JOURNAL TRANSCRIPT

---

RESEARCH AND PRACTICE

Longitudinal Effects of School Drug Policies on Student Marijuana Use in Washington State and Victoria, Australia | Tracy J. Evans-Whipp, PhD, Stephanie M. Plenty, PhD, Richard F. Catalano, PhD, Todd I. Herrenkohl, PhD, and John W. Toumbourou, PhD

Marijuana is the most widely used illicit drug worldwide,1,2 with an estimated 181 million (3.9%) of the world’s adults using it in 2011.3 Surveys in the United States and Australia have shown that marijuana use is particularly high among adolescents.4'5 Concern about mari­ juana use has increased in recent years as a result of improved understanding of the harmful health and psychological effects of frequent use, especially among adolescents and young adults.6,7 At the same time, many US states have passed marijuana laws making it legal for adults older than 21 years to possess small amounts of marijuana for medical pur­ poses. Two states—Colorado and W ashington— have legalized marijuana for recreational use by adults. Studies on the impact of marijuana legislation on marijuana use by US adoles­ cents have yielded mixed results, with some pointing to an increase in use and others to no change or to a decrease in marijuana use.8-13 M arijuana use is illegal in A ustralia.14 School-based prevention programs and pol­ icies have become the dominant mode of drug prevention for adolescents.15 School drug pol­ icies aim to reduce levels of adolescent sub­ stance use by restricting access to drugs and exposure to drug use during school hours. Studies measuring access to marijuana at the individual, school, and country levels have shown consistent associations between in­ creased access and higher rates of self-reported use by adolescents.16-18 An Australian study showed that high rates of school-level mari­ juana use (an indirect measure of exposure) are associated with increased risk of use by sec­ ondary students.19 In addition, students in Swiss schools with more incidents of marijuana intoxication (as reported by teachers) were more likely to report marijuana use, regardless of peer use,20,21 itself a salient risk factor.19,22 Even in the absence of direct exposure to others’ marijuana use, students may be influ­ enced by the general level of acceptability or

Objectives. We examined the longitudinal effect of schools' drug policies on student marijuana use. Methods. We used data from the International Youth Development Study, which surveyed state-representative samples of students from Victoria, Australia, and Washington State. In wave 1 (2002), students in grades 7 and 9 (n = 3264) and a school administrator from each participating school (n = 188) reported on school drug policies. In wave 2 (2003), students reported on their marijuana use. We assessed associations between student-reported and administrator-reported policy and student self-reported marijuana use 1 year later. Results. Likelihood of student marijuana use was higher in schools in which administrators reported using out-of-school suspension and students reported low policy enforcement. Student marijuana use was less likely where students reported receiving abstinence messages at school and students violating school policy were counseled about the dangers of marijuana use. Conclusions. Schools may reduce student marijuana use by delivering absti­ nence messages, enforcing nonuse policies, and adopting a remedial approach to policy violations rather than use of suspensions. (Am J Public Health. 2015;105: 994-1000. doi:10.2105/AJPH.2014.302421)

disapproval of marijuana use in the broader school environment.23,24 Thus, school drug policy may have a further potentially important function in addressing marijuana social norms in the school context. Although almost all secondary schools in the United States and Australia have illicit drug policies, school-to-school variation in policy content exists.25-27 Schools differ in how they develop, communicate, and enforce their policies as well as in policy intent (e.g., goals of abstinence vs harm minimization). In addition, schools vary with respect to their responses to incidents of student drug use, which range from highly punitive approaches such as expulsion and suspension to remedial responses such as coun­ seling.25-27 Despite calls from leading govern­ ment agencies for schools to implement evi­ dence-based, whole-school drug education policies and programs,28,29 empirical evidence of effective policy effects is relatively scarce. Studies examining the effectiveness of school drug policies in reducing student drug use have demonstrated mixed results, although there is some evidence of the importance of policy enforcement30

9 9 4 | Research and Practice | Peer Reviewed | Evans-Whipp et al.

The majority of studies have focused on the impact of tobacco policy on student smoking.31-42 Fewer studies have investigated the impact of policies on student alcohol43-46 and illicit drug use.47,48 W ith 1 exception,45 none of these studies has dem onstrated a longitudinal relationship betw een school policy and sub­ sequent drug use. Further research is re­ quired to understand how these policies affect student drug use. Particularly needed are studies addressing the predictive impact of various elem ents of school policy, includ­ ing punitive versus rem edial policies and responses, policy enforcement, and exposure to abstinence and harm minimization mes­ sages related to substance use in the school context. W e aimed to fill the existing knowledge gap by assessing the longitudinal impact of school illicit drug policies on student marijuana use. W e maximized variation in the measured policy components by using data from the International Youth Development Study (IYDS), an ongoing longitudinal cross-national study of schools and adolescents in Washing­ ton State and Victoria Australia which have

American Journal o f Public Health | May 2 0 1 5 , Vol 1 0 5 , No. 5

RESEARCH AND PR AC TICE

been shown to differ in their approach to school policy elements. Washington school policies have been more oriented toward total abstinence and more frequently enforced with harsh punishment (such as expulsion or calling law enforcement), whereas policies in Victoria schools have been more reflective of harm minimization principles.25 Previous studies investigating the validity of the IYDS school policy survey tools have shown that reports from school officials and students in the

as most knowledgeable of the school’s drug policies and procedures) from each participat­ ing school completed a school administrator

ranging from YES! (coded as 4) through yes (3) and no (2) to NO! (1). The mean response

mail survey (97.4% participation rate). In this study, we used data from participants in

W e measured perceived consequences of marijuana policy violation by asking students, “If a student was found using marijuana at school,

the grade 7 (middle) and grade 9 (oldest) cohorts, who completed a student survey in wave 1 (2002) and 1 year later in wave 2 (2003; n = 3850; 99% retention rate in both states) and from administrators at the schools they attended. Students were excluded if they did not have

