How do college students use their free time? A latent profile analysis of leisure activities and substance use.

Abstract

College is a critical period of transition to independence and the substantial amount of time that students have to participate in leisure activities may be conducive to substance use. However, little is known about the associations between leisure activities and substance use over time, or whether these associations differ by residential status (i.e., living with parents vs. on their own). Using latent profile analysis, this study found six distinct profiles of leisure activity participation in a racially/ethnically diverse sample of college students (N=1,207). Overall, profiles with medium levels of leisure activity participation were associated with more alcohol use, heavy drinking, and marijuana use one year later; whereas profiles with the lowest levels of leisure activity participation were associated with more cigarette use one year later. Identifying mechanisms through which leisure activities influence substance use can help inform prevention efforts to either reduce risks associated with participation or support protective effects.

Research on leisure activities has consistently found that participation is associated with both positive (e.g., psychological well-being) and negative (e.g., substance use) health outcomes (Piko & Vazsonyi, 2004; Trainor, Delfabbro, Anderson, & Winefield, 2010). The influence of leisure activities is particularly important to the health of emerging adults (ages 18 to 25) because emerging adults have more free time to participate in these types of activities. Emerging adults are less restricted by institutions (e.g., a set school schedule; having to follow family curfews and rules); at the same time, in recent decades, emerging adults are taking longer to transition into adult roles such as work, marriage, and parenthood (Kins & Beyers, 2010). Thus, they have less structure in their schedule, and may also have more free time than young people of their age did in prior generations. One hypothesized outcome of lack of structure and more free time during emerging adulthood is high rates of risky behaviors, including substance use (Arnett, 2005). Indeed, researchers have shown that alcohol use (Abadi, Shamblen, Thompson, Collins, & Johnson, 2011; Cho et al., 2015), heavy drinking (Patrick, Terry-McElrath, Kloska, & Schulenberg, 2016), and marijuana use (Abadi et al., 2011; Chen, Yu, Lasopa, & Cottler, 2017) during emerging adulthood. Although a lack of structure is one of the main characteristics of emerging adulthood, little is known about the association between what emerging adults may do in their free time and substance use during this developmental period. Moreover, most investigators studying the association between leisure activities and substance use have focused on adolescents. As such, little is known about longitudinal associations between leisure activities and substance use during the critical period of emerging adulthood.

Leisure Activities and Substance Use in College Students

College students have a substantial amount of free time, with an average of 20 hours per week spent on leisure activities, such as socializing with friends and participating in extracurricular activities (Fosnacht, McCormick, & Lerma, 2018), which is more than time spent on educational activities (U.S. Buerau of Labor Statistics, 2016). Researchers examining leisure activities in college students have found that participation in these types of activities is associated with both mental and behavioral health (Doerksen, Elavsky, Rebar, & Conroy, 2014). Despite the substantial amount of time that college students may spend participating in leisure activities, few studies have addressed how participation in activities during their free time may be associated with substance use. Of the studies that have examined the association between leisure activities and substance use, most have only assessed alcohol use, and have neglected to address other types of substances, such as marijuana and cigarettes (Finlay, Ram, Maggs, & Caldwell, 2012). The increase in substance use, including alcohol, marijuana, and cigarettes during emerging adulthood is especially pronounced in the transition from high school to college (Derefinko et al., 2016; Fromme, Corbin, & Kruse, 2008; Patrick et al., 2016; Tucker, Orlando, & Ellickson, 2003; White et al., 2006), as many young adults are moving away from home the first time, navigating a new social environment, and meeting new peers (Borsari, Murphy, & Barnett, 2007; Derefinko, Bursac, Mejia, Milich, & Lynam, 2018; Derefinko et al., 2016). For example, the National College Health Assessment of 2019 found that 55.8%, 23.1%, and 6.3% of college students reported using alcohol, marijuana, and cigarette in the last 30 days respectively (American College Health Association, 2019).

Social norms theory has been applied to understand drinking and drug use behaviors in undergraduate students. The social norms theory posits that college students’ drinking and drug use behaviors are highly influenced by peers (Berkowitz, 2004; Perkins, 2002). Peer influence appears to be even more influential during emerging adulthood than adolescence in the area of substance use (Abadi et al., 2011). Perceptions of peer use among college students influence drinking (Barnett et al., 2017; Borsari et al., 2007; DiGuiseppi et al., 2018), marijuana use (Edwards, Witkiewitz, & Vowles, 2019), and cigarette smoking (Noland et al., 2016), with the influence of perceived norms strongest during the first year of college (Noland et al., 2016; Turrisi, Padilla, & Wiersma, 2000). Furthermore, college students often overestimate peer alcohol use, and the exaggerated perception of peer use is associated with higher risk for drinking (Barnett et al., 2017; Cox et al., 2019; Dumas, Davis, & Neighbors, 2019; Perkins, Haines, & Rice, 2005). Recent research on the influence of social media provides further evidence to support the importance of peer norms in college student drinking. First-year college students were more likely to drink after viewing peers’ alcohol-related postings on social media (Boyle, LaBrie, Froidevaux, & Witkovic, 2016). In sum, college students’ alcohol use and drug use are highly subjected to peer norms and their overestimation of peers’ drinking and drug use sometimes lead to increased use of alcohol, marijuana, and cigarettes.

