Are you knowledgeable in educational action research?

To educate or to incarcerate: Factors in disproportionality in school discipline☆ Matthew L. Mizel, MSW a,b ,JeremyN.V.Miles,Ph.D. a, Eric R. Pedersen, Ph.D. a, Joan S. Tucker, Ph.D. a, Brett A. Ewing, MS a, Elizabeth J. D'Amico, Ph.D. a,⁎ aRAND Corporation, 1776 Main Street, Santa Monica, CA 90407-2138, United StatesbUniversity of California Los Angeles, Department of Social Welfare, Luskin School of Public Affairs, 3250 Public Affairs Building, Box 951656, Los Angeles, CA 90095-1656, United States abstract article info Article history:

Received 30 May 2016 Received in revised form 7 September 2016 Accepted 7 September 2016 Available online 9 September 2016 The school-to-prison pipeline describes the process by which school suspension/expulsion may push adolescents into the justice system disproportionately based on race/ethnicity, socioeconomic status, and gender. The current study moves theﬁeld forward by analyzing a survey of a diverse sample of 2539 students in 10th to 12th grade in Southern California to examine how demographic, individual, and family factors contribute to disparities in ofﬁce referral and suspension/expulsion. African Americans, boys, and students whose parents had less education were more likely to be suspended/expelled. Higher levels of student academic preparation for class, hours spent on homework, and academic aspiration were associated with less school discipline. Findings suggest that helping students engage in school may be protective against disproportionate school discipline.

Disciplinary disparities Disproportionality Out-of-school suspension Expulsion School discipline 1. Introduction Commonly referred to as the“school-to-prison pipeline”(American Civil Liberties Union, 2012)orthe“cradle-to-prison pipeline” (Children's Defense Fund, 2012), suspension and expulsion can push children out of school and into the juvenile justice system, a process that tends to more severely penalize students of color as well as those who are male, of lower socioeconomic status (SES), and who have dis- abilities (Krezmien, Leone, & Wilson, 2014). Over 2,000,000 secondary school students–or approximately 1 out of 9–were suspended from U.S. middle and high schools during the 2009–2010 school year (Losen & Martinez, 2013). Suspension has increased in frequency in re- cent years coinciding with an increase in the gap in racial disproportionality. Between the 1972–1973 and 2009–2010 school years, rates of suspension doubled for African American (11.8% to 24.3%) and Latino (6.1% to 12.0%) students, whereas rates increased only slightly for White students (6.0% to 7.1%) (Losen & Martinez, 2013).

Krezmien et al. (2014)describe two pathways, one direct and the other indirect, through which suspension and expulsion can lead to stu- dents entering the justice system. In the direct pathway, schools referstudents facing suspension/expulsion directly to the police and courts (Kupchik & Monahan, 2006; Krezmien, Leone, Zablocki, & Wells, 2010). In the indirect pathway, suspensions lead to the youth's discon- nection from school, reduced academic performance, increased delin- quent activity, and incarceration (Butler, Bond, Drew, Krelle, & Seal, 2005; Krezmien et al., 2014; Skiba & Rausch, 2006). Students who are suspended and/or expelled, especially those who are repeatedly disci- plined, are more likely to be held back a grade or to drop out than stu- dents not receiving such discipline (Arcia, 2006; Balfanz, Byrnes, & Fox, 2014; Fabelo et al., 2011; Skiba & Rausch, 2006). School suspension hinders academic growth and contributes to racial disparities in achievement, accounting for approximately one-ﬁfth of black-white dif- ferences in performance (Morris & Perry, 2016). School suspension is also associated with contact with the juvenile justice system the follow- ing year (Fabelo et al., 2011), antisocial behavior (Hemphill, Toumbourou, Herrenkohl, McMorris, & Catalano, 2006), and arrest in that same month versus months when youth had not been suspended or expelled (Monahan, VanDerhei, Bechtold, & Cauffman, 2014).

Most prior research has only included suspension and expulsion as outcome variables, but a smaller number of studies have also found of- ﬁce referrals (i.e., a teacher or school ofﬁcial sent a student to the ofﬁce for disciplinary purposes) for students to be disproportionate based on race, SES, and gender (Bradshaw, Mitchell, O'Brennan, & Leaf, 2010; Rocque, 2010; Skiba, Michael, Nardo, & Peterson, 2002). Ofﬁce referrals are an important form of discipline that can reduce student opportuni- ties to learn (Scott & Barrett, 2004) and increase the risk for future sus- pension and dropout (Morrison & Skiba, 2001 ). As a result, this study Children and Youth Services Review 70 (2016) 102–111 ☆ This work was funded by the National Institute on Alcohol Abuse and Alcoholism (R01AA020883; PI Elizabeth J. D'Amico). The authors wish to thank the districts and schools who participated and supported this project. We would also like to thank Kirsten Becker and Megan Zander-Cotugno for overseeing the survey administrations.

⁎Corresponding author.

E-mail address:[email protected](E.J. D'Amico).

http://dx.doi.org/10.1016/j.childyouth.2016.09.009 0190-7409/© 2016 Elsevier Ltd. All rights reserved. Contents lists available atScienceDirect Children and Youth Services Review journal homepage:www.elsevier.com/locate/childyouth examines both suspension/expulsion and ofﬁce referrals as outcomes.

The termschool disciplinewill refer to all of these outcomes. Doing so also allows a comparison of factors for the different forms of discipline as they can operate via different processes (Gregory, Skiba, & Noguera, 2010).

Despite extensive research linking the demographic factors of race, SES, and gender to school discipline, fewer studies address the role of multiple, varied risk and protective factors with disciplinary actions.

The current study addresses this gap by testing whether individual risk factors (e.g., delinquency, substance use), individual protective fac- tors (e.g., academic engagement and mental health), and family factors (e.g., alcohol and drug use, cultural values about family, parental moni- toring) along with demographic factors (e.g., race/ethnicity, parent ed- ucation, and gender) are associated with school disciplinary action.

