Binge Drinking and Educational Participation among Youth in the COMPASS Host Study (Year 5: 2016/2017): School Connectedness and Flourishing as Compensatory Factors

The current study investigated resilience factors influencing the associations between binge drinking and measures of educational participation among Canadian youth. Self reported data were collected during the 2016/2017 school year from 5238 students in Grades 9 through 12 (2744 females, 2494 males) attending 14 secondary schools in Ontario and British Columbia as part of the COMPASS study. Logistic regression analyses were conducted to determine relationships between binge drinking, school connectedness and flourishing on measures of educational participation. Binge drinking was associated with increased likelihood of skipping classes, going to class without completing homework, lower Math and English scores, and having educational and/or training expectations and aspirations beyond high school only. Decreased flourishing was linked to increased likelihood of going to class with incomplete homework, lower Math and English scores, and decreased likelihood of aspiring and expecting to achieve education and/or training beyond high school only. Increased school connectedness was associated with decreased likelihood of skipping classes and going to class with incomplete homework, higher Math and English scores, and increased the likelihood of aspiring to and expecting to achieve education and/or training beyond high school only. Lower flourishing was additive in its effect on current binge drinking in negatively impacting class attendance and homework completion and academic performance, while higher school connectedness was compensatory in its effect on these outcomes. This study suggests that, for high school students who are susceptible to binge drinking, those who are more connected to school and have a higher sense of wellbeing can maintain active participation in school and achieve their educational goals.


Introduction
Reports from the Canadian Centre on Substance Abuse have shown negative impacts of alcohol consumption on the mental and physical development in youth (Butt, Beirness, Gliksman, Paradis, & Stockwell, 2011). Binge drinking in adolescents, defined as having five or more alcohol drinks on one occasion (Wechsler & Nelson, 2001) has been associated with lower academic performance in school and increased tendency for smoking and use of illicit substances (Miller, Naimi, Brewer, & Jones, 2007). Data from the Youth Risk Behaviour Survey indicated that while binge drinking rates were similar among girls and boys, rates rose with increasing age and grade levels (Miller et.al., 2007). Previous work has shown negative effects of binge drinking on educational achievement among youth in the COMPASS Host Study, specifically the decreased likelihood for higher Math and English scores (Patte, Qian, & Leatherdale, 2017a). Binge drinking during adolescence has been highly predictive of binge drinking into early adulthood (McCarty et.al., 2004), which presents a problematic life trajectory. Current social norms are such that approximately 64% of Canadian adults have reported binge drinking at least once in a year (Paradis, Demers, & Picard, 2010).
The goal of the current investigation is to gain knowledge on youth who binge drink, from a resilience perspective, with resilience defined as an adaptation to adversity brought about by a developmental process (Fergus & Zimmerman, 2005). Resilience is here defined with a greater emphasis on the assets and resources that an individual has at their disposal to facilitate developmental growth, and less as a static individual trait. Such an approach proposes that youth possess assets and can access resources which, either alone or in combination, can compensate for, or protect against the risks that promote binge drinking behaviour and negatively impact adolescent development and normative functioning (Fergus & Zimmerman, 2005).
The COMPASS Host Study offers a comprehensive resource for hierarchical study of factors influencing health behaviours in Canadian youth. Interest in the COMPASS Host Study has been motivated in large part by high prevalence of binge drinking in students in grades 9 through 12. Data from the first cycle (2012/13) of the COMPASS study indicated that approximately 26% of respondents identified as current binge drinkers (Herciu et.al., 2014).
Further investigations have indicated that binge drinking is predictive of lower educational achievement, preparation, and engagement in Canadian youth, which may interfere with their future life aspirations (Patte et.al., 2017a).
Previous studies have shown that school structure and school engagement are key factors influencing youth academic performance (Costa et.al., 1999;Crosnoe et.al., 2002). Higher social support from teachers and parents has also predicted higher levels of school compliance (Wang & Eccles, 2012). Higher levels of school bonding have been shown to be associated with higher levels of subjective wellbeing (Rose, Lindsey, Xiao, Finigan-Carr, & Joe, 2017). Work by Griffin et al. (2002) also indicated that a student's sense of wellbeing is likely linked to their academic performance. The current study seeks to determine whether factors of school connectedness and flourishing influence the associations between youth binge drinking and frequency of skipping classes, completing assigned homework, current/most recent Math and English scores, and future educational expectations and aspirations via protective or compensatory mechanisms; with flourishing as a measure of overall wellbeing. Protective factors are tested using interaction effects, while compensatory factors contribute additively to the specified outcomes after accounting for risks in the model (Fergus & Zimmerman, 2005). Findings can inform policies and preventive programming for maladaptive alcohol use and poor academic achievement in youth.