By contrast, if college students spend their free time participating in leisure activities in which substance use is not central to that gathering and when they are not subjected to peer influence, such as spending time with family, they are less likely to engage in substance use (LaBrie et al., 2016; Rulison, Wahesh, Wyrick, & deJong, 2016; White et al., 2006). Although parents are expected to exert less influence on their children during emerging adulthood, there is evidence to suggest that parents still affect college students’ substance use during their first year in college (Borsari et al., 2007; Rulison et al., 2016). For example, one study found that consistent parent-child communication about risk behaviors reduced heavy drinking and marijuana use in college students during their transition to college (LaBrie et al., 2016). In addition, spending time with family in a substance-free environment may also protect college students from using substances as it could decrease students’ exposure to substance use.

Researchers have found that leisure activities, such as sports or volunteering, may also affect college student substance use, especially in their first year as they are experiencing increased independence and decreased adult supervision (e.g., (Meshesha, Dennhardt, & Murphy, 2015). For example, one study using the daily diary method to examine the association between different types of leisure activities and alcohol use in first year college students found that students who spent more time in volunteer activities reported lower levels of alcohol use during a 14-day period (Finlay et al., 2012). In comparison, students who spent more time in athletic activities reported higher levels of alcohol use (Finlay et al., 2012). Furthermore, participation in intercollegiate or intramural sports was associated with increased alcohol use among student athletes (Martens, Dams-O’Connor, & Duffy-Paiement, 2006; Mastroleo, Barnett, & Bowers, 2019). A recent qualitative review of studies examining drinking as a leisure activity found that social drinking is prevalent for youth who participate in sports, and social drinking often contributes to the socio-cultural aspects of sports (Burns & Gallant, 2018). Thus, it is crucial to distinguish between sports and non-sports related leisure activities when examining associations with substance use.

Residence Status, Leisure Activities, and College Student Substance Use

Recent social and economic changes have led to a substantial increase in the number of college students living at home. The most recent Sallie Mae’s National Study of College Students and Parents found that 50% of college students were living at home in 2017. Even though half of the college population is living at home, research on substance use in college students has mostly focused on students living on campus or living away from home (Kim & Rury, 2011). Researchers have found that living at home is associated with less frequent use of alcohol (Simons-Morton et al., 2016; White, Fleming, Kim, Catalano, & McMorris, 2008); however, the influence of residence status on other types of substances (i.e., cigarette smoking and marijuana) is unknown. No studies have examined the longitudinal association between leisure activities and substance use in college students who live at home. One cross-sectional study found that being friends with peers who participated in non-sports leisure activities (e.g., school clubs) was associated with lower levels of alcohol- and drug-related problems in college students who commuted to school, including students who lived at home (Wax, Hopmeyer, Dulay, & Medovoy, 2019). However, the study did not assess actual participation in leisure activities. Further research is needed to understand whether residence status might influence associations between leisure time use and substance use among college students.

Current Study

To date, no longitudinal studies have examined associations between leisure activities and substance use, despite college being a critical period of transition to independence, and the substantial amount of time that college students have to participate in leisure activities that may be conducive to substance use. As college students are likely to engage in multiple leisure activities, it is important to identify patterns of participation and how these patterns may be associated with substance use (Agans et al., 2014; Zarrett et al., 2009). We therefore used latent profile analysis to identify patterns of participation in four leisure activities (i.e., socializing with friends, spending time with family, sports-related activities, and non-sports related activities) among a sample of racially/ethnically diverse college students during freshmen year. We wanted to understand: (1) what are the associations between patterns of leisure activity participation identified during freshmen year and substance use (i.e., cigarette, alcohol use, heavy drinking, and marijuana use) during sophomore year of college, and (2) how do these associations differ depending on residence status? Informed by cross-sectional work addressing specific substances, we hypothesized that the profile with the highest levels of non-sports activities and time with family would report the lowest levels of longitudinal use across all four substances (LaBrie et al., 2016; Rulison et al., 2016); whereas the profile with the highest levels of sports activities and time socializing with friends would report the highest level of longitudinal use (DiGuiseppi et al., 2018; Martens et al., 2006; Mastroleo et al., 2019).