2. Literature review 2.1. Disproportionality by race/ethnicity African American and Latino students are negatively affected by dis- proportionate suspension/expulsion rates in comparison to Whites, whereas Asian Americans tend to experience a lower rate of punish- ment than Whites. In 2007, the National Center for Education Statistics (NCES) conducted a nationally representative survey of student disci- pline among public school students in grades 6 through 12. Based on pa- rental reports, lifetime suspension rates were 43% for African Americans, 22% for Latinos, 16% for Whites, 14% for Native American/ Alaskan Natives, and 11% for Asian Americans; lifetime expulsion rates were 13% for African Americans, 3% for Latinos, and 1% for Whites (Aud, Fox, & KewalRamani, 2010). Ofﬁcial school records show that school districts reported suspension rates of 17% for African American students, 8% for Native Americans, 7% for Latinos, 5% for Whites, and 2% for Asian Americans in the 2009–2010 academic year (Losen & Martinez, 2013). A large literature base corroborates these disparities in suspension and/or expulsion (e.g.,Fabelo et al., 2011; Krezmien, Leone, & Achilles, 2006; Raffaele Mendez & Knoff, 2003; Skiba et al., 2011; Wallace, Goodkind, Wallace, & Bachman, 2008). Moreover, Afri- can American students are more likely than students of other races/eth- nicities to experience ofﬁce referral (Bradshaw et al., 2010; Rocque, 2010; Skiba et al., 2002). Importantly, differences in student behavior have not justiﬁed disparities in school discipline across race (Bradshaw et al., 2010; McCarthy & Hoge, 1987; Skiba et al., 2002; Wehlage & Rutter, 1986; Wu, Pink, Crain, & Moles, 1982).

2.2. Disproportionality by socioeconomic status Students of low SES are also more likely to be suspended or expelled (Petras, Masyn, Buckley, Ialongo, & Kellam, 2011; Skiba et al., 2002; Skiba, Peterson, & Williams, 1997; Sullivan, Klingbeil, & Van Norman, 2013; Wu et al., 1982). In particular, level of parents' education as a measure of SES has been associated with suspension rates, with less ed- ucation predicting greater punishment (Hemphill, Plenty, Herrenkohl, Toumbourou, & Catalano, 2014; McCarthy & Hoge, 1987). In one study, when family SES comprised parental education, family income, and parent/guardian occupational prestige, SES was not associated with student misbehavior, but increased SES reduced the likelihood of suspension/expulsion (Peguero & Shekarkhar, 2011).

2.3. Disproportionality by gender Schools tend to suspend boys at a much greater rate than girls (Costenbader & Markson, 1998; Raffaele Mendez & Knoff, 2003; Skiba et al., 1997; Skiba et al., 2002; Wu et al., 1982). According to the 2007 NCES survey previously mentioned, almost twice as many boys thangirlsingrades6through12weresuspended(28%vs.15%) and expelled (4.5% vs. 2.3%) at least once in their lifetime (Aud etal., 2010). In addition, several studiesfound a powerful interaction of race/ethnicity and gender whereby the highest rates of suspension were for African American boys (Losen & Martinez, 2013; Losen & Skiba, 2010; Skiba et al., 2002; Raffaele Mendez & Knoff, 2003; Wallace et al., 2008).

2.4. Individual risk factors Even though schools discipline disproportionately based on race/ ethnicity, SES, and gender, student behavior inﬂ uences punishment (Skiba et al., 2014). In addition, research has indicated that alcohol and drug (AOD) use is associated with lower expectations for academic success (Donovan, 1996; Sutherland & Shepherd, 2001), reduced educa- tional achievement (Degenhardt et al., 2010; Engberg & Morral, 2006; Jeynes, 2002; Lynskey, Coffey, Degenhardt, Carlin, & Patton, 2003; Martins & Alexandre, 2009), and delinquency (D'Amico, Edelen, Miles, & Morral, 2008). To address this, we include both delinquency and AOD use in our analyses.

2.5. Individual protective factors Students with a greater interest in school achievement are less likely to have a history of suspension (Costenbader & Markson, 1998; Morrison, Anthony, Storino, & Dillon, 2001). Further, having high expec- tations for future educational achievement increases the likelihood of high school graduation (Ensminger & Slusarcick, 1992) and reduces dropout among children of immigrants (Rumbaut, 2005). In a study of school resiliency, a combination of teacher and student measures indi- cated that students who were suspended less (both in school and out of school) worked harder, engaged in more learning activities, attended more regularly, were more prepared for class, and expended more effort to complete assignments (Finn & Rock, 1997).

Some limited evidence suggests that mental health may affect the likelihood of school discipline. Data from the National Longitudinal Study of Adolescent Health indicate that students who report persistent depressive symptoms are more likely to be suspended one year later (Rushton, Forcier, & Schectman, 2002). Furthermore, optimism about the future is associated with a lower dropout rate, even after controlling for grade point average, suspension, and SES (Suh, Suh, & Houston, 2007). As prior research has only begun to measure the association be- tween mental health and school discipline, this study aims to address that gap.

2.6. Family factors A variety of research suggests the importance of family on adoles- cents' school outcomes. More family conﬂict and family AOD use are as- sociated with ofﬁce referral (Morrison et al., 2001). Another study found that parental involvement and discussing homework are associated with reduced student misbehavior but not associated with school punishment (Peguero & Shekarkhar, 2011). Familism is often protective for Mexican-American youth and bolsters academic achievement (Valenzuela & Dornbusch, 1994). Increased family cohesion, measured in part as familism, is associated with a greater level of engagement in school and schoolwork discipline (Rumbaut, 2005). Increased family re- spect is also associated with higher academic achievement (Fuligni, Tseng, & Lam, 1999). Given the inﬂuence that family can have on stu- dent academics, we include family factors in our model to increase understanding of how they may affect school discipline and its dispro- portionate application.

3. The present study In sum, a large body of research supports the conclusion that schools discipline students disproportionally based on race/ethnicity, SES, and gender, and that doing so may have severe consequences, such as 103 M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 entry into the justice system. Despite the extensive work examining these demographic factors, fewer studies have explored the following speciﬁc factors either individually or simultaneously: individual factors (academic engagement, mental health) and family relationships (family alcohol/marijuana use, cultural values about family, parental monitor- ing). Furthermore, much of the prior research on these factors has fo- cused on school achievement and dropout and has not addressed school discipline directly. This study adds to the existing literature by examining associations of these multiple individual and family factors with two key disciplinary outcomes, student ofﬁce referrals and suspen- sion/expulsion, thus affording the opportunity to compare results across a broader spectrum of outcomes. In addition, to test whether individual and family factors can reduce disproportionality across demographic factors, this research examines interactions between each of these var- iables and parent education.

Using self-report from a large, diverse sample of high school stu- dents primarily living in Southern California, this study answers the fol- lowing research questions:

1. What individual and family factors are associated with school discipline?

2. Do individual and family factors interact with parent education on the outcome of school discipline, suggesting a mechanism for disproportionality?

4. Methods 4.1. Participants and procedure Participants are part of a longitudinal study of adolescent risk behav- ior (D'Amico et al., 2012). Beginning in the fall of 2008, 16 schools across three districts were selected to participate to obtain a diverse sample.