United States are significantly different from those in Australia and accurately reflect their respective

corresponding school administrator survey data (n = 91 students from 5 school administrators

national policy approaches to youth alcohol and drug use.25'25 School official and student reports on IYDS school alcohol policy measures have

who did not complete the school survey) or if they changed schools between wave 1 and wave 2 (n=449). Honesty criteria resulted in the exclu­ sion of 46 students. The final sample consisted of

longitudinally predicted student alcohol use.45 In this study, we used IYDS school policy information collected from both school officials and students and self-reported student marijuana use 1 year later to address the following research questions: 1. Is student marijuana use predicted by the level of enforcement of school illicit drug policies? 2. Is student marijuana use predicted by different types of school responses to illicit drug use at school? 3. Is student marijuana use predicted by the degree to which school illicit drug policy is based on abstinence and harm minimiza­ tion principles?

M ETHO DS

The data used in this study were collected during the first and second years of the IYDS. Procedures for the IYDS sampling, school administrator survey, and student survey have been described in detail elsewhere.25,49,50 Briefly, a 2-stage cluster sampling approach

3264 students from 188 schools. Because of the 2 age cohorts in the sample, participants in wave 2 were aged approximately 14 or 16 years (Washington: m ean= 15.0 years; SD = 1.1; range= 13.0-18.2; Victoria: m ean= 14.9 years; SD = 1.0; range =12.9-17.2). M easu res

The self-reported measure of student mari­ juana use was adapted from the Monitoring the Future survey.51 The school policy measures in the school administrator and student surveys were developed by the IYDS to measure school drug policy environments in Washington and Victoria Many school administrator survey items were derived from existing measures of school policies in the United States and some items, as well as the student survey items, were developed by IYDS staff to measure additional dimensions of interest25,26 The cross-sectional and prospective validity of the school policy measures has been documented previously. 25-26-3445

schools approached) Washington schools and 154 (65.5%) Victoria schools agreed to par­ ticipate. In the second stage, 2885 (74.8%) of

Student-level outcome and school policy variables. The measure of current marijuana use at wave 2 asked students, “In the past 30 days on how many occasions (if any) have you used marijuana (pot weed, grass)?” A binary indicator of marijuana use was formed (none vs > 1 times). W e used responses regarding the most frequently used drugs, alcohol and cigarettes, to measure low policy enforcement. Students

Washington State parents and students and 2 8 8 4 (73.5%) of Victoria parents and students consented to participate. Students completed surveys during class time. The school principal (or a staff member nominated by the principal

indicated their agreement with the following 2 items, “Many students smoke on school grounds without getting caught” and “Many students drink alcohol on school grounds without getting caught,” on a 4-point scale from

was used to recruit state-representative sam­ ples of school students from 3 grade cohorts (grades 5, 7, and 9) in Washington State and Victoria. In the first stage, 153 (70.5% of

May 2 0 1 5 , Vol 10 5, No. 5 | American Journal o f Public Health

formed a measure of low policy enforcement.

which of the following would most likely hap­ pen? (circle all that apply).” Responses were as follows: (1) he or she would be talked to by a teacher about the dangers of using marijuana, (2) he or she would be suspended, (3) he or she would be expelled, and (4) the police would be called. Each response was coded as 1 if circled and 0 if not circled. W e measured abstinence and harm minimiza­ tion policies by asking students whether they agreed with the following 2 statements regarding their school: “We are taught to say no to alcohol” (abstinence) and “W e are taught how to use alcohol safely” (harm minimization). Response options were YES!(A), yes (3), no (2) and NO!(\). W e calculated a measure of honesty based on student reports of being “not honest at all” when completing the survey, using a fic­ tional drug, or using illicit drugs more than 120 times in the past 3 0 days.52 School-level school policy variables. School administrators were asked, “In your opinion, how strictly are the substance use policies being enforced at your school?” Responses options ranged from very strictly (1) to not at all strictly (4). W e determined penalties for illicit drug use by asking school administrators to indicate whether their schools had illicit drug policies. The 97.3% of administrators who responded yes were then asked to indicate the likelihood of issuing specific consequences when “students are caught using, possessing or being under the influence of illicit drugs on school grounds or at school events.” Responses were as follows: expelled from school; referred to legal authori­ ties (police); suspended from school; referred to a school counselor or nurse; recommended to participate in an assistance, education, or cessa­ tion program; or required to participate in an assistance, education, or cessation program. Re­ sponses were dichotomized as always or almost always (1) or sometimes, rarely, or never (0). W e measured abstinence and harm mini­ mization policy by asking administrators whether they agreed with the following 2 statements: “School policies emphasize total abstinence from drug use” (abstinence) and

Evans-Whipp et ai. \ Peer Reviewed | Research and Practice | 9 9 5

RESEARCH AND PR AC TICE

t test for continuous measures and the y 2 test for categorical variables. Then, we tested the bivariate

TABLE 1 -D e s c r ip tiv e S ta tis tic s for O utcom e, Predictor, and Control V ariab les by S ta te : In te rn a tio n a l Youth D evelop m ent Study; W ashington S ta te and V ic to ria , A ustralia;

(unadjusted) association between each school

2 0 0 2 and 2 0 0 3

administrator-reported policy variable and student marijuana use in a random effects logistic regres­

Victoria (n = 17 2 2 Students Variable

Washington (n - 15 42 Students

sion using maximum likelihood estimation. We

and 9 8 Schools), % (No.),

and 90 Schools), % (No.),

Mean ± S D , or Median (IQR)

Mean ± S D , or Median (IQR)

regressions to identify associations between each control variable and student marijuana use. We used partially adjusted models in which each

Control variables Family SES, median (interqu artile range)