Methods

Procedures and Participants

Participants in the current study were originally recruited from 16 middle schools in three school districts in Southern California to participate in an ongoing longitudinal study (D’Amico et al., 2012; D’Amico et al., 2012). Schools were selected to obtain a racially and ethnically diverse sample. A subset of the original sample (N = 6,509 6^th and 7^th graders from the original sample) has completed annual surveys since 2008. As youth graduated from middle school to high school between Waves 5 and 6, they transitioned from 16 middle schools to over 200 high schools. At wave 6 (Spring 2013-Spring 2014), 61% of the sample participated in the follow-up survey (D’Amico et al., 2016). We retained 80% of the sample from Wave 6 to 7, 91% from Wave 7 to 8, and 89% from Wave 8 to 9. If a participant did not complete a wave of data collection, they were still eligible to complete all subsequent waves. That is, they did not “dropout” of the study once they missed a survey wave; rather we fielded the full sample at every wave so that all participants had an opportunity to participate in each survey. Failure to complete a certain wave was not significantly associated with demographics or risk behaviors, such as drinking and marijuana use (D’Amico et al., 2016; Tucker et al., 2019). The analytic sample for present study includes participants enrolled in college as freshmen during Wave 8 (2015–2016) who were also enrolled as sophomores in college in Wave 9 (N = 1,207; 2016–2017). All procedures were approved by the institution’s IRB, and we have a certificate of confidentiality protecting the data.

Measures

Sociodemographic Measures.

Students answered questions about age, gender, race/ethnicity, and employment status. Ethnicity was measured by first assessing whether students identified as Hispanic. They then reported racial categories, including White, Black, Asian, and Other. We then categorized students in the following racial/ethnic groups: Hispanic, non-Hispanic Black, non-Hispanic White, non-Hispanic Asian, and non-Hispanic Other. Students reported their employment status (full time, part time, not working but looking for a job, and not working but not looking for job), and responses were recoded to create a binary variable (1=employed and 0=not employed). Data on mother’s education (didn’t finish high school, college, some college, don’t know) were also collected as a proxy measure of family socioeconomic status (Korupp, Ganzeboom, & Van der Lippe, 2002) and were used as covariate in the analyses. Students also reported on residence status (living with parents = 1 or not living with parents = 0). Finally, students reported Greek organization affiliation (1=Greek member and 0=non-Greek member). Greek affiliated college students reported more frequent alcohol use and heavy drinking during the first two years of college (Capone, Wood, Borsari, & Laird, 2007).

Leisure Activity Participation.

Students reported how often they: (1) go to parties or other social events with friends, (2) hang out with friends, (3) do things with family, (4) talk with their parent or guardian about personal experiences, (5) participate in sports-related activities (not counting regular P.E. class), and (6) participate in non-sports activities such as school clubs and band (but not counting sports) (Tucker, Ellickson, Collins, & Klein, 2006). All items were rated on a scale of 1 to 7 (1=Not at all, 2=less than once a month, 3=about once a month, 4=2–3 times a month, 5=once a week, 6=more than once a week but not everyday, and 7=everyday). Four leisure categories were included in the current study: (1) socializing with friends, (2) spending time with family, (3) sports, and (4) non-sports. Two items were averaged to obtain a score for socializing with friends (go to parties or other social events with friends, and hang out with friends); and similarly, two items were averaged to obtain a score for spending time with family (do things with family, and talk with their parent or guardian about personal experiences). Sports and non-sports activities were each assessed by one item as noted above.

Substance Use.

Past month substance use was assessed using items from Monitoring the Future: “During the past month (30 days), how many days did you use 1) cigarettes, 2) at least one drink of alcohol, 3) five or more drinks of alcohol in a row, that is, within a couple of hours, and 4) marijuana (pot, weed, grass, hash, bud, sins)” (Johnson, O’Malley, Miech, Bachman, & Schulenberg, 2016). Participants chose one of the following responses: 0=0 days, 1=1 day, 2=2 days, 3=3–5 days, 4=6–9 days, 5=10–19 days, and 6=20–30 days. These response options were rescaled to the mid-point of number of days (e.g., 3–5 days = 4). This established approach (NIAAA, 2003; Osilla et al., 2014) yields a continuous score ranging 0–25 which allows for model estimates to be interpreted as the number of days a youth used a substance in the past month.