Seventh- and eighth-grade students received parental consent forms to participate in the study. Ninety-two percent of parents replied with 71% giving permission for their child to participate in the study. As youth graduated from middle school to high school between waves 5 and 6, they transitioned to over 200 high schools nationally and interna- tionally. The current study uses Wave 6 data, which was completed be- tween May, 2013, and April, 2014, when participants were in grades 10 through 12 (n= 2539). Of the youth who were eligible for the Wave 6 survey (6th or 7th grader at Wave 1, could be located, were re- consented), 61% completed the survey. Dropout from prior waves was not associated with demographics or risk behaviors, such as alcohol and marijuana use (Ewing et al., 2015). Because this was theﬁrst wave to measure suspension/expulsion and ofﬁce referral, we could not determine whether those variables associated with survey dropout.

There were 59 (2.3%) African American participants, 532 (21.0%) Asian American, 1115 (43.9%) Latino, 300 (11.8%) Multiracial, and 533 (21.0%) White. Youth who received parental permission to participate and who assented to the surveys were representative of the school populations (D'Amico et al., 2012). No weighting was applied to the analysis. A Cer- tiﬁcate of Conﬁdentiality from the National Institutes of Health was ob- tained to protect survey responses. The institution's internal review board, the school districts, and the individual schools approved all ma- terials and procedures.

4.2. Measures 4.2.1. Sociodemographic characteristics Variables included race/ethnicity, gender, age, and parents' educa- tion. Youth were categorized using NIH guidelines into race/ethnic cat- egories: Hispanic/Latino, White, African American, Asian American, and Multiracial/Other (which included African American multiracial, Native American, Native Hawaiian). For parent education, participants an- swered the question“How far did your father/mother go in school?” for each parent by selecting one of“Did notﬁnish high school”,“Graduated from high school”,“Some college”,“Graduated from col- lege”,“Don't know. ”Answers to the questions were scored on a 1 to 4 scale (4 = graduated from college) for mother and father and then av- eraged into one variable to represent parent education.

4.2.2. Individual risk factors Variables included delinquency and AOD use in the past year. Delin- quency was measured with 4 items that asked how often in the past year the respondent had“been involved inﬁghts”,“taken something from a store that did not belong to you”,“damaged something on pur- pose that did not belong to you”, and“written things or sprayed paint on walls or sidewalks or cars where you were not supposed to” (Ellickson, Tucker, & Klein, 2001; Tucker, Orlando, Ellickson, 2003). Par- ticipants answered on the following scale: not at all, 1–2 times, 3–5 times, 6–9 times, 10–19 times, 20 or more times. Because few partici- pants indicated any level of afﬁrmative response (79.5% answered “not at all”to all 4 questions), answers for the four measures were coded as a dichotomous yes/no and then averaged to form the delin- quency score (α= 0.77). Alcohol and marijuana use in the past year were assessed using Monitoring the Future survey items by asking “During the past year, how many times have you used or tried”each one, specifying with alcohol“at least one drink.”Original responses ranged from 1 =“0times”to 6 =“N20 times”. Responses were dichot- omized (1 =“any use”versus 0 =“no use”) due to infrequent re- sponses at high levels of use (Ellickson, McCaffrey, Ghosh-Dastidar, & Longshore, 2003; WestEd., 2008).

4.2.3. Individual protective factors These variables included four indicators of academic engagement and a measure of mental health. Academic engagement measures were taken from the High School Longitudinal Study of 2009 (Ingels et al., 2011). Participants reported whether theirgrades were important by indicating“getting good grades in school is important to you”on a 5 point scale from strongly agree to strongly disagree.Academic pre- paredness was assessed by asking youth how often they“go to class without your homework done”,“go to class without pencil or paper”, “go to class without books”,and“go to class late”Participants chose from: 1 = never, 2 = rarely, 3 = sometimes, 4 = often. Items were summed to create a total score with lower values indicating more pre- paredness (α=0.75).Forhomework hours,respondentsanswered the question“During a typical weekday during the school year how many hours do you spend working on homework and studying for your classes?”by selecting from:b1h,1to2h,2to3h,3to4h,4to 5 h, 5 or more hours.Academic aspirationwas measured with the ques- tion“What is the highest level of school that you plan toﬁnish?”Partic- ipants chose from:“I may notﬁnish high school; I plan toﬁnish high school; I plan to go to vocational or trade school after high school grad- uation; I plan to go to college, but may notﬁnish; I plan to graduate from college; I plan to go to graduate school or professional school, such as medical or law school.”Answers to the questions were scored on a 1 to 6 scale (1 = I may notﬁnish high school, 6 = I plan to go to graduate school or professional school).Mental healthstatus was assessed using the Mental Health Inventory-5 (MHI-5) (Stewart, Ware, Sherbourne, & Wells, 1992). Participants answered 5 questions about their overall emotional functioning (e.g.,“How much of the time have you felt down- hearted and blue?”) using the following scale: 1 = all of the time, 2 = most of the time, 3 = a good bit of the time, 4 = some of the time, 5 = a little bit of the time, 6 = none of the time. As is customary with the MHI-5, the mean of the 5 responses is the mental health score (α= 0.75).

4.2.4. Family factors Variables included family AOD use, family-related cultural values, and parental monitoring.Sibling alcohol usewas measured by asking “Do any of your older brothers or sisters drink alcohol sometimes?”Sib- ling marijuana usewas measured by asking“Do any of your older 104M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 brothers or sisters smoke marijuana sometimes?”Participants chose from: A = I don't have any older brothers or sisters, B = no, C = yes (D'Amico & Fromme, 1997).Adult alcohol usewas measured by asking “How often does the adult who is most important to you drink alcohol?” Adult marijuana usewas measured by asking“How often does the adult who is most important to you use marijuana?”Respondents used the following scale: 1 = never, 2 = less than once a week, 3 = 1–3daysa week, 4 = 4–7 days a week (Tucker, Ellickson, & Klein, 2003).Familism (e.g.,“if anyone in my family needed help, we would all be there to help them;”α= 0.83) andrespect(e.g.,“it is important to respect my par- ents;”α= 0.90) were assessed with an updated version of the Cultural Values Scale (Unger, Ritt-Olson, Huang, Hoffman, & Palmer, 2002).

Youth responded to 4 items for each construct using the following scale: 1 = strongly agree, 2 = sort of agree, 3 = sort of disagree, 4 = strongly disagree. Items were averaged such that a higher score indicat- ed higher levels of familism and respect.Parental monitoringwas mea- sured using the Communities that Care - Risk and Protective Factor Scale to determine the extent that parents had rules (e.g.,“Rules in my family are clear;”) and monitored youth (e.g.,“My parent(s) or guard- ian(s) ask if I've gotten my homework done.”)(α= 0.79) (Arthur, Hawkins, Pollard, Catalano, & Baglioni, 2002). Youth rated 5 items on the following scale: 1 = strongly agree, 2 = sort of agree, 3 = sort of disagree, 4 = strongly disagree, and items were coded and averaged with higher scores indicating greater parental monitoring.