Study Description
The COMPASS Host Study is a prospective cohort study for collection of hierarchical data from a convenience sample of secondary schools and the students in grades 9 through 12 who attend these schools. Student-level assessment of current binge drinking and other health behaviours was conducted via use of the COMPASS Student Questionnaire (Leatherdale et.al., 2014). Data for the current investigation utilized a sub-sample of students who completed the Mental Health Module during the 2016/2017 academic year. Comprehensive details on the COMPASS Host Study, including sampling and data collection procedures, are available online (www.compass.uwaterloo.ca) and in print (Leatherdale et.al., 2014).

Participants
School boards and associated schools were selected according to whether they permitted active-information, passive-consent parental permission protocols (Leatherdale et.al., 2014). Student-level assessment of alcohol use and binge drinking were conducted via the COMPASS Student Questionnaire (Leatherdale et.al., 2014), with 46957 student respondents in Year 5 (Y5:2016/2017) from Ontario, Alberta, Quebec, British Columbia and Nunavut. The current study used student-level data from the Mental Health Pilot study, a subsample of the COMPASS Host Study in Year 5: 8344 students completed the Mental Health Pilot of the COMPASS Student Questionnaire (77% participation) at 9 schools in Ontario and 5 schools in British Columbia. Missing respondents primarily resulted due to absenteeism or free/study periods.

Measures
To assess the frequency of skipping classes, students were asked, "In the last 4 weeks, how many classes did you skip when you were not supposed to?" Responses were grouped as No skipping for "0 classes"; and Some skipping for "1 or 2 classes" or "3 to 5 classes" or "6 to 10 classes" or "11 to 20 classes" or "More than 20 classes". To assess the frequency of not completing assigned homework, students were asked, "How often do you go to class without your homework complete?" Responses were grouped as: Seldom for "Never" or "Seldom" and Often for "Often" or "Usually". To assess educational achievement students were asked, "In your current or most recent Math course, what is your approximate overall mark?" and "In your current or most recent English course, what is your approximate overall mark?" Responses for both variables were grouped as: less than 70%; 70-79%; 80-89%; and 90-100%.
To assess the level of post-secondary education and/or training that students expect to achieve, students were asked, "What is the highest level of education you think you will get?" Responses were grouped as: High School only for "Some high school or less" or "High school diploma or graduation equivalency"; College, Trade, Vocational for "College/trade/vocational certificate"; Bachelor degree for "University Bachelor's degree"; and Graduate degree for "University Master's/PhD/law school/medical school/teachers' college degree". To assess levels of post-secondary education and/or training students would like to achieve, students were asked, "What is the highest level of education you would like to get?" Responses were grouped as: High School only for "Some high school or less" or "High school diploma or graduation equivalency"; College, Trade, Vocational for "College/trade/vocational certificate"; Bachelor's degree for "University Bachelor's degree"; and Graduate degree for "University Master's/PhD/law school/medical school/teachers' college degree".
To assess binge drinking status, students were asked, "In the last 12 months, how often did you have 5 drinks of alcohol or more on one occasion?" Responses were grouped as: Nonbinger for "I have never done this" or "I did not have 5 or more drinks on one occasion in the last 12 months" or "Less than once a month"; and Current for "Once a month" or "2 to 3 times a month" or "Once a week" or "2 to 5 times a week". To assess current smoking status, students were asked, "On how many of the last 30 days did you smoke one or more cigarettes?" Responses were grouped as: Non-smoker for "None"; and Current smoker for "1 day" or "2 to 3 days" or "4 to 5 days" or "6 to 10 days" or "11 to 20 days" or "21 to 29 days" or "30 days (every day)".
To assess psychosocial wellbeing, students were questioned using the Flourishing Scale (Diener et.al., 2010). This scale assessed students' agreement with the following statements: "I lead a purposeful and meaningful life"; "My social relationships are supportive and rewarding"; "I am engaged and interested in my daily activities"; "I actively contribute to the happiness and well-being of others"; "I am competent and capable in the activities that are important to me"; "I am a good person and live a good life"; "I am optimistic about my future"; and "People respect me". This measure is reverse coded with higher scores indicating increased languishing and lower scores indicating increased flourishing; scores range between 8 and 40. Cronbach's α for this measure was 0.91. To assess the levels of school connectedness, a six-item derived measure was used. These items assessed students' agreement with the following statements, as previously reported (Elton-Marshall et.al., 2011): "I feel close to people at my school"; "I feel I am part of my school"; "I am happy to be at my school"; "I feel the teachers at my school treat me fairly"; "I feel safe in my school"; and "Getting good grades is important to me". Scores range between 6 and 24, with higher scores indicating higher levels of school connectedness. Cronbach's α for this measure was 0.83.