Data Analytic Strategy

To investigate the association between leisure activity participation during freshmen year and substance use one year later, we conducted latent profile analysis (LPA) using the 3-step procedure with auxiliary variables (Asparouhov & Muthen, 2014) in Mplus v8 (Muthén & Muthén, 1998–2017). Models were estimated using maximum-likelihood with robust standard errors allowing for missing data to be handled by full-information maximum likelihood. We specified an ascending number of group solutions up to ten groups. To evaluate the fit of each group solution, we examined the following fit statistics: log-likelihood, Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), sample size-adjusted Bayesian Information Criterion (aBIC), Parametric Bootstrapped Likelihood Ratio Test (BLRT), and Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (VLMRT). A solution with lower log-likelihood, AIC, BIC, and aBIC is a better fit. BLRT and VLMRT test whether adding a latent group significantly improves the overall fit of the solution (Nylund, Asparoutiov, & Muthen, 2007). In addition, we considered interpretability and conceptual relevance when determining the final solution.

Prior to testing the predictive power of the groups, we first examined whether students who lived with parents and students who did not live with parents were equally represented across the groups. Then, we tested for meaningful differences between groups on substance use outcomes. This was accomplished within the 3-step LPA estimation using the BCH procedure with auxiliary variables to test the equality of means on substance use measures between two groups. We tested whether groups would predict subsequent substance use, controlling for gender, race/ethnicity indicators, mother’s education, employment status, and Greek membership. Age was not included as a covariate because there was relatively little variation in age (98% of participants were between the ages of 17 and 19).

Results

Descriptive Statistics

Table 1 reports baseline characteristics of participants. The average age of students who did not live with their parents at Wave 8 was 18.40 (SD = 0.62), and the average age of the students who lived with parents at Wave 8 was 18.11 (SD = 0.7). Approximately 92% of participants who did not live with their parents reported that their mother completed at least high school. In comparison, 82% of participants who lived with their parents reported that their mother completed at least high school. At Wave 8, similar percentages of students who did not live with their parents and their counterparts were employed at least part-time (42%). Students who did not live with their parents and those who lived with their parents did not differ in heavy drinking (t = 1.28, p > .05), marijuana use (t = 1.83, p > .05), or cigarette smoking (t = 1.30, p > .05) at Wave 9. However, students who did not live with their parents reported more frequent alcohol use than their counterparts (t = 4.25, p < .01).

Table 1.

Baseline Characteristics of the College Student Sample

	LIVING WITH PARENTS (N = 1010)	NOT LIVING WITH PARENTS (N = 197)
Variable	M(SD) or n(%)	M(SD) or n(%)
Wave 8 Demographics
Age	18.11 (0.70)	18.40 (0.62)
Female n(%)	575 (56.9%)	116 (58.9%)
Hispanic n(%)	412 (40.8%)	42 (21.3%)
Non-Hispanic White n(%)	203 (20.1%)	64 (32.5%)
Asian/Pacific Islander n(%)	259 (25.6%)	52 (26.4%)
Other n(%)	118 (11.7%)	33 (16.8%)
African American n(%)	18 (1.8%)	6 (3.1%)
Employed n(%)	428 (42.4%)	81 (41.1%)
Wave 8 Substance Use
Cigarette Use	2.10 (5.63)	3.96 (7.44)
Alcohol Use	2.50 (3.81)	4.71 (5.13)
Heavy Drinking	1.96 (3.45)	3.87 (5.04)
Marijuana Use	4.95 (7.79)	6.04 (8.56)
Wave 9 Substance Use
Cigarette Use	2.58 (5.36)	3.22 (6.10)
Alcohol Use	3.06 (4.33)	4.55 (4.99)
Heavy Drinking	2.05 (2.99)	2.41 (3.21)
Marijuana Use	4.81 (7.72)	6.91 (9.57)
Wave 8 Participation in Recreational Activities
Go to parties or other social events	2.92 (1.65)	3.78 (1.61)
Hang out with friends	4.93 (1.66)	5.66 (1.48)
Do things with family	4.34 (1.65)	3.71 (1.77)
Talk with parent about personal experiences	4.32 (2.06)	4.43 (1.95)
Do homework 2 more hours a day	5.12 (1.91)	5.55 (1.61)
Participate in extra-curricular school activities	3.61 (2.24)	3.84 (2.19)
Participate in sports	3.53 (2.36)	3.62 (2.27)

Latent Profile Analysis of Free Time Participation

Table 2 presents the log-likelihood, AIC, BIC, aBIC, BLRT, and VLMRT statistics of all solutions (one-group to ten-group). Although the seven-, eight-, nine-, and ten-group solutions had the lowest aBIC, these solutions also had lower entropy scores. In contrast, the entropy scores of the five- and six-group solutions were both higher (.93), which indicates high accuracy of classification. Also, the aBIC scores of five- and six-group solutions were acceptable. We therefore further examined the five- and six-group solutions for interpretability and conceptual clarity, deciding on the six-group solution on that basis.