4.2.5. School discipline Ofﬁce referralandsuspension/expulsionwere measured with two items that asked how often in the past year the respondent had been “sent out of the classroom for causing trouble”and“suspended or ex- pelled from school”(Tucker, Orlando & Ellickson, 2003; Ellickson et al., 2003). Answer choices were: Not at all, 1–2times,3–5 times, 6–9 times, 10– 19 times, 20 or more times. Consistent with prior research, the variables were converted into a dichotomous yes/no for analysis.

Moreover, the distribution of each of these variables was skewed with few reporting ofﬁce referral or suspension/expulsion, preventing treating the variable as ordered.

4.3. Analysis If a participant did not answer a question for a measure, then that participant was removed from the sample for only the regression anal- ysis that used that variable. This allowed for the maximum sample size for each regression but limited between regression comparisons. We conducted logistic regression analyses andﬁrst regressed school disci- pline outcomes (ofﬁce referral, suspended/expelled) independently on the demographic variables of race/ethnicity, gender, parent education, and age. For race/ethnicity, we used African American as the reference group because prior research has consistently found African American students to be disproportionately impacted relative to other racial/eth- nic groups.

To answer research question #1 focused on what individual and family factors are associated with school discipline, we conducted mul- tivariable logistic regressions on both forms of school discipline inde- pendently with the factors grouped into blocks based on their constructs. Model 1 (the base model) included the covariates of race/ ethnicity, gender, parent education, and age. Building on the base model, Block Model 2 added the three individual risk factors, Block Model 3 added theﬁve individual protective factors, and Block Model 4 added the seven family factors to the covariates. For theﬁnal model (Model 5), we conducted separate regressions with ofﬁce referral and suspension/expulsion, including only variables that were statistically signiﬁcant (pb0.05) in the block models for that speciﬁc dependent variable. To answer research question #2, whether individual and fam- ily factors interact with parent education on the outcome of school dis- cipline, we tested interactions with each of the individual risk, individual protective, and family factors while controlling for covariates.Respondents within schools were too sparse to justify clustering within school. Youth in the sample attended over 200 high schools with 103 schools containing only a single student and with only 11 schools hav- ing 30 or more participants, which does not support a multilevel model- ing approach (Bell, Morgan, Kromrey, & Ferron, 2010; Clarke & Wheaton, 2007; Hox, 1998; Maas & Hox, 2004; Maas & Hox, 2005).

5. Results 5.1. School discipline by demographic factors As shown inTable 1, main effects for race, gender, and parent educa- tion existed for both ofﬁce referral and suspension/expulsion. Latinos were most likely to receive ofﬁce referral, and African Americans were most likely to be suspended/expelled. Boys were more likely than girls to receive ofﬁce referral and to be suspended/expelled. Higher parent education associated with less ofﬁce referral and suspension/expulsion.

Age was not a signiﬁcant predictor of ofﬁce referral or suspension/ expulsion.

5.2. Demographic factors and school discipline Including each of the demographic factors/covariates in a single analysis revealed several statistically signiﬁcant effects (Table 2). Stu- dents were more likely to have received an ofﬁce referral if they were male, younger, and if their parents had lower education levels; howev- er, this form of discipline was not signiﬁcantly associated with being Af- rican American relative to other race/ethnicities. Suspension/expulsion was more likely among students who were male, were African Ameri- can (vs. Asian American), and had parents with lower education levels.

5.3. Block models We performed a series of regression analyses with blocks of related factors. Block Model 2 (Individual Risk Factors), Model 3 (Individual Protective Factors) and Model 4 (Family Factors) are shown inTables 3and4. For individual risk factors, delinquency in the past year was as- sociated with a higher likelihood of ofﬁce referral, and both delinquency and marijuana use in the past year were associated with a higher likeli- hood of suspension/expulsion. For individual protective factors, greater academic preparedness and more time spent on homework were both associated with a lower likelihood of ofﬁce referral. In addition, greater academic preparedness, more homework hours and greater academic aspirations were all associated with a lower likelihood of suspension/ expulsion. Sibling alcohol use was associated with a higher likelihood of ofﬁce referral, and adult marijuana use was associated with a higher likelihood of suspension/expulsion.

5.4. Final model For theﬁnal analysis (Model 5 inTables 3 and 4), increased delin- quency was associated with a higher likelihood of ofﬁce referral, where- as increased academic preparedness and homework hours were associated with a lower likelihood of ofﬁce referral. For demographic factors, boys were more likely to receive an ofﬁce referral than girls, and younger students were more likely than older students. Increased delinquency and prior year marijuana use were associated with higher rates of suspension/expulsion. Greater homework hours and academic aspirations were associated with a lower likelihood of suspension/ex- pulsion. African American students were more likely to be suspended/ expelled than Whites and Asian Americans, and those with parents who had lower parent education were also more likely to be suspended/expelled. Gender was no longer a signiﬁcant correlate of suspension/expulsion in theﬁnal model. 105 M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 5.5. Secondary analyses: Interactions of parent education and predictor variables To answer research question #2, we tested interactions between parent education and each predictor variable while controlling for co- variates. There were no statistically signiﬁcant interactions with ofﬁce referral as the outcome, but two variables had signiﬁcant interactions for suspension/expulsion: Grades Important × Parent Education (β= 0.27, SE = 0.11, p = 0.018;Table 5,Fig. 1a) and Parental Monitoring × Parent Education (β= 0.26, SE = 0.13, p = 0.040; Table 5,Fig. 1b). For students with parents who had higher parent edu- cation, rates of suspension/expulsion were unrelated to the importance of grades or amount of parental monitoring. However, for students with parents with lower parent education, suspension/expulsion was more likely among students who placed less emphasis on grades or reported less parental monitoring. Two other interactions are worth noting: Mar- ijuana Use × Parental Education (β= 0.36, SE = 0.19, p = 0.054; Table 5,Fig. 1c) and Adult Alcohol Use × Parental Education (β= 0.18, SE = 0.09, p = 0.064;Table 5,Fig. 1d). For students withparents with higher parental education, rates of suspension/expulsion were unrelated to marijuana use in the past year or adult alcohol use, but for students with parents with lower parental education, suspen- sion/expulsion was more likely among students who reported more marijuana use and adult alcohol use.

6. Discussion This is theﬁrst study to simultaneously examine associations of a broad range of demographic, individual, and family factors with two key forms of school discipline: ofﬁce referral and suspension/expulsion.

Understanding the role that these factors may play in school discipline may help remedy the“school-to-prison pipeline.”Ourﬁrst research question addressed whether speciﬁc individual and family factors were associated with the two outcomes of school discipline while con- trolling for demographic factors of race/ethnicity, parental education, and gender. Results indicate that youth were more likely to report an of- ﬁce referral if they had been involved in more delinquency during the prior year, had lower academic preparedness, and spent fewer hours doing homework. For suspension/expulsion, youth were more likely to report this outcome if they had been involved in more delinquency and if they had used marijuana during the prior year, but also if they spent fewer hours doing homework and had lower academic aspirations.