Statistical Analyses
Descriptive statistics were performed to describe the distribution of the study variables. Two-sample t-tests and Chi-square tests were conducted to examine whether study variables differed between males and females. Marginal logistic regression models were used to examine whether flourishing and school connectedness moderate the impact of binge drinking status on school engagement, educational achievement, and educational ambitions in youth. Full (with interactions) and reduced (without interactions) models were fitted for each outcome, based on resilience models described elsewhere (Fergus & Zimmerman, 2005). Models 1 and 2 are for binary outcomes for frequency of skipping classes and for frequency of not completing homework, respectively. Models 3 and 4 are for multinomial outcomes on current/most recent Math and English scores, respectively. Models 5 and 6 are for multinomial outcomes on the level of education a student would like to achieve and the level of education a student expects to achieve, respectively. Models were controlled for grade (9 through 12), sex (male or female), ethnicity (White, Non-White) and current smoking status (Non-current or Current smoker), and accounted for within-school clustering.
All analyses were conducted using the statistical software package SAS 9.4 (Cary, NC). Marginal logistic regression models were fitted by using the SAS PROC GEE procedure with a binomial distribution and a logit function for models with a binary outcome and a generalized logit function for models with a multinomial outcome. The models for binary outcomes used an exchangeable working correlation structure based on the results of initial analyses, while the models for multinomial outcomes used an independent working correlation. Empirical standard error estimates were used to calculate confidence intervals and test statistics. Statistical significance was set at P < 0.05.

Ethics
This study has been approved by the University of Waterloo's Office of Research Ethics (ORE # 17264) and participating school boards.

Demographics
As shown in Table 1, 40.6% (n = 1114) of females and 40.0% (n = 998) of males reported current binge drinking, while 10.7% (n = 293) of females and 12.1% (n = 301) of males identified as current smokers. Students reported moderate levels of flourishing with mean scores of 16.8 (5.9) for females and 15.7 (5.6) for males, based on the 2 nd tertile of the Flourishing Scale. Students also reported high levels of school connectedness with mean scores of 18.0 (3.2) for females and 18.3 (3.3) for males, based on the 3 rd tertile of this derived measure. Regarding school engagement, 43.0% (n = 1180) of females and 39.7% (n = 991) of males reported some skipping of classes, while 25.0% (n = 685) of females and 31.9% (n = 795) of males reported often going to class without completing their homework. For current/most recent Math score, 28.1% (n = 771) of females and 33.0% (n = 822) of males reported scores of less than 70%. Further, 22.2% (n = 610) of females and 19.4% (n = 483) of males reported scores between 90%-100%. For current or most recent English score, 11.9% (n = 327) of females and 25.3% (n = 632) of males reported scores of less than 70%. Further, 26.6% (n = 729) of females and 13.0% (n = 325) of males reported scores between 90%-100%. In terms of educational aspirations, 55.4% (n = 1520) of females and 35.7% (n = 890) of males aspired to a graduate degree. For levels of education students expected to achieve, 41.5% (n = 1139) of females and 26.5% (n = 661) of males reported a graduate degree.

Skipping Classes and Incomplete Homework
Results for skipping classes (Model 1) and not completing assigned homework (Model 2) are given in Table 2. Compared to students who identified as non-bingers, students who reported current binge drinking were more likely to skip classes [OR 3.39, 95% CI: 2.93-3.92]; and were more likely to go to class with incomplete homework often [OR 1.42, 95% CI: 1.20-1.68]. For every 1-unit decrease in flourishing there was an associated 5% increase in likelihood of going to class with incomplete homework [OR 1.05; 95% CI: 1.04-1.06]. For every 1-unit increase in school connectedness, there was an associated 10% decrease in likelihood of skipping classes [OR 0.90, 95% CI: 0.88-0.92], and an associated 9% decrease in likelihood of going to class with incomplete homework often [OR 0.91, 95% CI: 0.89-0.94]. There were no significant interactions between flourishing and binge drinking status, or between school connectedness and binge drinking status, on predicting likelihoods of skipping classes and not completing assigned homework.

