from 11 colleges in North Carolina and Virginia at six time points across the college career. Group-based trajectory models were used to characterize longitudinal marijuana use patterns during college. Associations between marijuana trajectory groups and academic outcomes were modeled using random-effects linear and logistic regressions.
Five marijuana trajectory groups were identified: non-users (69.0%), infrequent users (16.6%), decreasing users (4.7%), increasing users (5.8%), and frequent users (3.9%). Decreasing users and frequent users were more likely to drop out of college and plan to delay graduation when compared to non-users. All marijuana user groups reported lower GPAs, on average, than non-users.
These results identify marijuana use patterns that put students at risk for poor academic performance in college. Students who use marijuana frequently at the beginning of the college career are especially at risk for lower academic achievement than non-users, suggesting that early intervention is critical.
Marijuana is the most commonly used illicit substance among college students, with 48.5% reporting lifetime use, 20.8% past month use, and 5.9% reporting daily use in 2013 (Johnston et al., 2015). Daily and past 30 day marijuana use among college students has risen steadily since 2007. Daily marijuana users exhibit more characteristics of dependence than less frequent users (Hammersley and Leon, 2006), which makes the increase in daily use particularly concerning. At the same time that daily use is increasing, perceptions of harm associated with regular marijuana use are declining; only 35.1% of young adults think smoking marijuana regularly places the user at great risk compared to 57.2% a decade ago (Johnston et al., 2015).
Prior research has found that college student marijuana users are more likely to be white, male, single, members of fraternities or sororities, non-athletes, not religious, cigarette smokers, and heavy episodic drinkers (Bell et al., 1997; Johnston et al., 2015; Buckman et al., 2011; Wechsler et al. 1997; Yusko et al., 2008; McCabe et al., 2005; Mohler-Kuo et al., 2003). Students who initiate marijuana prior to age 16 are more likely to continue to use marijuana in college and be regular users (Mohler-Kuo et al., 2003), and early age of initiation has been shown to be associated with problems later in life such as depression and drug dependence (Green and Ritter, 2000; Ellickson et al., 2005; Chen et al., 2009). In one study, initiation of marijuana use during freshman year was found to be associated with living on campus, using cigarettes or alcohol, and Hispanic ethnicity (Suerken et al., 2014).
Acute effects of marijuana use among college students include impaired driving (Whitehill et al., 2014) and engaging in risky sexual activity (Bell et al., 1997) as well other high risk behaviors (Shillington and Clapp, 2001; Kouri et al., 1995). Several studies have linked marijuana use with impaired mental functioning and reduced psychological well-being. College student marijuana use has been found to be associated with anxiety, depression, hostility, interpersonal sensitivity, paranoia, and psychoticism (Buckner et al., 2010). Marijuana use is associated with the impairment of many cognitive functions that affect academic performance, including attention, concentration, memory, verbal fluency, processing speed, planning, and decision making (Caldeira et al., 2008; Churchwell et al., 2010; Hermann et al., 2007; McHale et al., 2008; Ramaekers et al., 2006; Shillington and Clapp, 2001; Vadhan et al., 2007; Wadsworth et al., 2006). Marijuana use reduces brain volume, affects brain metabolism, alters brain circuitry, and restricts blood flow to the brain, thereby reducing cognitive performance (Battistella et al., 2014; Block et al., 2002; Churchwell et al., 2010; Hermann et al., 2007; Verdejo-García et al., 2006; Yücel et al., 2008). Chronic marijuana use poses even more risks. Heavy and long-term marijuana users experience even greater difficulties with cognitive functioning, compared to light users and non-users (Block et al., 2002; Bolla et al., 2002; Kouri et al., 1995; Pope and Todd, 1996; Solowij et al., 1995, 2002, 2011; Verdejo-García et al., 2006; Whitlow et al., 2004; Yücel et al., 2008). Chronic marijuana users report higher levels of sensation seeking as well as more problems with self-control and externalizing behavior (Brook et al., 2011).