Among the individual risk factors, delinquency had a more powerful association than prior year alcohol or marijuana use. It is important to note that the delinquency measure was not indicative of the reason for school discipline, but rather was a measure of overall delinquent be- havior during the prior year both during and after school. Thus, as one might expect, youth who were engaging in problem behaviors were also more likely to be disciplined at school. Prior year alcohol use was not related to either ofﬁce referral or suspension/expulsion, but prior year marijuana use was associated with a greater likelihood of suspen- sion/expulsion. Alcohol or drug infractions typically encompassb5% of suspensions (Krezmien et al., 2014; Rausch & Skiba, 2004 ). Therefore, theﬁnding that marijuana use associates with suspension/expulsion suggests that marijuana may be a problem (or a marker for another problem) for youth even if they are not speciﬁcally caught with it on Table 1 School discipline by demographic variables.

Category Number Percent # ofﬁce referral % ofﬁce referral # suspended/expelled % suspended/expelled Race Asian American 532 21.0 31 5.8 10 1.9 Multiracial 300 11.8 36 12.0 11 3.7 White 533 21.0 67 12.6 23 4.3 Latino 1115 43.9 165 14.8 67 6.0 African American 59 2.3 6 10.2 5 8.5 Race main effectχ 2(4) = 27.8, pb0.001χ 2(4) = 16.8, pb0.01 Gender Boys 1162 45.8 187 16.1 69 5.9 Girls 1377 54.2 118 8.6 47 3.4 Gender main effectχ 2(1) = 33.0, pb0.001χ 2(1) = 8.7, pb0.01 Parent education a 1 230 9.6 39 17.0 20 8.7 1.5 120 5.0 25 20.8 4 3.3 2 274 11.4 36 13.1 20 7.3 2.5 184 7.7 23 12.5 10 5.4 3 345 14.4 44 12.8 25 7.3 3.5 191 8.0 22 11.5 5 2.6 4 1054 44.0 97 9.2 23 2.2 Parent education main effectχ 2(6) = 18.3, pb0.001χ 2(6) = 36.26, pb0.001 Age 16 243 11.9 30 12.4 6 2.5 17 971 47.5 121 12.5 39 4.0 18 829 40.6 76 9.2 33 4.0 Age main effectχ 2(2) = 5.3, p = 0.069χ 2(2) = 1.37, p = 0.505 aNote: for an individual respondent for parent education, 1 =“Did notﬁnish high school”,2=“Graduated from high school”,3=“Some college”,and4=“Graduated from college.” Table 2 Model 1–logistic regression of demographic factors on ofﬁce referral and suspension/ expulsion.

Demographic Factor Ofﬁce referral n= 2398Suspended/expelled n= 2397 Beta SE Beta SE (Intercept) 2.01 1.49 1.22 2.28 Age 0.24⁎⁎ 0.09 0.01 0.14 African American Ref. Ref.

White 0.23 0.45 0.70 0.52 Latino 0.27 0.45 0.83 0.50 Asian American 0.76 0.48 1.58⁎⁎ 0.58 Multiracial 0.18 0.47 0.83 0.56 Boys 0.72⁎⁎⁎ 0.13 0.69⁎⁎⁎ 0.20 Parent education 0.20⁎⁎ 0.07 0.42⁎⁎⁎ 0.11 ⁎⁎ pb0.01.

⁎⁎⁎ pb0.001. 106M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 school grounds. Given that youth tend to view marijuana use as having fewer consequences than alcohol use (D'Amico et al., 2013),ﬁndings highlight the importance of focusing on this drug in prevention and in- tervention efforts, especially given recent changes to marijuana policies in several states.

Academic preparedness and homework hours had signiﬁcant associ- ations with ofﬁce referral; academic aspirations and homework hours were associated with suspension/expulsion. Although parental educa- tion was also associated with ofﬁce referrals in the block models, it did not remain signiﬁcant when including delinquency, academic pre- paredness, and homework hours in theﬁnal model. This highlights the importance of delinquent behavior in increasing risk, and also sug- gests that academic preparedness and homework hours may be protec- tive factors that can reduce disproportionality with regards to parent education. This is especially notable as prior research has found higher SES to be associated with a lower likelihood of ofﬁce referral (Rocque, 2010; Skiba et al., 2002). Our results for suspension/expulsion did not indicate a reduction in disproportionality for parent education when the individual risk, individual protective, or family factors were includ- ed. Even though suspension/expulsion was more likely among boys than girls, this gender difference was no longer signiﬁcant when con- trolling for homework hours, academic aspirations, delinquency, and marijuana use. Certainly the risk factors of delinquency and marijuana use are important, but the protectiveﬁndings regarding homework hours and academic aspirations suggest that these indicators of aca- demic engagement merit further attention.Although it is an expectedﬁnding that spending more time on homework may reduce the likelihood of receiving any form of school discipline, it is important to note that the association between both more parent education (with ofﬁce referral) and being a boy (with sus- pension/expulsion) were no longer signiﬁcant with inclusion of home- work hours. Tangible efforts to increase student time on homework may therefore be a protective factor to reduce at least two forms of disproportionality in discipline. Thus, school and community based ef- forts to assist with homework, such as tutoring or increased library hours to provide safe places to study, may be practical ways to reduce unequal school discipline. Furthermore, providing these types of after school activities may also reduce the likelihood that youth will engage in other risk behaviors, such as delinquency and marijuana use (D'Amico et al., 2008). It is not clear the mechanism through which having high aspirations reduced suspension/expulsion, but it may help youth stay more involved in school or affect the way teachers and administrators treat the student. Future longitudinal studies could explore this as well as test whether speciﬁc programs can enhance aca- demic aspirations (e.g., adult mentoring of youth), thereby reducing disproportionality in suspension/expulsion.

Our measure of mental health was not related to either form of school discipline. There is limited work in this area, although one study (Rushton et al., 2002) found that students who reported persis- tent depressive symptoms were more likely to be suspended. Our mea- sure was not an in-depth measure of symptoms or diagnosis and therefore may not have captured the severity of symptoms. Speciﬁcally, the MHI-5 asks respondents about their depressive symptoms for only Table 3 Multivariable logistic regressions on ofﬁce referral.

Factor Model 2 n= 2396Model 3 n= 2392Model 4 n= 2377Model 5 n= 2397 Beta (SE) Beta (SE) Beta (SE) Beta (SE) Demographic (Intercept) 1.42 (1.61) 1.96 (1.66) 3.70 (1.63) 0.31 (1.63) Age 0.34 (0.10) 0.24 (0.09)⁎⁎⁎ 0.27 (0.09)⁎⁎ 0.28 (0.10)⁎⁎ African American Ref. Ref. Ref. Ref.