Math and English Scores
As shown in Model 3 in Table 3, students who identified as current binge drinkers were less likely than those identifying as non-current binge drinkers to report a As shown in Model 4 in Table 3, students who identified as current binge drinkers were less likely than those identifying as non-current binge drinkers to report an English score of 90-100% versus a score of less than 70%; [OR 0.70, 95% CI: 0.60-0.82]. For every 1-unit decrease in flourishing, there was an associated 2% decrease in likelihood of reporting an English score of 70-79% versus less than 70%; [OR 0.98, 95% CI: 0.97-0.99]. Further, for every 1-unit increase in school connectedness, there was an associated 23%, 16% and 8% increase in likelihood of reporting an English score of 90-100% [OR 1.23, 95% CI: 1.18-1.29], 80-89% [OR 1.16, 95% CI: 1.12-1.20], or 70-79% [OR 1.08, 95% CI: 1.05-1.11] respectively versus a score of less than 70%. There were no significant interactions between flourishing and binge drinking, or between school connectedness and binge drinking, on predicting current/most recent English scores.

Higher Education Aspirations and Expectations
Results on educational aspirations (Model 5) and educational expectations (Model 6) are given in Table 4. Compared to students who identified as non-bingers, students who reported current binge drinking were more likely to aspire to graduate school [OR 1.43 , and a 10% increase in likelihood of aspiring to a Bachelor degree [OR 1.10; 95% CI: 1.04-1.17] over high school only. For every 1-unit increase in school connectedness there was a 17% increase in likelihood of expecting to achieve a graduate degree [OR 1.17, 95% CI: 1.11-1.24], and a 13% increase in likelihood of expecting to achieve a Bachelor's degree [OR 1.13, 95% CI: 1.07-1.19] over high school only. There were no significant interactions between flourishing and binge drinking, or between school connectedness and binge drinking, on predicting the level of education a student aspired to achieve, or on the level of education a student expected to achieve.
Compared to Grade 9 students, students in higher grades were more likely to aspire to a Bachelor's degree and postsecondary training (P < 0.05 for all). Students in higher grades were also more likely to expect to achieve a graduate degree, a Bachelor's degree, or postsecondary training when compared to Grade 9 students (P < 0.05 for all).

Discussion
The current study provides insight into some resilience factors influencing associations between binge drinking and measures of educational participation among youth.
Current binge drinking was associated with increased likelihood to skip classes and go to class without completing homework. Students who smoked tobacco products were also more likely to skip classes and go to class without completing their homework and were less likely to aspire to and expect to achieve postsecondary education and training over high school only. It is probable that concomitant smoking is additive alongside binge drinking to negatively impact school engagement and may further impact educational ambitions. Students with lower levels of flourishing (i.e. higher levels of languishing) were more likely to go to class without completing their homework, and less likely to aspire to or expect to achieve any level of postsecondary education or training beyond high school. Students with higher levels of school connectedness had lower likelihoods of skipping classes and going to class without completing homework. Further, increased school connectedness was associated with an increased likelihood to aspire to and expect to achieve postsecondary education over high school only. These factors contribute to a significant risk-resilience profile which may impact short-term and long-term educational achievement among youth (Hill & Wang, 2015;Wang & Eccles, 2012).
Decreasing flourishing in students was associated with lower self-reported Math and English scores. Being deprived of this asset contributed additively to current binge drinking in negatively impacting affecting educational outcomes, aligning with previous work (Griffin et.al., 2002). Moreover, increasing school connectedness in the sample of students was associated with higher self-reported Math and English scores for all levels. The influence of school connectedness is striking as it indicates that students may leverage this resource to better their academic performance. Previous work has shown positive associations between social support from teachers and higher levels of school compliance (Wang & Eccles, 2012), along with positive associations between school bonding and higher levels of subjective wellbeing (Rose et.al., 2017). Given the moderate levels of flourishing and school connectedness in the current sample, it is anticipated that any improvement in these factors may effectively compensate for the effects of binge drinking behaviour on educational outcomes.
The current study shows that youth who reported current binge drinking were more likely to aspire to and expect to achieve postsecondary education and training over high school only. Previous work by Patte and colleagues (2017) has also shown that students who engaged in rare/sporadic to monthly binge were more likely to aspire to achieve a Bachelor's degree or postsecondary training over high school only (Patte et.al., 2017a). Additional study on binge drinking onset in youth has shown that, compared to students who started binge drinking in Grade 10 or earlier, students who started binge drinking in Grades 11 and 12 were more likely to aspire to a graduate degree and to expect to achieve to postsecondary training, but less likely to expect to achieve a Bachelor's degree (Patte, Qian, & Leatherdale, 2017b). It is important to note that, unlike these previous studies (Patte et al, 2017a;Patte et.al., 2017b), the current study accounts for school connectedness and flourishing levels among respondents, making the consistency of these results intriguing.
The present findings of lower school participation coupled with higher educational aspirations and expectations among students who binge drink are consistent with conceptual work on the "social clock" for contested behaviours in youth. The social clock has been defined as the societal expectations on the appropriate ages for youth to experience major life events, with an individual being either "early", "late" or "on-time" (Neugarten, Moore, & Lowe, 1965). Indeed, educational aspirations and expectations beyond high school were more likely for students in later grades. Research on parenting practices show them as indirectly influencing enrollment in post-secondary education and training (Hill & Wang, 2015). Using a sample of Ontario adults, Paglia and Room (1998) have also shown that parents who were current users of alcohol had higher tolerance for youth alcohol use compared to those who abstained. Social norms may drive parents and guardians to emphasize completing high school and furthering education and training, while assuming a more liberal role in terms of alcohol use. It is possible that youth who experiment with binge drinking during high school may exhibit this behaviour as a rite of passage that is socially acceptable within Western society (Ievers-Landis, Greenley, Burant, & Borawski, 2006). Such behaviour may be viewed as a life-change event marking their ability to take on more personal responsibility (Paglia & Room, 1998), including shaping their future life goals. Further investigations that account for parents' perspective of youth alcohol use and educational participation beyond high school are warranted.
There were no interactions between flourishing and binge drinking status or school connectedness and binge drinking status on measures of educational participation. These results indicate that the asset of flourishing and the resource of school connectedness may influence the associations between binge drinking and youth educational outcomes via compensatory mechanisms rather than protective ones (Fergus & Zimmerman, 2005). Based on a resilience framework, increasing flourishing and school connectedness may buffer the associations between binge drinking and indicators of educational participation (Fergus & Zimmerman, 2005). Flourishing, school connectedness and binge drinking are likely associated with a latent construct that directly influences students' school engagement and their subsequent educational aspirations and expectations. Indeed, higher levels of school bonding in youth have been associated with higher levels of subjective wellbeing (Rose, Lindsey, Xiao, Finigan-Carr, & Joe, 2017). In turn, increased levels of social support from parents and teachers have been associated with higher levels of school compliance and increased subjective valuing of learning (Wang & Eccles, 2012). As students navigate the latter years of adolescence, they may leverage the asset of flourishing and the resource of school connectedness to improve their levels of school engagement and maintain greater educational ambitions. Future work may determine whether the factors of flourishing and school connectedness are protective over varying degrees of binge drinking frequency.