Table 2.

Model Fit Statistics and Entropy for Latent Profile Analyses (LPAs) of Recreational Activities

Model	Free Parameters	Log-likelihood	AIC	BIC	aBIC	BLRT	VLMRT	Entropy
One Class	8	19648	19664	19704.80	19679.40	-	-	-
Two Class	13	18601.90	18627.90	18694.10	18652.80	0	0	.92
Three Class	18	18148.40	18184.40	18276.10	18218.90	0	0	.95
Four Class	23	17983.90	18029.90	18147	18074	0	0	.96
Five Class	28	17671.50	17727.50	17870.10	17781.10	0	0	.93
Six Class	33	17480.80	17546.80	17714.90	17610.10	0	0	.93
Seven Class	38	17395.10	17471.10	17664.60	17543.90	0	0	.88
Eight Class	43	17299.70	17385.70	17604.70	17468.10	0	0	.87
Nine Class	48	17220.70	17316.70	17561.20	17408.70	0.0006	0.0008	.90
Ten Class	53	16703.50	16809.50	17079.40	16911	0.08	0.09	.93

Note. AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; aBIC = Sample size-adjusted BIC; BLRT = Parametric Bootstrapped Likelihood Ratio Test; VLMRT = Vuong-Lo-Mendell-Rubin Likelihood Ratio Test.

Of the four leisure activities tested, participants from all six groups spent about the same amount of time engaging in non-structured activities (socializing with friends and spending time with family). The main difference among the six groups was the varying levels of involvement in campus structured activities (CSA): sports and non-sports activities. Hence, the profiles were characterized by participation in sports and non-sports activities. The six groups are: 1) No CSA; 2) Low Overall CSA; 3) Medium Overall CSA; 4) High Overall CSA; 5) Sports Only; and 6) Non-Sports Only (see Figure 1).

Figure 1.

Average Friends, Family, Non-sports Activities, and Sports participation for six-class latent profile analysis model. SA = Structured Activities.

Students in the No CSA group (30.7% of the sample) were generally not involved in any campus structured activities. Students in the Low Overall CSA (7.4% of the sample) did participate in some campus structured activities; however, they spent a relatively smaller amount of time participating in sports and non-sports activities in comparison to the other groups. Relative to the rest of the participants, students in the Medium Overall CSA (13.4% of the sample) reported a medium level of participation in both campus structured activities (sports and non-sports). The fourth group, High Overall CSA, (23.5% of the sample) reported the highest levels of campus structured activity participation. Sports Only made up 11.1% of the sample. These students reported the highest level of sports activity participation but had low levels of participation in non-sports activities. Finally, the Non-sports Only includes 13.8% of the sample. Students in this group reported high levels of non-sports activity participation but very little involvement in sports activities. There was no significant association between residential status and leisure group membership, χ²(5) = 6.76, p > .05; thus, students who lived with their parents and students who did not live with their parents were equally represented across the six groups.

Leisure Activities and Substance Use – Students Living with Parents

Table 3 presents both concurrent analyses (Wave 8 leisure activities and Wave 8 substance use) and longitudinal analyses (Wave 8 leisure activities and Wave 9 substance use) by groups. Furthermore, the table presents the results separately for students who lived with parents and for students who did not live with parents. As can been seen from the table, the concurrent and longitudinal associations are very similar. Given the lack of longitudinal studies in this area, we therefore focus on the longitudinal associations in the results section.

Table 3.