White 0.02 (0.47) 0.35 (0.46) 0.33 (0.47) 0.14 (0.47) Latino 0.26 (0.46) 0.32 (0.46) 0.47 (0.46) 0.31 (0.46) Asian American 0.57 (0.49) 0.58 (0.49) 0.52 (0.49) 0.49 (0.49) Multiracial 0.09 (0.49) 0.27 (0.48) 0.38 (0.48) 0.14 (0.49) Boys 0.64 (0.14)⁎⁎⁎ 0.55 (0.14)⁎⁎⁎ 0.79 (0.14)⁎⁎⁎ 0.52 (0.14)⁎⁎⁎ Parent education 0.18 (0.07)⁎ 0.13 (0.07) 0.19 (0.07)⁎⁎ 0.11 (0.08) Individual risk Delinquency 1.78 (0.19)⁎⁎⁎ 1.62 (0.17)⁎⁎⁎ Alcohol prior year 0.24 (0.18) Marijuana prior year0.36 (0.20) Individual protective Grades important 0.13 (0.09) Academic preparedness 0.19 (0.02)⁎⁎⁎ 0.16 (0.03)⁎⁎⁎ Homework hours 0.23 (0.05)⁎⁎⁎ 0.24 (0.05)⁎⁎⁎ Academic aspirations 0.05 (0.07) Mental health 0.004 (0.003) Family Sibling alcohol use 0.44 (0.16)⁎⁎ 0.26 (0.15) Sibling marijuana use 0.16 (0.19) Adult alcohol use 0.01 (0.07) Adult marijuana use0.33 (0.11) Familism 0.06 (0.13) Respect 0.22 (0.13) Monitoring 0.19 (0.11) ⁎ pb0.05 ⁎⁎ pb0.01 ⁎⁎⁎ pb0.001 Table 4 Multivariable logistic regressions on Suspended/Expelled.

Factor Model 2 n= 2395Model 3 n= 2391Model 4 n= 2376Model 5 n= 2379 Beta (SE) Beta (SE) Beta (SE) Beta (SE) Demographic (Intercept) 1.49 (2.48) 0.85 (2.43) 0.69 (2.46) 1.84 (2.53) Age 0.10 (0.15) 0.01 (0.14) 0.04 (0.14) 0.07 (0.15) African American Ref. Ref. Ref. Ref.

White 1.10 (0.55)⁎ 0.64 (0.53) 0.53 (0.53) 1.09 (0.54)⁎ Latino 0.92 (0.52) 0.77 (0.51) 0.56 (0.52) 0.92 (0.52) Asian American 1.36 (0.59)⁎ 1.39 (0.59)⁎ 1.24 (0.60)⁎ 1.20 (0.60)⁎ Multiracial 1.10 (0.59) 0.78 (0.57) 0.58 (0.58) 1.05 (0.60) Boys 0.48 (0.22)⁎ 0.48 (0.22)⁎ 0.75 (0.21)⁎⁎⁎ 0.33 (0.23) Parent education 0.42 (0.12)⁎⁎⁎ 0.33 (0.11)⁎⁎ 0.41 (0.11)⁎⁎⁎ 0.33 (0.12)⁎⁎ Individual risk Delinquency 1.42 (0.19)⁎⁎⁎ 1.28 (0.18)⁎⁎⁎ Alcohol prior year 0.50 (0.31) Marijuana prior year1.21 (0.32)⁎⁎⁎ 0.73 (0.25)⁎⁎ Individual protective Grades important 0.15 (0.13) Academic preparedness 0.12 (0.04)⁎⁎ 0.05 (0.04) Homework hours 0.17 (0.08)⁎ 0.17 (0.08)⁎ Academic aspirations 0.24 (0.08)⁎⁎ 0.24 (0.09)⁎⁎ Mental health 0.005 (0.005) Family Sibling alcohol use 0.22 (0.26) Sibling marijuana use0.42 (0.29) Adult alcohol use 0.07 (0.11) Adult marijuana use0.38 (0.14)⁎⁎ 0.05 (0.17) Familism 0.04 (0.20) Respect 0.22 (0.25) Monitoring 0.30 (0.16) ⁎ pb0.05 ⁎⁎ pb0.01.

⁎⁎⁎ pb0.001.107 M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 the last month. Future research could assess mental health more in depth.

Family AOD use, cultural values about family, and parental monitor- ing were not associated with ofﬁce referral or suspension/expulsion in theﬁnal model. Our results are consistent withPeguero and Shekarkhar (2011), who reported that parental involvement, discussing homework, and engaging in activities with children were not related to school punishment. Nevertheless, prior research has shown that these factors are associated with school achievement and dropout (Balfanz, Herzog, & Mac Iver, 2007; Christenson & Thurlow, 2004; Ensminger & Slusarcick, 1992; Rosenthal, 1998). This suggests that for these family factors, the role of family is more important for school completion than for school discipline. It is also important to note that our respect measure was heavily skewed with most students reporting a max level of respect, which likely weakened the ability toﬁnd an effect (Miles, Shih, Tucker, Zhou, & D'Amico, 2012).

Consistent with prior research, African American students reported rates of suspension/expulsion that were signiﬁcantly higher than for White and Asian American youth. The effect of race/ethnicity became even stronger when controlling for students' prior year delinquency and marijuana use. These results suggest that even when youth report similar levels of engagement in these behaviors, African American youth are more likely to be suspended/expelled. Prior work in this area has shown that African American students who committed the same infractions as White youth were more likely to be suspended and expelled (McCarthy & Hoge, 1987; Skiba et al., 2002; Wallace et al., 2008; Wu et al., 1982). Our results add to the research as this dispar- ity occurred even when controlling for individual factors that appeared to reduce disproportionality based on gender. Although African Ameri- can students were more likely to be suspended/expelled, we did not ﬁnd that they were signiﬁcantly more likely than other racial/ethnic groups to report an ofﬁce referral. Thisﬁnding differs from prior studies that utilized ofﬁcial school reports (Rocque, 2010; Bradshaw et al., 2010; Skiba et al., 2002), and it may be the case that our self-report mea- sure (“sent out of the classroom for causing trouble”) captures some- thing different (and perhaps less severe) than ofﬁcial school reports of youth being referred to the principal's ofﬁce. Another explanation is that due to the small number of African American students in thesample uncertainty was introduced into the results, preventing this ef- fect from emerging. The disproportionality in rates of both ofﬁce referral and suspension/expulsion with regards to parent education and gender without covariates are consistent with prior research. Notably, boys were punished at nearly double the rate as girls, similar to the prior na- tional study of suspension/expulsion (Aud et al., 2010).