Strengths and Limitations
The current study utilized a large sample of students from schools based in two Canadian provinces (Ontario and British Columbia), providing in-depth evaluation of binge drinking and other health behaviours in youth. However, the cross-sectional design means that causality and sequence of events cannot be inferred. Data from the COMPASS Host Study's student questionnaire are self-reported, and though bias may have been introduced, this method provides an emic representation of students' perceived flourishing and school connectedness. The data collection procedures further limit bias by use of an active-information, passive-consent permission approach which has been found to maintain perceived anonymity, improve response rates, and minimize underreporting of antisocial behaviour (Thompson-Haile, Bredin, & Leatherdale, 2013).

Implications
This study provides insight on youth who engage in binge drinking from a resilience perspective. Based on the current results, increasing flourishing and school connectedness in students may be useful targets for substance-use intervention programs. Efforts to provide developmental supports which maintain student confidentiality, parental autonomy, and build on school-based resources would be most promising (Hill & Wang, 2015). Increased flourishing may be achieved by enabling students to form a positive identity, increase their self-awareness, and maintain a sense of belonging within the school environment through peer-peer and peermentor relationships, along with improved teacher support (Compton, Jackson, & Dimmock, 2016;Fergus & Zimmerman, 2005;Mcguire, 1961). Increased school connectedness may be achieved via clear messaging of behavioural and academic standards, and an overall personalization of the school experience for students (Catalano, Berglund, Ryan, Lonczak, & Hawkins, 2004;Rose et.al., 2017;Wang & Eccles, 2012).

Conclusion
The current study found that youth who binge drink had poorer educational participation in terms of skipping classes, not completing assigned homework, and lower Math and English scores, while maintaining educational expectations and aspirations beyond high school. As an asset, decreased flourishing may have an additive effect on how binge drinking negatively impacts educational participation. As a resource, increased school connectedness may compensate for the negative impact of binge drinking on educational participation. School-based efforts to improve flourishing and school connectedness may enable youth to improve their school engagement, successfully meet their educational ambitions, and develop a resilient trajectory over the long-term.