1.9 (1 .5 -2 .4 )

Older cohort

49 .9 (860)

Wave 1 current marijuana use, past 30 d

3.6 (62)

2 .0 (1 .9 -2 .5 )* * 48 .9 (754)

policy variable was entered simultaneously with

7.9 (1 2 1 )* *

the control variables. These analyses modeled the

Outcom e variable (wave 2): Current m arijuana use, past 30 d

7.2 (123)

random effects at the school (cluster) level. We

11.7 (1 8 0 )* *

performed random effects logistic regression analysis using the xtlogit command of Stata

School ad m inistrator-re po rted variables (wave 1) Punitive penalties fo r illic it drug use 2 8 .4 (25)

Call police

also performed a series of random effects logistic

The bivariate associations between studentreported policy components, as well as each

6 9 .8 (6 0 )* *

Expulsion

10.8 (9)

15.9 (13)

Out-of-school suspension

62.9 (56)

89.7 (7 8 )* *

control variable and marijuana use, were tested in bivariate logistic regressions. W e then per­

Recommend program

3 3 .3 (26)

5 2 .4 (4 3 )*

formed a series of partially adjusted logistic regressions to estimate the predictive associa­

Remedial penalties fo r illic it drug use

Require program

40.2 (33)

6 7 .8 (5 9 )* *

Refer to nurse o r counselor

78.9 (71)

77.7 (66)

Low policy enforcem ent3

1.5 ± 0 . 6 * *

1.1 ± 0 . 3

Abstinence policy11

3.9 ± 1 . 3

4.9 ± 0 . 4 * *

3 .3 ± 1 . 3 * *

2.1 ± 1 . 3

the clustering of students within classes using the svy command in Stata. We evaluated interactions between schooland student-reported policy components and

Harm m inim ization policy11

tion between each policy component and marijuana use while accounting for control variables. All logistic regressions accounted for

S tudent-reported variables (wave 1) Talked to by teacher

38.1 (6 5 6 )* *

2 9 .6 (457)

Suspension

3 8 .8 (668)

3 9 .7 (612)

Expulsion

57.7 (9 9 4 )* *

5 1 .2 (7 90)

Police called

4 4 .0 (758)

53.7 (8 2 8 )* *

Low policy enforcem ent3

2.2 ± 0 . 7 * *

1.9 ± 0 . 8

Abstinence policy11

3.0 ± 0 . 9

3.4 ± 0 . 7 * *

Harm m inim ization policy11

2.7 ± 1 . 0 * *

2.4 ± 1 . 1

the variables state, cohort gender, and wave 1 marijuana use to determine any differential effects. Of the 68 comparisons, only 3 were statistically significant (at P< .05). We therefore

Note. IQR = interquartile range; SES = socioeconom ic status. Statistics are based on nonm issing values. The range o f sam ple sizes fo r student variables was 1 5 2 6 -1 7 2 2 fo r Victoria and 1 4 7 1 -1 5 4 2 fo r W ashington; the range of sam ple sizes fo r school variables was 7 8 -9 8 fo r Victoria and 8 2 -9 0 fo r Washington. “ On a scale ranging from 1 to 4, on which 1 - very strictly, 2 - moderately strictly, 3 = not very strictly, and 4 = not a t all

strictly. b0n a scale ranging from 1 to 5, on which 1 = not at all, 2 - a little, 3 = some, 4 = quite a bit, and 5 * a lot. c0 n a scale ranging from 1 to 4, on which 1 - NO!, 2 -n o , 3 -ye s, and 4 - YES!

(harm minimization). Response options ranged from not at all (1) to a lot (5). Control variables. W e controlled for several variables in examining the impact of aspects of school policy on marijuana use: state (Victoria vs Washington), gender, cohort (oldest vs middle), family socioeconomic status (SES),45 and previous-year marijuana use. The binary

of the school administrator-reported penalties for illicit drug use items, fewer than 3°/o of cases were missing data for each variable; we therefore excluded missing data from the analyses. Missing

Statistical Analysis

data for the 6 school-reported penalties for illicit drug use items ranged from 3.4% to 12.1%. Student-level cases missing for these items were more likely to be from Victoria and in the older cohort although they did not differ with respect to current marijuana use at wave 2. For 2 of the items (recommend or require program), cases with missing data had a lower SES. These differences may have biased the results slightly.

W e performed all analyses using Stata ver­ sion 12.1 (StataCorp LP, College Station, TX).

RESULTS

First, we summarized the school policy, mari­ juana use, and control variables for each state separately and compared differences using the

Table 1 presents the sample characteristics for Washington State and Victoria On both survey

* P < . 05 ; * * P < . 001 in state comparison.

“School policies are based on the assumption that most youth will experiment with drugs”

present the analyses for the nonstratified sample (with subgroup analyses performed for the 3 significant interaction cases). With the exception

measure of previous year (past 30 days mari­ juana use in wave 1) was identical to the marijuana use measure in wave 2.

9 9 6 | Research and Practice | Peer Reviewed | Evans-Whipp et at

American Journal o f Public Health | May 2 0 1 5 , Vol 10 5, No. 5

RESEARCH A ND PR AC TICE

TABLE 2 -U n a d ju s te d and Adjusted Associations for School-Level Policy Variables as Predictors of Student Current M arijuana Use 1 Year Later: International Youth Development Study; Washington State and Victoria, Australia; 2 0 0 2 and 2 0 0 3

students. Interactions for state (Victoria vs

Unadjusted Variable

No.

OR (95% Cl)

Tests of interactions provided no strong evidence of differential effects betw een the 2 grade cohorts and betw een male and female

Adjusted® No.