Especially relevant to college students is the immediate impact that marijuana use has been shown to have on academic performance. Marijuana use is associated with dropping out of college (Braun et al., 2000; Degenhardt et al., 2010; Fergusson et al., 2003; Fergusson and Boden, 2008; Fleming et al., 2012; Hunt et al., 2010; Schulenberg et al., 2005; Tucker et al., 2005, 2006), having a lower GPA (Arria et al., 2013a, 2015; Bell et al., 1997; Buckner et al., 2010), poorer performance on exams and projects (Shillington and Clapp, 2001), spending less time studying for classes (Bell et al., 1997), and lower class attendance (Caldeira et al., 2008; Arria et al., 2013a, 2015; Shillington and Clapp, 2001). Marijuana craving has been shown to be negatively associated with time spent studying and academic motivation in college, and more frequent marijuana use has been found to be negatively associated with college GPA (Phillips et al., 2015; Martinez et al., 2015). Another study found that the likelihood of earning a college degree declines with more frequent marijuana use (Horwood et al., 2010).
Patterns of frequency of marijuana use may vary over the course of the college career, so it is important to study the complete trajectory of marijuana use during those years. Several studies report the impact of any past year or past month marijuana use on college performance but do not measure how often students use marijuana (Bell et al., 1997; Braun et al., 2000; Shillington and Clapp, 2001). Several studies on the impact of marijuana trajectories on academic performance and educational aspirations follow a cohort of adolescents from adolescence into young adulthood but do not focus specifically on college students (Brook et al., 2011; Degenhardt et al., 2010; Fleming et al., 2012; Flory et al., 2004; Schulenberg et al., 2005; Tucker et al., 2005; Windle and Wiesner, 2004). Only one study has focused on college student frequency of marijuana use over time. Arria et al. (2013b) found that infrequent marijuana users, increasing users, and chronic/heavy users are more likely to have a gap in college enrollment compared to minimal users. They also found that increasing marijuana use over the college career was associated with a drop in GPA and that marijuana use frequency during the first year of college had an enduring effect on delaying graduation, via its influence on the path from skipping class to GPA at baseline (Arria et al., 2015). However, this study only included students at one college. More research is needed in order to understand the impact of the frequency of marijuana use across the college career on academic outcomes.
Data were obtained as part of the Smokeless Tobacco Use in College Students study. The objective of this study was to assess trajectories and correlates of smokeless tobacco use among a cohort of college students by surveying them at multiple points during their college career (Wolfson et al., 2014). All first year students enrolled at 11 colleges in North Carolina and Virginia were recruited through school email to participate in a brief web-based screener survey in fall 2010 in order to determine study eligibility. Nine participating colleges are public schools, and two are private schools. Five colleges are located in rural areas, four colleges are located in suburban communities, and two colleges are in urban areas. Thirty-six percent (10,528) of eligible students participated in the screener survey (Spangler et al., 2014).
A random sample of eligible participants was selected two weeks after the screener survey. This sample was invited to participate in the longitudinal cohort study. Students were selected within each school with an objective of 285 completions per school in order to have sufficient power to detect differences in smokeless tobacco use for various predictors in the parent study. Due to the goal of the parent study, students at higher risk for using smokeless tobacco were oversampled, including lifetime smokeless tobacco users, current cigarette smokers, and males. Data were collected each semester of the students' freshman and sophomore year, and during the fall of the students' junior and senior years (Wolfson et al., 2014). Students had the opportunity to update their contact information at each wave. Students who did not initially complete the survey via the URL emailed to them received follow-up text messages and phone calls with reminders to complete the survey. Attempts were made to contact all students who participated at baseline, including participants who dropped out of college. Among the 4,190 students who were invited to participate, 3,146 (64%) eligible students completed the first survey. Of the students who participated in the first survey, 2,520 (80.1%), 2,459 (78.2%), 2,507 (76.7%), 2,516 (80.0%), and 2,500 (79.5%) students participated in the second, third, fourth, fifth, and sixth surveys, respectively. Almost two thirds of the sample (65.4%) participated in all 6 waves, another 10.1% participated in 5 waves, 5.6% participated in 4 waves, and 18.9% of the sample participated in fewer than 4 waves. Females (p = 0.019) and students whose mothers do not possess at least a college degree (p = 0.041) were more likely to be missing at least one wave of data.
There was a $15 incentive for completing the first survey, and this incentive increased by $5 for each subsequent survey. The Wake Forest School of Medicine Institutional Review Board approved study protocol. Several participating schools also had their own Institutional Review Board approvals. A Certificate of Confidentiality by the Department of Health and Human Services was obtained in order to protect the privacy of the participants (Wolfson et al., 2014).