Past Month Substance Use Means and Standard Deviations by Groups

Students living with parents
Substance	No CSA	Low Overall CSA	Medium Overall CSA	High Overall CSA	Sports Only	Non-sports Only
Concurrent
Cigarettes	2.22(.74)	11.29(6.59)	1.80(1.10)	1.06(.47)	.78(.50)	.94(.78)
Alcohol	2.13(.30)	4.19(1.73)	2.55(.45)	3.08(.42)	2.19(.46)	1.87(.35)
Heavy drinking	1.79(.34)	3.13(1.78)	1.58(.39)	2.47(.41)	2.17(.66)	.82(.21)
Marijuana	4.96(.81)	5.43(2.31)	3.67(1.02)	4.26(.78)	6.05(1.48)	3.57(1.06)
Longitudinal
Cigarettes	3.33(.83)	4.90(4.66)	2.19(.89)	1.21(.34)	1.74(.69)	1.23(.56)
Alcohol	2.56(.32)	2.26(.76)	3.90(.60)	3.47(.41)	2.29(.37)	3.25(.52)
Heavy drinking	1.83(.28)	1.95(1.10)	2.63(.52)	2.03(.28)	1.73(.34)	2.29(.52)
Marijuana	5.01(.80)	5.20(2.10)	3.78(.95)	4.61(.74)	4.85(1.00)	3.40(.99)
Students not living with parents
Substance	No CSA	Low Overall CSA	Medium Overall CSA	High Overall CSA	Sports Only	Non-sports Only
Concurrent
Cigarettes	4.59(1.39)	1.03(2.28)	4.07(1.86)	1.88(1.09)	4.92(2.96)	.20(.18)
Alcohol	4.23(.74)	2.80(1.52)	6.60(1.20)	5.11(.95)	3.44(.86)	3.85(.92)
Heavy drinking	2.52(.67)	1.81(1.94)	7.19(1.44)	3.80(.93)	3.46(.76)	3.32(1.32)
Marijuana	6.44(1.56)	7.78(4.17)	8.27(1.98)	3.53(1.07)	4.15(1.67)	3.33(1.51)
Longitudinal
Cigarettes	4.46(1.37)	12.42(11.69)	2.89(1.41)	.96(.39)	2.61(1.50)	1.57(.63)
Alcohol	3.73(.73)	1.98(.85)	6.73(1.18)	3.93(.66)	7.95(1.29)	2.02(.44)
Heavy drinking	2.15(.56)	.88(.67)	3.51(.83)	2.15(.45)	3.34(.93)	1.50(.63)
Marijuana	6.83(1.48)	4.13(2.78)	8.99(2.04)	5.33(1.34)	4.75(1.92)	3.30(1.47)

Note. CSA = Campus Structured Activities. “Concurrent” refers to analyses that assessed the association between leisure activities and substance use in the same year (wave 8); “Longitudinal” refers to analyses examining the association of leisure activities in wave 8 with substance use in wave 9. Participants chose one of the following responses for past month substance use: 0=0 days, 1=1 day, 2=2 days, 3=3–5 days, 4=6–9 days, 5=10–19 days, and 6=20–30 days. These response options were rescaled to the mid-point of number of days (e.g., 3–5 days = 4).

Cigarette Smoking.

Analysis found that the Low Overall CSA group reported the highest frequency of smoking one year later (Table 3), but this group did not statistically differ from the others due to its large standard deviation. However, the No CSA group reported more frequent cigarette smoking one year later than the Non-sports Only group (p < .05) and High Overall CSA group (p < .05).

Alcohol Use.

Overall, we found that students who participated in medium to high levels of overall campus structured activities reported drinking the most frequently one year later. Medium Overall CSA drank more frequently than Sports Only (p < .05). In addition, High Overall CSA group reported more frequent alcohol use than Sports Only (p < .05).

Heavy Drinking.

No significant differences were found in heavy drinking among the six groups.

Marijuana Use.

Frequency of marijuana use did not significantly differ between groups.

Summary.

In summary, students who did not participate in campus structured activities, including both sports and non-sports activities, during freshmen year tended to report more frequent use of cigarettes during sophomore year. Students who participated in medium to high levels of overall campus structured activities in their freshmen year reported the highest frequency of alcohol use in their sophomore year. Participation in leisure activities was not associated with heavy drinking or marijuana use one year later.

Leisure Activities and Substance Use – Students Not Living with Parents

Cigarette Smoking.

The No CSA group reported more frequent cigarette smoking than the High Overall CSA group (p < .05). Other groups did not significantly differ from each other.

Alcohol Use.

Groups differ significantly in their risk for alcohol use during sophomore year. Students from the Medium Overall CSA group and Sports Only group reported the most frequent drinking during sophomore year. Students from the Medium Overall CSA group and Sports Only group drank more frequently than those from the No CSA (p < .01), Low Overall CSA (p < .01), High Overall CSA (p < .01), and Non-Sports Only groups (p < .01). Finally, No CSA group and High Overall CSA group reported more frequent drinking than Non-sports Only group (p < .05).

Heavy Drinking.

Longitudinally, the Medium Overall CSA and Sports only groups reported the most frequent heavy drinking. Students in both groups reported more frequent heavy drinking than students in Low Overall CSA group (p < .05).

Marijuana Use.

Students in Medium Overall CSA group reported more frequent marijuana use than students in Non-sports Only (p < .05) during sophomore year.

Summary.