OR (95% Cl)

W ashington) indicated differential effects on m arijuana use for the low policy enforcem ent variable only. Recalculation of the odds ratios (ORs) for Victoria and W ashington

M ultilevel model State (Washington)

32 43

1 .7 6 * (1 .30, 2.39)

Cohort (older)

32 43

2 .6 3 * (1 .97, 3.51)

Family SES

31 5 0

Wave 1 current m arijuana useb

32 1 6

State separately (including the control vari­ ables state, gender, cohort, and family SES) showed th at low policy enforcem ent p re­

0 .7 3 * (0 .55, 0.98) 1 6 .5 9 * (1 1.72 , 23 .4 8 )

dicted higher student m arijuana use in the

School administrator-reported policy variables

Victoria sample only (O R = 1.50; 95 % con­ fidence interval [Cl] = 1.08, 2.08). Interac­

Punitive penalties Call police

3032

1.32 (0 .96, 1.82)

29 26

1.0 (0 .72, 1.38)

Expulsion

28 84

1.31 (0 .83, 2.06)

27 82

0 .9 8 (0 .65, 1.49)

Out-of-school suspension

30 58

2 .3 6 * (1 .53, 3.66)

29 52

1 .6 2 * (1 .06, 2.49)

Remedial penalties Recommend program

2781

1.32 (0 .93, 1.87)

26 8 3

1.03 (0 .75, 1.41)

Require program

29 49

1 .8 6 * (1 .34, 2.5 8)

2847

1.29 (0 .93, 1.77)

Refer to nurse or counselor

30 2 8

1.13 (0 .76, 1.69)

29 22

1.14 (0 .79, 1.63)

Low policy enforcem ent

31 7 6

1.05 (0 .77, 1.42)

3061

1 .3 5 * (1 .01, 1.82)

Abstinence policy

32 02

1 .2 6 * (1 .08, 1.47)

30 83

1.14 (0.97, 1.34)

Harm m inim ization policy

32 22

0 .9 2 (0 .82, 1.03)

31 05

0 .9 7 (0 .86, 1.09)

N ote. Cl - confidence interval; OR - odds ratio; SES - socioeconom ic status. “ Partially adjusted models controlled for the effects of state, cohort (grade), fam ily SES, and wave 1 current m arijuana use. bCurrent m arijuana use was defined as > 1 tim e in the past 30 days. * P < .05.

occasions, the prevalence of current marijuana use was significantly higher among Washington students than among Victoria students. School administrator reports of illicit drug policy revealed that Washington schools, compared with Victoria schools, were more likely to call police, use out-of-school suspen­ sions, and recommend or require students to attend programs in response to illicit drug incidents. Victoria schools, on average, reported lower policy enforcement than W ashington schools. None of the W ashington schools rated their policy enforcement in the not-very-strictly or not-at-all-strictly cate­ gories. Victoria schools, on average, reported higher levels of harm minimization policy orientation and lower levels of abstinence policy than W ashington schools. Students reported that the most common responses to breaches of illicit drug policy were expulsion for the Victoria students and calling the police for the Washington students. About 40% of students in both states dted suspension

as an option. Being counseled by a teacher about the dangers of using marijuana was another common option reported by the Victoria students. Victoria students were more likely to report low enforcement of school drug and alcohol policy. Similar to school administrator responses to policy orientation, significantly more Washington students reported an abstinence approach, and significantly more Victoria students reported a harm minimization approach. Results of the random effects regression models used to investigate the predictive associations between school administratorreported policy variables and student mari­ juana use 1 year later are presented in Table 2. W e found no statistically significant effects of several aspects of school policy on student marijuana use: calling the police, expulsion, recommending a program, or referring to a nurse or counselor. Use of out-of-school suspensions and low policy enforcement each predicted increased odds of student marijuana use in partially adjusted models.

May 2 0 1 5 , Vol 10 5, No. 5 | American Journal o f Public Health

tions for wave 1 m arijuana use status showed differential effects on m arijuana use only for the out-of-school suspension variable. Recalculation of the odds ratios for time 1 m arijuana user and nonuser groups sepa­ rately (including the control variables state, gender, cohort, and family SES) showed that school reports of using out-of-school sus­ pension predicted student m arijuana use only in the time 1 marijuana-using sample (OR = 4.36; 95 % C I= 1 .3 9 -1 3 .6 4 ). Table 3 presents the predictive associations between student-reported policy components and current marijuana use. In the unadjusted models, the odds of student marijuana use were reduced when students reported that being talked to by a teacher, expulsion, and calling the police were likely responses to illicit policy violations and for abstinence policy orientation. W ith the exception of expulsion, these signifi­ cant effects were retained in the partially adjusted models. Student reports of low policy enforcement predicted increased marijuana use in the unadjusted and partially adjusted models. Tests of interactions provided no strong evidence of differential effects betw een the 2 grade cohorts, male and female students, or Time 1 m arijuana users and nonusers. Ex­ amination of the interactions for state showed differential effects only for the low enforcem ent variable. Recalculation of the ORs for Victoria and W ashington separately (including the control variables tim e 1 m ar­ ijuana use, gender, cohort, and family SES) showed that low enforcem ent predicted increased student m arijuana use only in the Victoria sample (O R = 1.94; 95% Cl = 1 .4 6 , 2.58).

Evans-Whipp et at. | Peer Reviewed | Research and Practice | 99 7

RESEARCH AND PR AC TICE

TABLE 3 -U n a d ju s te d and A d justed A ssociations for Student-Level Policy V ariables as P redictors o f S tu d e n t C u rren t M a riju a n a Use 1 Year L ater: In te rn a tio n a l Youth Developm ent Study; W ashington S ta te and V ic to ria , A u stralia ; 2 0 0 2 and 2 0 0 3 Unadjusted Variable

No.

OR (95% Cl)

Adjusted3 No.