Three academic outcomes were measured during the fall semester of the participants’ senior year (Wave 6): current enrollment in college, plans to graduate from college on time, and grade point average. All academic outcomes were self-reported. Students were considered to be still enrolled in college if they reported a college where they were enrolled or had already graduated from college (since this indicates that they did not drop out of college). They were considered to not be enrolled in college if they reported taking a leave of absence or were no longer enrolled in an academic institution. Students were also asked to report the month and year that they planned to graduate from college. They were considered to be planning to graduate on time if they had already graduated or if their expected graduation date was May 2014 or earlier, since all participating schools hold spring commencement in May. The third college outcome that students reported was college grade point average. Grade point average was reported on a scale of 0–4, with any values over 4 being rounded down to 4.0.
During the first wave, students were asked if they had ever used marijuana. At each subsequent time point, students were asked if they had used marijuana within the past six months. If they answered affirmatively to either version of the question, then they were asked on how many days out of the past 30 days that they used marijuana, with the following response options: 0, 1–2, 3–5, 6–9, 10–19, 20–29, and all 30. Responses to this question were recoded to the midpoint of the category (i.e., a response of “6–9” was coded as “7.5”).
Demographic characteristics measured during fall 2010 (Wave 1) include gender, race (white and non-white), ethnicity (Hispanic and non-Hispanic), and mother’s education (4 year college degree or higher vs. less than a 4 year college degree). Spending money available in an average month (at least $100 per month vs. less than $100 per month) was measured at Wave 6.
Social characteristics were measured at Wave 6 and included participation in campus athletics (varsity, club, or intramural sports) within the past six months (yes vs. no); current membership or pledge status in a fraternity or sorority (yes vs. no); participation in religious activities at least twice per month over the past six months (yes vs. no); current residential status (on campus vs. off campus or studying abroad); and relationship status (steady partner or married vs. single, separated/divorced, or widowed). College graduates were not asked about their residential status and were assumed to be living off campus.
For cigarettes and hookah tobacco, students were considered a user if they reported using the substance at least once within the past 30 days at Wave 6. National Institute of Alcohol Abuse and Alcoholism guidelines (2004) were used to define heavy episodic drinking. Male and female students were denoted as heavy episodic drinkers if they drank at least five or four drinks in a row during the past 30 days, respectively. Illicit drug use was defined as using cocaine, methamphetamines, hallucinogens, rohypnol, ecstasy, or heroin at least once within the past six months at Wave 6.
The Brief Sensation Seeking Scale (Hoyle et al., 2002) was administered at Wave 6. Sensation seeking scores were computed by averaging eight five-point Likert scale items (1= strongly disagree to 5 = strongly agree) for all participants who answered at least 5 items in the scale. Cronbach’s alpha for the Brief Sensation Seeking Scale was 0.82. Stress was measured at Wave 6 using the Perceived Stress Scale (Cohen and Williamson, 1988). Scores were computed by summing ten items on a scale from 0=never to 4=very often. Two items were reverse coded. If only one or two items were missing, the mean of the remaining items was substituted for the missing items. Cronbach’s alpha for the Perceived Stress Scale was 0.86.
Group based trajectory modeling was used to identify the most common patterns of past-30-day marijuana use frequency during college (Nagin, 1999). Models used a zero-inflated Poisson distribution to account for the large number of students who did not use marijuana. Linear and quadratic terms for each trajectory group were included and compared. One- to eight-group models were considered. The best model was selected based on a combination of the Bayesian information criterion (BIC), group interpretability, and having reasonably large groups (at least 5% of the sample). Trajectory models were constructed using PROC TRAJ in SAS Version 9.4. Maximum likelihood estimation was used to estimate model parameters. Students were assigned to the marijuana trajectory group with the highest probability of membership.
Descriptive statistics on all demographic and social characteristics, substance use rates, and mental health and psychological factors are presented. The prevalence of being currently enrolled in college, graduating on time and the mean GPA were estimated by school to examine variation in academic outcomes. Bivariate associations between trajectory groups and all covariates were assessed via Chi square tests.