In summary, the medium level of participation in campus structured activities was associated with more frequent alcohol use, heavy drinking, and marijuana use overtime. In addition, students who were not participating in any campus structured activities smoked more frequently than those who participated in campus structured activities at high level. Finally, participation in sports only activities during freshmen year was associated with more frequent alcohol use, heavy drinking, and marijuana use during sophomore year.

Discussion

Following the transition out of high school, emerging adults who attend college are at risk of increased substance use. One of the many new tasks required of college students is to learn how to utilize their free time wisely and effectively. The present study tested longitudinal associations between participation in leisure activities and substance use in college students during their first two years at school. Latent profile analysis identified six distinct groups of students based on their involvement with two categories of leisure activities during their freshmen year: non-structured activities (socializing with friends and spending time with family) and campus structured activities (participation in sports and non-sports activities).

Leisure Activities, Substance Use, and Residence Status

The longitudinal association between leisure activity participation and substance use varied slightly depending on student’s residence status. Leisure activity participation was associated with using all four types of substances for students who did not live with their parents, whereas it was only associated with two types of substances (cigarette smoking and alcohol use) for students who lived with their parents. Results from the 2013–2014 National Survey of Student Engagement indicate that first year students who live farther than walking distance from campus spend fewer hours per week on extracurricular activities than students who live on campus or within walking distance from campus (Gonyea, Graham, & Fernandez, 2015).Thus, it is likely that campus structured activities, as measured in this study, are not as relevant to students who lived with their parents because they may be less likely to participate in them. Off-campus leisure activities, on the other hand, were not captured in the current study, and these activities might be more prevalent among students who lived at home. Thus, future research should include community-based structured leisure activities to fully capture the types of activities that college students may participate in who live at home with their parents. Participation in community service activities is associated with lower levels of alcohol use in first year college students (Finlay et al., 2012).

Campus Structured Leisure Activity Participation and More Frequent Alcohol Use, Heavy Drinking, and Marijuana Use

In contrast to the constructive activity hypothesis (Kleiber, Larson, & Csikszentmihalyi, 1986; Larson & Kleiber, 1993), the present study found that medium levels of participation in campus structured activities (sports and non-sports) were associated with more frequent substance use regardless of residential status. Findings somewhat contradict previous work suggesting that participation in non-sports structured leisure activities is protective against substance use (Correia, Benson, & Carey, 2005; White et al., 2008), but the motivation behind participation may be important to consider. For example, college students who participate in campus structured activities at a medium level (which is only 2 to 3 times a month in this study), might be more motivated to engage in these activities to socialize with peers compared to those who participate at a higher level (or do not participate at all). A recent study found that almost 25% of college students reported socializing and having a good time with friends as their motivations to use alcohol (Cooper, Weybright, Bumpus, Hill, & Agley, 2018). In addition to measuring frequency of participation, research that explores motivations behind college students’ participation in sports and non-sports activities is needed to illuminate the current findings.

Participation in sports only structured activities in freshmen year was associated with more frequent alcohol use and heavy drinking during sophomore year among students who did not live with their parents. Overall, findings support other work showing an association between leisure and intramural sports participation and heavy drinking in college students (Andes, Poet, & McWilliams, 2012; Barry, Howell, Riplinger, & Piazza-Gardner, 2015; Ward & Gryczynski, 2007). Peer norms and behaviors are important predictors of heavy drinking in college students (Borsari et al., 2007; DiGuiseppi et al., 2018). Moreover, college students have a tendency to overestimate alcohol consumption among their peers (Barnett et al., 2017; Hartzler & Fromme, 2003), including their teammates (Grossbard, Hummer, LaBrie, Pederson, & Neighbors, 2009), and these misperceptions associate strongly with actual drinking behavior (Hummer, LaBrie, & Lac, 2009). Social norms campaigns that aim to change misconceptions about peer alcohol use can be particularly effective for college students who are involved with sports (Fearnow-Kenney et al., 2016; Labrie, Hummer, Huchting, & Neighbors, 2009).

Campus Structured Leisure Activity Participation and Less Frequent Cigarette Use

The association between leisure activity participation and cigarette use followed a different pattern than the other three substances. Participation in campus structured leisure activities (sports and non-sports) was protective against cigarette smoking in college students regardless of residence status. College students with low levels of campus structured leisure activity participation may be experiencing social isolation and having difficulty transitioning to college (Bohnert, Aikins, & Edidin, 2007). Indeed, tobacco use is associated with higher levels of stress and depressive symptoms and poorer sleep quality in college students (Bandiera, Loukas, Wilkinson, & Perry, 2016; Boehm, Lei, Lloyd, & Prichard, 2016). Tobacco prevention efforts could therefore also focus on social adjustment/coping and provide students with opportunities to engage with peers in substance-free environments.