OR (95% Cl)

0 .6 1 * (0 .45, 0.83)

Logistic regression model State (Washington)

32 43

1 .7 1 * (1 .27, 2 .2 9)

Cohort (older)

32 43

2 .5 8 * (1 .90, 3 .5 0)

Family SES

31 50

0 .8 0 (0 .61, 1.04)

Wave 1 current m arijuana useb

32 16

1 5 .6 7 * (1 1.39 , 21 .5 4 )

S tu dent-reported policy variables

Talked to by teacher

32 43

0 .5 2 * (0 .39, 0.69)

31 2 4

Suspended

32 43

1.24 (0 .97, 1.58)

31 24

1.12 (0.85, 1.48)

Expelled

32 4 3

0 .7 4 * (0 .57, 0.96)

31 24

0 .8 8 (0 .65, 1.18)

Police called

32 43

0 .7 3 * (0 .55, 0.97)

31 24

0 .7 4 * (0 .55, 1.00)

lo w policy enforcem ent

32 11

1 .7 8 * (1 .52, 2.08)

30 92

1 .5 0 * (1 .2 2 , 1.82)

Abstinence policy

32 1 8

0 .6 8 * (0 .59, 0 .7 7)

30 99

0 .6 8 * (0 .59, 0.79)

Harm m inim ization policy

32 01

0 .9 0 (0 .80, 1.02)

30 8 5

0 .9 4 (0 .82, 1.09)

N ote. Cl - confidence interval; OR - odds ratio; SES - socioeconom ic status. aPartially adjusted models controlled fo r the effects o f state, cohort (grade), fam ily SES, and wave 1 current m arijuana use. 6Current m arijuana use was defined as > 1 tim e in the past 30 days. * P < .05.

DISCUSSION This study is one of the first to analyze the longitudinal effects of school illicit drug policy on student marijuana use. Both student and school administrator reports of school policy were investigated and found to be predictive of student marijuana use 1 year later. The first research question was related to policy enforcement. Enforcement has been identified as a key factor in studies of school tobacco31,35,53-56 and alcohol45 policy, and our findings indicate that it is similarly impor­ tant as a predictor of student marijuana use. Both school administrator and student reports of low policy enforcement predicted an in­ crease in the likelihood of later marijuana use. The second research question concerned the differential impact of school responses to breaches of illicit drug policy. Of particular note is the finding that students who attended schools that reported always or almost always using out-of-school suspensions for illicit drug policy violations were 1.6 times as likely to be marijuana users 1 year later. Accumulating evidence has shown that suspensions are re­ lated to unintended negative outcomes for the

suspended student, such as disengagement from school, delinquency or antisocial behav­ ior, smoking, and alcohol and drug use,52,57,58 and concerns have been raised as to the value of such practices 59 Our findings also reveal that school use of suspensions is associated with increased risk of marijuana use for the entire student body, not just for those who are suspended. However, student reports of likely suspen­ sions for marijuana policy violations, although indicative of elevated risk of marijuana use, were not statistically predictive, suggesting the elevated risk of marijuana use shown in the school-report model may be attributable to other co-occurring school factors. We tested the reverse causality hypothesis, in which schools with greater numbers of marijuana-using students are more likely to use suspensions, in additional analyses controlling for the total number of illicit policy violations in the school in the past year. This alternative hypothesis was not supported because we observed no significant attenuation in the association between school use of suspensions and student marijuana use (data available on request).

9 9 8 | Research and Practice | Peer Reviewed | Evans-Whipp et al.

Student-reported teacher counseling for policy violators predicted an almost 50% re­ duction in the likelihood of later marijuana use. Some previous studies have found protective effects of student counseling on the risk of student harmful drinking45 and smoking.60 However, school administrators’ reports of re­ ferral to a nurse or counselor were not related to student marijuana use. Whether this was the result of differences in the wording of the measure between the student and school ad­ ministrator surveys (referral to a teacher vs a nurse or counselor) or whether students’ and school administrators’ reports are capturing different dimensions of school policy and en­ forcement is not clear. Similar percentages of schools in Victoria and Washington reported using counseling responses, whereas Victoria students were more likely than Washington students to report teacher counseling. It is also interesting to note that the proportion of schools reporting referring student offenders to a nurse or counselor was about double that of students reporting a teacher counseling re­ sponse, which might suggest that schools are overreporting their use of counseling remedial approaches. Further longitudinal research on the impacts of various remedial approaches to drug policy violations is warranted. The reduced likelihood of marijuana use among students reporting punitive penalties, such as calling the police (adjusted OR [AOR] = 0.74; 95% Cl = 0.55, 1.00) and ex­ pulsion (AOR = 0.88; 95% Cl = 0.65, 1.18) might be indicative of such policies acting as a deterrent. However, we did not specifically measure marijuana use on school grounds, where a deterrent effect would most likely be observed. Punitive penalties might also help schools shape student norms by sending out a strong negative message about illicit drug use. This concept is supported by the finding that student reports of strong school abstinence messages predicted lower marijuana use. The final research question focused on the relative impact of abstinence-based and harm minimization-based policies on stu­ dent marijuana use. There is some evidence that student perceptions of abstinence ap­ proaches are protective against marijuana use (AOR = 0.68; 95% Cl = 0.59, 0.79), al­ though school reports of abstinence policies are not (AOR= 1.14; 95% Cl = 0.97, 1.34). Harm

American Journal o f Public Health | May 2 0 1 5 , Vol 10 5, No. 5

RESEARCH A ND PR AC TICE

minimization did not have an impact on the likelihood of student marijuana use in the

students’ reports might be subject to nonpolicy factors such as stories they have heard, and school

school-reported or student-reported models. However, harm minimization approaches

reports might be subject to response desirability

might be expected to exert maximal effects on harmful patterns of marijuana use rather than any use in the past 30 days. This was observed

attempt to control for previous-year marijuana use adds to the rigor of the tests conducted.