Random-effects linear and logistic regression models were fit in order to explore associations between marijuana trajectories and academic outcomes measured during the students’ senior year of college. School was treated as a random effect to account for the inter-school correlation of academic outcomes (Donner et al., 1981; Murray and Short, 1995, 1996). Bivariate models were constructed for each covariate and academic outcome. Multivariable models predicting college outcomes from marijuana trajectory groups were estimated, adjusting for characteristics that had a marginal bivariate association (p < 0.20) with the outcomes. Covariates included basic demographic variables and factors shown to be associated with marijuana use in the literature (Bell et al., 1997, Johnston et al., 2015; Buckman et al., 2011; Wechsler et al. 1997; Yusko et al., 2008; McCabe et al., 2005; Mohler-Kuo et al., 2003). Adjusted and unadjusted odds ratios and 95% confidence intervals are presented for both dichotomous outcomes (enrollment in college and graduation on time). Regression coefficients, standard errors, and p-values are presented for the linear model for college grade point average. Models predicting graduation on time and college grade point average were restricted to only students who were currently enrolled or had graduated. Analyses for missing data were carried out using multiple imputation methods (Royston, 2009). First we assessed whether the sample with complete data differed from those with some missing data who additionally contributed to the multiple imputation analysis. This was done for each of the academic outcome models. Results revealed that the sample with full data were more likely to be white compared to those with missing data for the enrollment and GPA outcome models and were more likely to be enrolled in a public institution for the model for graduation on time. No differences were found with regards to gender, Hispanic ethnicity, mother’s education or spending money for any of the academic outcome models. We then conducted multiple imputation analysis on our regression models of marijuana trajectory group predicting academic outcomes. Models were estimated using the GLLAMM procedure and imputations were performed using the ICE procedure in Stata Version 13.1. All analyses use a 5% level of significance.
Since some groups of students were oversampled, all prevalence estimates, bivariate tests, and regression models use weights. Only univariate descriptive statistics on demographic and social characteristics are reported unweighted in order to describe the sample. Sampling weights reflect the inverse probability of selection from the screener survey and include a non-response adjustment. The weights were scaled using the approach of Pfefferman et al. (1998) to account for the students-within-schools design.
Almost half (48.6%) of the sample participating at Wave 6 was female (Table 1). About 16% were nonwhite, and 7% were Hispanic. Nearly 63% reported that their mother earned at least a four year college degree, and 82.6% had at least $100 of spending money per month. Around one third (34.6%) of the sample participated in campus athletics, 26.0% were members or pledges of a sorority or fraternity, and 21.8% participated in religious services on at least a biweekly basis. About one-fourth of students lived on campus, and 45.3% were in a committed relationship. Fifteen percentwt of students reported using cigarettes, 64.2%wt reported heavy episodic drinking, and 8.8%wt used hookah within the past month. In the past six months, 29.8%wt of students used marijuana and 6.6%wt used other illicit drugs. Mean sensation seeking and stress scores were 3.0wt (SD = 0.8wt) and 15.9wt (SD = 6.9wt), respectively.
Most students (97.2%wt) were either enrolled in college or had graduated as of Wave 6. Among students who were still enrolled, 73.9%wt planned to graduate on time. The mean GPA among students enrolled in college as of wave 6 was 3.29wt (SD = 0.47wt). Academic outcomes varied by school and ranged from 92.5%wt to 99.6%wt for college enrollment, 51.1%wt to 92.8%wt for graduating on time, and 3.15wt to 3.39wt for the mean GPA.
The Bayesian Information Criteria statistic increased with the addition of each trajectory group (Table 2). As noted by Nagin and Tremblay (2001), in some applications, the BIC continues to improve, often resulting in the splitting of a large trajectory group into two smaller ones with parallel trajectories. In this instance, it is best to choose the best model based on interpretability and group sizes (no trajectory group significantly below 5% of the sample, though some weighted estimates may fall below 5%). We stopped at five groups because adding a sixth group would have split one of the groups in the five group model into two parallel groups that would not have improved interpretability.
The five group model trajectories are plotted in Figure 1. Among the 2,500 students who participated at Wave 6, 1,495 (69.0%wt) were classified as non-users of marijuana throughout their college careers. The trajectory for non-users remained relatively flat, with 0.03 days of marijuana use, on average, at Wave 1, and 0.04 days of use by Wave 6. Infrequent users (n=460, 16.6%wt) used marijuana occasionally over time. They averaged 0.9 days of marijuana use per month at Wave 1 and increased use slightly over time, to an average of 1.7 days per month by Wave 6. Decreasing users (n=178, 4.7%wt) used marijuana more frequently during their first semester of college (8.9 days per month, on average), and their use declined over time to an average of 1.0 day per month by Wave 6. Increasing users (n=196, 5.8%wt) used marijuana rarely during their first year of college (1.1 days per month, on average), and their use increased during their time in college to an average of 16.6 days per month by Wave 6. Frequent users (n=171, 3.9%wt) used marijuana often throughout their entire college careers, averaging 15.7 days per month at Wave 1, steadily increasing to 21.3 days per month by Wave 5, and dropping slightly to 19.8 days per month, on average, at Wave 6.