Findings from our second research question on interactions between parental education and the protective/risk factors indicated that the likelihood of being suspended/expelled from school was signiﬁcantly associated with using marijuana, exposure to adult role model alcohol use, less belief that grades were important, and lower parental monitor- ing only among students who had parents with less education. For those with more highly educated parents, these factors were unrelated to sus- pension/expulsion status. Thus, being from a higher parental education family seemed to prevent these students from receiving severe disci- pline. This suggests that interventions for students from lower parental education backgrounds that address these factors, such as reducing marijuana use or bolstering the belief that grades are important, may be especially helpful at reducing disproportionality in suspension/ expulsion.

It is important to note some limitations of the current study. First, the small number of African American students in the sample reduced the power to ﬁnd race effects, although several effects were statistically signiﬁcant, and power limitations also prevented examining interac- tions with risk/protective factors. Second, we relied exclusively on self-report data, including the measurement of school discipline (it was not feasible to collect ofﬁcial school reports given that participants were from over 200 high schools). The measure of ofﬁce referral asked students if they had been“sent from class,”which could be measuring something different than an ofﬁcial report of ofﬁce referral. Third, the survey did not include certain types of information that would have allowed for moreﬁne-grained analyses. For example, we did not have information on reasons for school discipline and cannot infer that stu- dents who reported delinquent behaviors or substance use necessarily faced disciplinary action due to these behaviors. In addition, we were not able to examine whether students received ofﬁce referrals or sus- pension/expulsion for different infractions based on race, as other stud- ies have shown (e.g.Skiba et al., 2002). We also lacked data on school Table 5 Interactions of parent education and select predictor variables.

Factor Interaction 1 n= 2395Interaction 2 n= 2394Interaction 3 n= 2396Interaction 4 n= 2390 Beta (SE) Beta (SE) Beta (SE) Beta (SE) Covariates (Intercept) 3.61 (2.67) 3.43 (2.63) 0.81 (2.36) 1.73 (2.32) Age 0.00 (0.13) 0.05 (0.14) 0.09 (0.14) 0.01 (0.13) African American Ref Ref Ref Ref White 0.79 (0.52) 0.68 (0.52) 0.75 (0.53) 0.68 (0.52) Latino 0.85 (0.50) 0.77 (0.50) 0.80 (0.51) 0.80 (0.50) Asian American 1.68 (0.58)⁎⁎ 1.60 (0.58)⁎⁎ 1.41 (0.59)⁎ 1.59 (0.58)⁎⁎ Multiracial 0.89 (0.56) 0.81 (0.56) 0.83 (0.58) 0.82 (0.56) Boys 0.62 (0.21)⁎⁎ 0.66 (0.21)⁎⁎ 0.80 (0.21)⁎⁎⁎ 0.70 (0.21)⁎⁎⁎ Parent education 1.63 (0.53)⁎⁎ 1.30 (0.44)⁎⁎ 0.28 (0.14)⁎ 0.26 (0.13) Predictor variables Grades important 1.06 (0.30)⁎⁎⁎ Monitoring 1.20 (0.35)⁎⁎⁎ Marijuana prior year2.44 (0.53) ⁎⁎⁎ Adult alcohol use0.54 (0.27)⁎ Interactions with parent educ.

Grades important 0.27 (0.11)⁎ Monitoring0.26 (0.13)⁎ Marijuana prior year 0.36 (0.19) Adult alcohol use 0.18 (0.09) ⁎ pb0.05.

⁎⁎ pb0.01.

⁎⁎⁎ pb0.001. 108M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 level factors (e.g., suspension rate, weak and inconsistent adult leader- ship, school racial composition;Christle, Jolivette, & Nelson, 2005; Flannery, 1997; Krezmien, 2007; Theriot, Craun, & Dupper, 2010; Welch & Payne, 2010; Wu et al., 1982), as well as staff attitudes and be- havior, which could affect school discipline (Skiba et al., 2014). The sur- vey also did not ask about disability, and prior research has found disproportionality in suspension/expulsion is related to this (U.S.

Department of Education, 2014; Vincent, Sprague, & Tobin, 2012).

Fourth, the cross-sectional analyses preclude us being able to draw any causal conclusions; other factors may have been driving the ob- served relationships, and one cannot determine the way school disci- pline related to the observed variables even when associations existed. Fifth, it is possible that the sample does not include some stu- dents who may have already been expelled or dropped out. However, the sample includes students from schools that focus on at-risk youth and continuation schools, so many who may have been expelled from or dropped out of traditional schools still participated.

Results from this study suggest that when controlling for demo- graphic, individual, and family variables, schools disproportionally sus- pend/expel by student race/ethnicity and parental education, and that schools discipline with ofﬁce referral disproportionally by gender. Our ﬁndings with academic engagement emphasize the need to bolster stu- dent academic preparedness (e.g., teaching study skills), provide safe places to do homework (e.g., restoring library hours that were reduced during budget shortfalls), and support academic aspirations (e.g., through mentoring and university outreach/recruitment programs). In terms of individual risk factors, we found that teen marijuana use was associated with suspension/expulsion, so improved drug prevention and treatment may reduce that punishment. In particular, each of these efforts could be made in communities and with populations that experience disproportionate punishment and that have been historical- ly underserved.Even when controlling for delinquent behavior and marijuana use, rates of suspension/expulsion were higher for African American stu- dents and for those from families with low parental education back- grounds. In order to reduce the disproportionality associated with the demographic factors, school districts could assess and change policies that rely on or reinforce inequality, whether the policies do so inadver- tently or not. As with much prior research, this study suggests that Afri- can American students are suspended and expelled at a greater rate irrespective of behavior. Therefore, schools could better train teachers and administrators to accurately and fairly assess student behavior and apply discipline equally regardless of background. Because dispar- ities increase in schools with police presence and with rigid zero toler- ance policies (Carter, Fine, & Russell, 2014), schools could adopt alternative forms of discipline that are less disruptive than suspension/ex- pulsion (Anyon et al., 2014), such as School-Wide Positive Behavioral Interventions and Supports (Flannery, Fenning, Kato, & McIntosh, 2014; Horner, Sugai, & Anderson, 2010) or restorative justice (Gregory, Clawson, Davis, & Gerewitz, 2014). For a more thorough discussion of school-based interventions, seeGregory, Bell, and Pollock (2014).