in a previous study of alcohol use in which

Implications

harm minimization policies were not associated with the likelihood of any drinking in the past 30 days but reduced the likelihood of student

dicating that schools should take measures to increase the enforcement of a no-use policy for

binge drinking and alcohol-related harms.45

substance use on school grounds. This might be

Further research on the impacts of school harm minimization policies on marijuana use pat­

achieved through intentional efforts to com­ municate in schools what the policies are for substance use and rule violations and by in­ creasing monitoring of substance-using behav­

terns and behaviors would be beneficial. Limitations This study has a number of limitations. First, the study was observational, not experimental, and so causal effects cannot be firmly estab­ lished. Second, we did not include a measure of self-reported marijuana use on school grounds, which is where the strongest deterrent effects of policy might be expected. Third, the studentand school administrator-reported policy measures require further validation and opti­ mization. There were some differences in wording between the student and administrator items, rendering direct comparisons problem­ atic. In some cases, the policy items specified not marijuana use but rather illicit drug use more broadly. Student assessment of strict policy enforcement, abstinence, and harm minimiza­ tion were based on responses to tobacco and alcohol policy items. We chose these items to provide more variation because use of these substances is legal at older ages, whereas use of illicit drugs is never legal. However, further improvement in specificity of the items in future research would be beneficial. The use of selfreport data may give rise to response bias and inaccuracies. This study also has major strengths. It drew on data from large representative samples of sec­ ondary students in 2 states that differ in their policies regarding substance use, thereby in­ creasing the variation in the policy variables. Survey procedures and instruments were matched between the 2 states, and attrition was extremely low.50 We used reports of school policy from both school administrators and stu­ dents to overcome some of the limitations asso­ ciated with using just 1 data source. For example,

bias. Finally, the use of longitudinal data and the

Our findings confirm previous research in­

iors on school grounds. In addition, delivery of strong abstinence messages relating to illicit drugs through policy and curriculum is impor­ tant and might be reinforced by the use of some punitive penalties, such as notifying the police. However, our finding related to the negative impact of school suspensions is of concern and worthy of further research. Rather than rely only on punitive responses, schools may be advised to provide education and counseling to students. ■

About the Authors Tracy J. Evans-Whipp, Stephanie M. Plenty, and John W. Toumbourou are with the Centre fo r Adolescent Health, Murdoch Children’s Research Institute, Parkville, Victoria, Australia. Tracy J. Evans-Whipp and Stephanie M. Plenty are also with the University o f Melbourne Department o f Paediatrics, Royal Children’s Hospital, Parkville. John W. Toumbourou is also with the Centre fo r Mental Health and Wellbeing Research and School o f Psychology, Deakin University, Geelong, Victoria. Richard F. Catalano and Todd I. Herrenkohl are with the Social Development Research Group, School o f Social Work, University o f Washington, Seattle. Correspondence should be sent to Tracy J. Evans-Whipp, Centrefo r Adolescent Health, Murdoch Children’s Research Institute, Parkville, Victoria 3052, Australia (e-mail: tracy. [email protected]). Reprints can be ordered at http://www.ajph.org by clicking the ‘Reprints’’ link. This article was accepted October 26, 2014.

Contributors T.J. Evans-Whipp conceptualized the analytic design, conducted the analysis, and prepared the original draft of the article. S. M. Plenty assisted with data analysis and participated in the preparation of the original article and subsequent revisions. R. F. Catalano designed the overall International Youth Development Study (IYDS) and assisted with the analytic design, interpretation of results, and revision of the article. T. I. Herrenkohl assisted with interpretation of results and revision of the article. J. W. Toum bourou designed the overall IYDS and assisted with interpretation of results and revision of the article.

May 2 0 1 5 , Vol 10 5, No. 5 | American Journal o f Public Health

Acknowledgments This work was supported by the National Institute on Drug Abuse (R01-DA012140-05) and the Australian National Health and Medical Research Council (Project No. 491241). It was also supported by the Victorian Govern­ ment’s Operational Infrastructure Support Program. W e thank the participants and project staff of the IYDS. The study is led and managed in Australia by the Centre for Adolescent Health and in W ashington State by the Social Development Research Group, University of Washington. Further information is available at the IYDS W eb site (http://www.iyds.org). W e are also grateful for helpful discussions with Eric C. Brown, PhD, at the Department of Public Health Sciences, University of Miami. Note. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.

Human Participant Protection All study procedures were approved by the institutional review board at the University of W ashington and the Ethics in Human Research Office at the Royal Children’s Hospital in Victoria and relevant education authorities in each state.

References 1. Degenhardt L, Bucello C, Calabria B, et al. Review: w hat data are available on the extent of illicit drug use and dependence globally? Results of four systematic reviews. Drug Alcohol Depend. 2 011; 1 1 7 (2 -3 ):8 5 - 101. 2. Hall W, D egenhardt L. Prevalence and correlates of cannabis use in developed and developing countries. Curr Opin Psychiatry. 2007;20(4 ):3 9 3 -3 9 7 . 3. United Nations Office on Drugs and Crime. World Drug Report 2 0 1 3 (United Nations publication no. E.13. XI.6). Vienna, Austria: United Nations Office on Drugs and Crime; 2013. 4. Johnston LD, O’Malley PM, Bachman JG, Schulenberg JE. Monitoring the Future National Results on Drug Use: 2 0 1 2 Overview, Key Findings on Adolescent Drug Use. Ann Arbor, MI: Institute for Social Research, University of Michigan; 2013. 5. Australian Institute of Health and Welfare. 2 0 1 0 National Drug Strategy Household Survey Report. Can­ berra, Australian Capital Territory, Australia: Australian Institute of Health and Welfare; 2011. 6. Room R, Fischer B, Hall WD, Lenton S, Reuter P. Cannabis Policy: Moving Beyond Stalemate. The Global Cannabis Commission Report. Oxford, UK: Beckley Foundation; 2008. 7. Hall W. The adverse health effects of cannabis use: what are they, and what are their implications for policy? Int J Drug Policy. 2 0 0 9 ;20(6):458-466. 8. Anderson DM, Hansen B, Rees DI. Medical Mari­ juana Laws and Teen Marijuana Use. Bonn, Germany: Institute for the Study of Labor; 2012. Discussion Paper No. 6592. 9. H arper S, Strum pf EC, Kaufman JS. Do medical m arijuana laws increase m arijuana use? Replication study and extension. A n n Epidemiol. 2 0 1 2;22(3): 2 0 7 -2 1 2 . 10. Khatapoush S, Hallfors D. “Sending the wrong message”: did medical marijuana legalization in Califor­ nia change attitudes about and use of marijuana? J Drug Issues. 2 0 0 4 ;34(4):751-770.