Marijuana trajectory groups varied greatly across demographic groups (Table 3). Only 28.6% of frequent users were women, while 68.1% of non-users were female (p < 0.001). Non-users and increasing users had the highest percentages of nonwhites (17%–18%; p=0.002). Sixteen percent of frequent users were Hispanic, while the other four trajectory groups were 4%–8% Hispanic (p=0.006). Non-users were less likely to have more than $100 per month in spending money than the other four groups (78% vs. 85%–89%, p < 0.001).
We also observed differences in social characteristics across marijuana trajectory groups. Non-users were less likely than users to be a member or pledge of a sorority or fraternity (23% vs. 29%–34%, p < 0.001) and more likely to participate regularly in religious activities (33% vs. 5%–11%, p < 0.001), live on campus (30% vs. 8%–16%, p < 0.001), and be in a committed relationship (49% vs. 34%–42%, p = 0.020).
Non-users were also far less likely to use cigarettes (6% vs. 21%–58%, p < 0.001), partake in heavy episodic drinking (54% vs. 86%–94%, p < 0.001), use hookah tobacco (6% vs. 11%–21%, p < 0.001), or use other illicit drugs (1% vs. 7%–44%, p < 0.001) than members of the four marijuana user groups. Mean age of initiation was higher for infrequent (17.4) and increasing users (17.0) than for decreasing (16.2) and frequent (15.7) users (p < 0.001). Trajectory groups also differed by mean sensation seeking score, with non-users having the lowest average (2.9), and frequent users having the highest average (3.7, p < 0.001).
In a bivariate model, both decreasing marijuana users (OR=0.4; CI: 0.2, 0.6) and frequent users (OR=0.4; CI: 0.2, 0.97) were less likely than non-users to be still enrolled or have graduated from college (Table 4). After adjusting for covariates, decreasing marijuana users (AOR=0.3; CI: 0.2, 0.7) and frequent users (AOR=0.4; CI: 0.2, 0.97) were still less likely than non-users to be still enrolled in college or to have graduated. In these adjusted models, students who attended religious services often, students who were not in a committed relationship, heavy episodic drinkers, and hookah users were more likely to be still enrolled in college or to have graduated.
In an unadjusted model, infrequent marijuana users (OR=0.7; CI: 0.6, 0.96), decreasing users (OR=0.5; CI: 0.3, 0.9), increasing users (OR=0.6; CI: 0.4, 0.9), and frequent users (OR=0.4; CI: 0.3, 0.8) were all less likely to plan to graduate from college on time than non-users. After adjusting for covariates, only decreasing users (AOR=0.6; CI: 0.4, 0.99) and frequent users (AOR=0.5; CI: 0.3, 0.97) were still less likely than non-users to plan to graduate from college on time. In these adjusted models, white students and students who lived on campus were more likely to plan to graduate on time.
In the unadjusted linear regression model, infrequent marijuana users (β=−0.10, SE=0.04, p=0.009), decreasing users (β= −0.20, SE=0.06, p=0.001), increasing users (β= −0.34, SE=0.04, p<0.001), and frequent users (β= −0.29, SE=0.05, p<0.001) all had lower GPAs, on average, than non-users. In an adjusted linear regression model, infrequent users (β= −0.08, SE=0.04, p=0.030), decreasing users (β= −0.14, SE=0.06, p=0.015), increasing users (β= −0.25, SE=0.04, p<0.001), and frequent users (β= −0.18, SE=0.05, p=0.001) all had lower GPAs, on average, than non-users. When allowing for multiple comparisons in the multivariable model, increasing users were found have lower GPAs, on average, than infrequent users (β= −0.17, SE=0.06, p=0.003). In the adjusted models, females, white students, and students whose mothers possessed at least a four year college degree had higher GPAs, on average. There was a negative association between GPA and both cigarette use and stress.
We identified five marijuana trajectory groups that were similar to those found in prior research on college students, though we observed fewer students in each user group and more students in the non-user group [69.0% in our study vs. 60.2% “minimal users” reported by Arria et al. (2013b)]. This difference may be due to the fact that the prior study only included students from one college that is not representative of our sample.