The present study found that participation in campus structured leisure activities may be either beneficial or harmful depending on the type of substance. Findings highlight the importance of context, including the environment in which college students may choose to participate in leisure activities and their use of different types of substances. For example, alcohol use among college students typically happens as part of social activity (Borsari et al., 2007); in contrast smoking cigarettes tends to be more of a solitary activity (Nichter, Nichter, Carkoglu, & Network, 2007). If campus structured leisure activities provide opportunities for college students to spend time with peers who are prosocial and allow them to learn effective coping skills, such participation may prevent college students from both drinking and smoking. For example, Caldwell and Darling (1999) found that youth were only more likely to drink and use marijuana when participating in leisure activities if they perceived their peers were endorsing substance use. In addition, when leisure activities promote social support and positive mood, participation in such activities is related to short-term reduction of stress and long-term well-being (Iwasaki, 2003, 2006). In addition, an interview study found that male college students reported experiencing less stress when they participated in action-oriented (both physical and mental) leisure activities (Blanco & Robinett, 2014). Our findings provide additional evidence supporting that participation in campus structured leisure activities can be beneficial to student well-being if the activities allow for the development of positive coping skills (e.g., interaction with prosocial peers, receiving social support).

Implications

Findings from the current study could be used to design structured leisure activities for college students during their transition to college to prevent substance use and increase overall academic success. Our findings suggest that it is important to have activities for college students as those who participate in leisure activities tend to report less cigarette use; however, these activities should provide an opportunity for students to get together in an alcohol-free environment. For example, a program called LateNight Penn State (LNPS), a prevention program that provided alcohol-free leisure activities to undergraduate students, was effective in decreasing students’ alcohol use (Layland, Calhoun, Russell, & Maggs, 2018). Students reported drinking less alcohol on days when they participated LNPS activities than on days when they did not participate. Another qualitative study found that when students learned about the importance of positive, drug-free leisure activities in class and they had time to reflect on their behaviors, they made healthier choices around substances (Yarnal, Qian, Hustad & Sims, 2013). The key is to provide college students with leisure activities that give them opportunities to develop a healthy lifestyle and achieve academic success.

Limitations and Future Directions

The present study identified significant patterns between structured leisure activities and substance use in college students with different residence statuses during the first two years of college, but it is not without limitations. First, the leisure participation measure is limited as we only measured leisure participation once. However, college students’ participation in leisure activities can change weekly (Doerksen et al., 2014); as such, assessing participation more frequently may allow a more precise estimate of the association between leisure activity participation and substance use. Furthermore, it may benefit future research to use more precise measures of different sports and non-sports activities, as well as include items assessing community-based structured leisure activities and (for students who live off campus with their parents) leisure activities in their neighborhood or close to home (Stuber, 2011), to better understand what may be protective for substance use during the early college years. Second, 85% of participants reported that they lived with their parents. While this suggests that many were community college students (Ma & Baum, 2016), we did not have information about the type of school they were attending (e.g., community college, four-year college). Such additional information is needed in future studies to examine whether patterns of leisure activities and substance use differ by type of school. For example, community students may have less time for structured leisure activities as many of them work part-time and need to care for dependents (McClenney, 2007). Unstructured leisure activities, such as spending time with family, may be of greater importance. One study found that 77% of community college students reported talking to family and friends as their most common choice of stress coping strategy (Pierceall & Keim, 2007).

Moreover, a more balanced sample of students who live with their parents and those who live without their parents is needed in future studies to ensure more accurate estimates of group differences. With a more balanced sample, an interaction between profiles and residential status could be tested to examine the extent to which profile configuration differ significantly across the two groups of students. In addition, previous research suggests that college students are often motivated to participate in different types of sports activities (e.g., intramural, recreational, intercollegiate) for different intrinsic and extrinsic reasons, and the motivation for participation can explain associations between participation and substance use (Rockafellow & Saules, 2006).

This is one of the few studies that assessed the longitudinal association between leisure participation and substance use among college students. Overall, we found that participation in leisure activities was positively associated with alcohol and marijuana use but negatively associated with smoking during the early college years. Participation in leisure activities increases college students’ exposure to peers and peer use is associated with more frequent alcohol and marijuana use (Barnett et al., 2014). In contrast, cigarette smoking is perceived to be less socially acceptable among college students (Noland et al., 2016). The next step would be to identify mechanisms through which participation influences substance use to either reduce risks associated with participation or support the protective effects.