Evans-Whipp et al. | Peer Reviewed | Research and Practice | 9 9 9

RESEARCH AND PR AC TICE

11. Lynne-Landsman SD, Livingston MD, Wagenaar AC. Effects of state medical marijuana laws on adolescent marijuana use. Am JPublic Health. 2013;103(8):15001506. 12. Wall MM, Poh E, Cerda M, Keyes KM, Galea S, Hasin DS. Adolescent marijuana use from 2002 to 2008: higher in states with medical marijuana laws, cause still unclear. Ann Epidemiol. 2011 ;21(9):714-716. 13. Choo EK, Benz M, Zaller N, Warren O, Rising KL, McConnell KJ. The impact of state medical marijuana legislation on adolescent marijuana use.J Adolesc Health. 2014;55(2): 160—166. 14. National Cannabis Prevention and Information Centre. Fact sheet 2: cannabis and the law. Available at: https:// ncpic.org.au/ncpic/publications/factsheets/ article/cannabis-and-the-law. Accessed September 16, 2014.

the School Health Policies and Programs Study 2006. J Sch Health. 2007;77(8):522-543.

alcohol policy on student drinking. Health Educ Res. 2013;28(4):651-662.

28. Ministerial Council on Drug Strategy. National Drug Strategy 2010-2015. Canberra, ACT, Australia: Ministerial Council on Drug Strategy; 2011. Report no. D0224.

46. Desousa C, Murphy S, Roberts C, Anderson L. School policies and binge drinking behaviours of school-aged children in Wales—a multilevel analysis. Health Educ Res. 2008;23(2):259-271.

29. National Institutes of Health Substance Abuse and Mental Health Services Administration. Healthy People 2010. Available at: http://www.healthypeople.gov. Accessed June 4, 2014. 30. Galanti MR, Coppo A, Jonsson E, Bremberg S, Faggiano F. Anti-tobacco policy in schools: upcoming preventive strategy or prevention myth? A review of 31 studies. Tob Control. 2014;23(4):295—301. 31. Adams ML, Jason LA, Pokomy S, Hunt Y. The relationship between school policies and youth tobacco use. J Sch Health. 2009;79(l):17-23.

15. Porath-Waller AJ, Beasley E, Beimess DJ. A metaanalytic review of school-based prevention for cannabis use. Health Educ Behav. 2010;37(5):709-723.

32. Barnett TA, Gauvin L, Lambert M, O’Loughlin J, Paradis G, McGrath JJ. The influence of school smoking policies on student tobacco use. Arch PediatrAdolesc Med. 2007; 161 (9) :842-848.

16. Piontek D, Kraus L, Bjamason T, Demetrovics Z, Ramstedt M. Individual and country-level effects of cannabis-related perceptions on cannabis use: a multi­ level study among adolescents in 32 European countries. J Adolesc Health. 2013;52(4):473-479.

33. Darling H, Reeder AI, Williams S, McGee R. Is there a relation between school smoking policies and youth cigarette smoking knowledge and behaviors? Health Educ Res. 2 0 06;2 1(1 ):10 8 —115.

17. Barrett ME. Increases in marijuana use among eighth grade students in Texas. Subst Use Misuse. 1999;34(12): 1647—1663. 18. Swaim RC. Individual and school level effects of perceived harm, perceived availability, and community size on marijuana use among 12th-grade students: a random effects model. Prev Sci. 2003;4(2):89-98. 19. Coffey C, Lynskey M, Wolfe R, Patton GC. Initiation and progression of cannabis use in a population-based Australian adolescent longitudinal study. Addiction. 2000;95(11):1679-1690. 20. Kuntsche E, Jordan MD. Adolescent alcohol and cannabis use in relation to peer and school factors: results of multilevel analyses. Drug Alcohol Depend. 2006;84 (2): 167-174. 21. Kuntsche E. When cannabis is available and visible at school—a multilevel analysis of students’ cannabis use. Drugs Educ Prev Pol. 2010;17(6):681-688. 22. Hawkins JD, Catalano RF, Miller JY. Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: implications for sub­ stance abuse prevention. Psychol Bull. 1992;112(1 ):64— 105. 23. Ennett ST, Flewelling RL, Lindrooth RC, Norton EC. School and neighborhood characteristics associated with school rates of alcohol, cigarette, and marijuana use. J Health Soc Behav. 1997;38(1):55—71. 24. Kumar R, O’Malley PM, Johnston LD, Schulenberg JE, Bachman JG. Effects of school-level norms on student substance use. Prev Sci. 2002;3(2):105-124. 25. Beyers JM, Evans-Whipp T, Mathers M, Toumbourou JW, Catalano RF. A cross-national com­ parison of school drug policies in Washington State, United States, and Victoria, Australia. J Sch Health. 2005;75(4): 134-140.