All four marijuana use groups reported significantly lower GPAs, on average, than non-users. Even students who used marijuana infrequently exhibited lower academic performance. Students who increased marijuana use had the lowest adjusted average GPA, at 0.25 points lower than non-users. These findings are consistent with prior studies that have established a link between marijuana use and lower GPAs among college students (Arria et al., 2013a, 2015; Bell et al., 1997), even among infrequent users (Buckner et al., 2010). This association may be explained by the fact that marijuana users tend to skip more classes and in turn, earn lower GPAs (Arria et al., 2013a, 2015). The relationship between marijuana use and impaired mental functioning (Caldeira et al., 2008; Churchwell et al., 2010; Hermann et al., 2007; McHale et al., 2008; Ramaekers et al., 2006; Shillington and Clapp, 2001; Vadhan et al., 2007; Wadsworth et al., 2006) could also explain poorer performance among marijuana users.
Decreasing and frequent marijuana users were both less likely to be currently enrolled in college as of senior year and were less likely to plan to graduate on time. These findings suggest that students who frequently use marijuana early in their college career are most at risk of not completing college and delaying graduation. These results are consistent with Arria et al.’s finding that frequent marijuana use during the freshman year of college is associated with delayed graduation from college (2015). Campus prevention efforts should focus on early intervention in order to increase retention.
Males, whites, Hispanics, students whose mothers do not possess a four year degree, and students who report having at least $100 per month in spending money are more likely to be a member of at least one of these two most at-risk marijuana trajectory groups, compared to non-users. Students who participate in a fraternity or sorority, have low religious participation, live off campus, are not in a committed relationship, and use other substances are also more likely to be classified as a decreasing or frequent marijuana user instead of a non-user. Frequent and decreasing users also initiate marijuana at an earlier age than infrequent and increasing users, further underscoring the need for early intervention. In an earlier study of this sample, we found that students who have more spending money, attend church rarely or never, and use other substances are more likely to have initiated marijuana before attending college (Suerken et al., 2014). These students might have started to use regularly prior to college entry and might especially benefit from screening and early intervention. Our previous research also showed that Hispanic students were more likely to initiate marijuana during the first year of college (Suerken et al, 2014). In the current study, Hispanic students were more likely to be frequent marijuana users during college, further illustrating the need for more research into the association between ethnicity and marijuana use.
We acknowledge several limitations to this study. Our study included students from 11 four-year colleges in 2 states and may not be generalizable to college students in other areas of the country or to small private schools, given that most of the colleges that participated in our study were public institutions. Future research should consider the school environment, given that some types of institutions may monitor student performance more carefully. Graduation date was self-reported and based on future plans that may not come to fruition. Some students may take longer to graduate than expected, and some students who plan to delay graduation may eventually drop out of college. GPA was also self-reported and may have been rounded or remembered incorrectly. Self-reported college GPAs are commonly inflated, especially among students with lower grades (Kuncel et al., 1995). Findings on graduation time and GPA may be conservative because they are restricted to the subset of students who were enrolled or had graduated. Marijuana use may have been underreported (Akinci et al., 2001; Delaney-Black et al., 2010; Gruenwald and Johnson, 2006; Wagenaar et al., 1993), although measuring drug use in a school or research setting is more likely to produce honest responses (Kandel et al., 2006). The sample may also be subject to selection bias, given that substance users may have been less likely to continue participation in the study (McCoy et al., 2008).
Although all marijuana user groups exhibited lower academic performance, students who use marijuana frequently at the beginning of the college career are at the highest risk for dropping out or delaying graduation. Campus prevention efforts should focus on early intervention and target demographic and social groups who are most at risk for early frequent use. Academic assistance centers should screen students for frequent marijuana use during the first semester of college in order to identify students who may be struggling academically. Future research should investigate the association between marijuana trajectories and post-college outcomes.
Role of Funding Source
This research was supported by Award Number R01CA141643 from the National Cancer Institute. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.
ContributorsCynthia Suerken wrote the first draft of the manuscript and conducted the literature search and statistical analyses. Beth Reboussin oversaw the statistical analyses. Beth Reboussin, Kate Egan, Erin Sutfin, Kimberly Wagoner, John Spangler, and Mark Wolfson contributed to the study design. All authors reviewed and edited drafts of the manuscript and approved of the final version.
Conflict of Interest
All authors declare that they have no conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.