34. Evans-Whipp TJ, Bond L, Ukoumunne OC, Toumbourou JW, Catalano RF. The impact of school tobacco policies on student smoking in Washington State, United States and Victoria, Australia. Int J Environ Res Public Health. 2010;7(3):698-710. 35. Kumar R, O’Malley PM, Johnston LD. School tobacco control policies related to students’ smoking and attitudes toward smoking: national survey results, 19992000. Health Educ Behav. 2005;32(6):780-794. 36. Lovato CY, Pullman AW, Halpin PZC, et al. The influence of school policies on smoking prevalence among students in grades 5-9, Canada 2004-2005. Prev Chronic Dis. 2010;7(6):A129. 37. Mumaghan DA Leatherdale ST, Sihvonen M, Kekki P. School-based tobacco-control programming and student smoking behaviour. Chronic Dis Can. 2009;29(4): 169-177. 38. 0verland S, Aar0 LE, Lindbak RL. Associations between schools’ tobacco restrictions and adolescents’ use of tobacco. Health Educ Res. 2010;25(5):748-756. 39. Piontek D, Buehler A, Rudolph U, et al. Social contexts in adolescent smoking: does school policy matter? Health Educ Res. 2008;23(6):1029-1038. 40. Poulin CC. School smoking bans: do they help/do they harm? Drug Alcohol Rev. 2007;26(6):615-624. 41. Reitsma AH, Manske S. Smoking in Ontario schools: does policy make a difference? Can J Public Health. 2004;95(3):214-218. 42. Wiium N, Burgess S, Moore L. Brief report: multilevel analysis of school smoking policy and pupil smoking behaviour in Wales. J Adolesc. 2 0 1 1;34 (2):385-389. 43. Monshouwer K, Van Dorsselaer S, Van Os J, et al. Ethnic composition of schools affects episodic heavy drinking only in ethnic-minority students. Addiction. 2007;102(5):722-729.

26. Evans-Whipp TJ, Bond L, Toumbourou JW, Catalano RF. School, parent, and student perspectives of school drug policies. J Sch Health. 2007;77(3):138-146.

44. Maes L, Lievens J. Can the school make a difference? A multilevel analysis of adolescent risk and health behaviour. Soc Sci Med. 2003;56(3):517-529.

27. Jones SE, Fisher CJ, Greene BZ, Hertz MF, Pritzl J. Healthy and safe school environment, part I: results from

45. Evans-Whipp TJ, Plenty SM, Catalano RF, Herrenkohl TI, Toumbourou JW. The impact of school

1 0 0 0 | Research and Practice | Peer Reviewed | Evans-Whipp et al.

47. Roche AM, Bywood P, Pidd K, Freeman T, Steenson T. Drug testing in Australian schools: policy implications and considerations of punitive, deterrence and/or prevention measures. Int J Drug Policy. 2009;20(6):521-528. 48. Yamaguchi R, Johnston LD, O’Malley PM. Rela­ tionship between student illicit drug use and school drug-testing policies. J Sch Health. 2003;73(4): 159-164. 49. Patton GC, McMorris BJ, Toumbourou JW, Hemphill SA, Donath S, Catalano RF. Puberty and the onset of substance use and abuse. Pediatrics. 2004; 114 (3):e300-e306. 50. McMorris BJ, Hemphill SA, Toumbourou JW, Catalano RF, Patton GC. Prevalence of substance use and delinquent behavior in adolescents from Victoria, Aus­ tralia and Washington State, United States. Health Educ Behav. 2007;34(4):634-650. 51. Bachman JG, Johnston ID, O’Malley PM. Monitoring the Future: Questionnaire Responses From the Nation's High School Seniors, 1988. Ann Arbor, MI: University of Michigan, Institute of Social Research; 2001. 52. Hemphill SA, Toumbourou JW, Herrenkohl TI, McMorris BJ, Catalano RF. The effect of school suspen­ sions and arrests on subsequent adolescent antisocial behavior in Australia and the United States. J Adolesc Health. 2006;39(5):736-744. 53. Moore L, Roberts C, Tudor-Smith C. School smoking policies and smoking prevalence among adolescents: multilevel analysis of cross-sectional data from Wales. Tob Control. 2001; 10(2): 117-123. 54. Lipperman-Kreda S, Grube JW. Students’perception of community disapproval, perceived enforcement of school antismoking policies, personal beliefs, and their cigarette smoking behaviors: results from a structural equation modeling analysis. Nicotine Tob Res. 2009; 11 (5):531—539. 55. Lovato CY, Sabiston CM, Hadd V, Nykiforuk CIJ, Campbell HS. The impact of school smoking policies and student perceptions of enforcement on school smoking prevalence and location of smoking. Health Educ Res. 2007;22(6):782-793. 56. Watts AW, Lovato CY, Card A, Manske SR. Do students’ perceptions of school smoking policies influence where students smoke? Canada’s Youth Smoking Survey. Cancer Causes Control. 2010;21(12):2085-2092. 57. Arcia E. Achievement and enrollment status of suspended students: outcomes in a large, multicultural school district. Educ Urban Soc. 2006;38(3):359-369. 58. Hemphill SA, Heerde JA, Herrenkohl TI, Toumbourou JW, Catalano RF. The impact of school suspension on student tobacco use: a longitudinal study in Victoria, Australia, and Washington State, United States. Health Educ Behav. 2012;39(l):45-56. 59. American Academy of Pediatrics Committee on School Health. Out-of-school suspension and expulsion. Pediatrics. 2003;112(5):1206-1209. 60. Hamilton G, Cross D, Lower T, Resnicow K, Williams P. School policy: what helps to reduce teenage smoking? Nicotine Tob Res. 2003;5(4):507-513.

American Journal o f Public Health | May 2 0 1 5 , Vol 10 5, No. 5

Copyright of American Journal of Public Health is the property of American Public Health Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.