Using an intersectionality theoretical framework in which such fac- tors as race, gender, class, sexual orientation, and disability are consid- ered simultaneously could provide great insight (Cho, Crenshaw, & McCall, 2013; Rios, 2011; Shedd, 2015). For example, the school disci- pline of African American girls can be quite different than for African American boys in relation to their gender and race as well as their place (e.g., school, neighborhood) (Morris, 2012). Overall, efforts to re- duce disproportionality should address both the protective/risk factors indicative of academic engagement identiﬁed in the study as well as the demographic factors of race/ethnicity, parental education, and gen- der. Most importantly, the school-to-prison pipeline is our own crea- tion. Because we have grown it in the last 40 years, we can also dismantle it. a.Grades Important x Parent Education b.Parental Monitoring x Parent Education c.Prior Year Marijuana x Parent Education d.Adult Alcohol Use x Parent Education 0 0.1 0.2 0.3 0.4 0.5 0.6 1234 Predicted Probability of Suspension/Expulsion Grades Important 0 0.2 0.4 0.6 0.8 1234 Predicted Probability of Suspension/Expulsion Parental Monitoring 0 0.1 0.2 0.3 0.4 0.5 No Yes Predicted Probability of Suspension/Expulsion Prior Year Marijuana 0 0.050.1 0.150.2 0.250.3 1234 Predicted Probability of Suspension/Expulsion Adult Alcohol Use High Parent Education Low Parent Education Fig. 1.Interactions with parent education.109 M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 References American Civil Liberties Union (2012). Written statement of the American Civil Liberties Union for a hearing on“Ending the School-to-Prison Pipeline.”. Submitted to the Sen- ate Judiciary Subcommittee on the Constitution, Civil Rights and Human Rights. Re- trieved fromhttps://www.aclu.org/ﬁles/assets/aclu_statement_for_sjc_subcomm_ hearing_on_the_school_to_prison_pipeline_12_2012.pdf Anyon, Y., Jenson, J. M., Altschul, I., Farrar, J., McQueen, J., Greer, E., ... Simmons, J. (2014).

The persistent effect of race and the promise of alternatives to suspension in school discipline outcomes.Children and Youth Services Review,44, 379–386.http://dx.doi.

org/10.1016/j.childyouth.2014.06.025.

Arcia, E. (2006). Achievement and enrollment status of suspended students' outcomes in a large, multicultural school district.Education and Urban Society,38(3), 359–369.

http://dx.doi.org/10.1177/0013124506286947.

Arthur, M., Hawkins, J., Pollard, J., Catalano, R., & Baglioni, A., Jr. (2002). Measuring risk and protective factors for substance use, delinquency, and other adolescent problem behaviors: The communities that care youth survey.Evaluation Review,26(6), 575–601.http://dx.doi.org/10.1177/0193841X0202600601.

Aud, S., Fox, M. A., & KewalRamani, A. (2010).Status and trends in the education of racial and ethnic groups, July, 2010 (NCES 2010-015).Washington, D.C.: U.S. Department of Education, National Center for Education Statistics Retrieved fromhttp://nces.ed.

gov/pubs2010/2010015.pdf Balfanz, R., Herzog, L., & Mac Iver, D. J. (2007). Preventing student disengagement and keeping students on the graduation path in urban middle-grades schools: Early iden- tiﬁcation and effective interventions.Educational Psychologist,42(4), 223–235.http:// dx.doi.org/10.1080/00461520701621079.

Balfanz, R., Byrnes, V., & Fox, J. (2014). Sent home and put off-track: The antecedents, disproportionalities, and consequences of being suspended in the ninth grade.

Journal of Applied Research on Children: Informing Policy for Children at Risk,5(2), 13 Retrieved fromhttp://digitalcommons.library.tmc.edu/childrenatrisk/vol5/iss2/13 Bell, B. A., Morgan, G. B., Kromrey, J. D., & Ferron, J. M. (2010).The impact of small cluster size on multilevel models: A Monte Carlo examination of two-level models with binary and continuous predictors. JSM proceedings, survey research methods section., 4057–4067 Retrieved fromhttp://www.amstat.org/sections/srms/Proceedings/ y2010/Files/308112_60089.pdf Bradshaw, C. P., Mitchell, M. M., O'Brennan, L. M., & Leaf, P. J. (2010). Multilevel explora- tion of factors contributing to the overrepresentation of black students in ofﬁce disci- plinary referrals.Journal of Educational Psychology,102(2), 508.http://dx.doi.org/10.

1037/a0018450.

Butler, H., Bond, L., Drew, S., Krelle, A., & Seal, I. (2005).Doing it differently: Improving young people's engagement with school.Melbourne, Australia: Brotherhood of St Laurence.

Carter, P., Fine, M., & Russell, S. (2014). Discipline disparities series: Overview.Bloomington, IN: The Equity Project at Indiana University Retrieved fromhttp://rtpcollaborative.

indiana.edu/brieﬁng-papers/ Children's Defense Fund (2012). Cradle to prison pipeline campaign. Retrieved from http://www.childrensdefense.org/campaigns/cradle-to-prison-pipeline/ Cho, S., Crenshaw, K. W., & McCall, L. (2013).Toward aﬁeld of intersectionality studies:

Theory, applications, and praxis.Signs,38(4), 785–810.

Christenson, S. L., & Thurlow, M. L. (2004). School dropouts prevention considerations, in- terventions, and challenges.Current Directions in Psychological Science,13(1), 36–39.

http://dx.doi.org/10.1111/j.0963-7214.2004.01301010.x.

Christle, C. A., Jolivette, K., & Nelson, C. M. (2005). Breaking the school to prison pipeline:

Identifying school risk and protective factors for youth delinquency.Exceptionality, 13(2), 69–88.http://dx.doi.org/10.1207/s15327035ex1302_2.

Clarke, P., & Wheaton, B. (2007). Addressing data sparseness in contextual population re- search using cluster analysis to create synthetic neighborhoods.Sociological Methods & Research,35(3), 311–351 Retrieved fromhttp://smr.sagepub.com/content/35/3/ 311.full.pdf Costenbader, V., & Markson, S. (1998). School suspension: A study with secondary school students.Journal of School Psychology,36(1), 59–82.http://dx.doi.org/10.1016/S0022- 4405(97)00050-2.

D'Amico, E. J., & Fromme, K. (1997).Health risk behaviors of adolescent and young adult siblings.Health Psychology,16(5), 426–432.

D'Amico, E. J., Edelen, M. O., Miles, J. N., & Morral, A. R. (2008).The longitudinal associa- tion between substance use and delinquency among high-risk youth.Drug and Alcohol Dependence,93(1), 85–92.

D'Amico, E. J., Tucker, J. S., Miles, J. N., Zhou, A. J., Shih, R. A., & Green, H. D., Jr. (2012).

Preventing alcohol use with a voluntary after-school program for middle school stu- dents: Results from a cluster randomized controlled trial of CHOICE.Prevention Science,13(4), 415–425.

D'Amico, E. J., Houck, J. M., Miles, J. N. V., Hunter, S. B., Osilla, K. C., & Ewing, B. A. (2013).

What happens during group and does it matter? An analysis of change talk of at-risk adolescents in a teen court setting.Alcoholism: clinical & experimental research. 37.

(pp. 276A).

Degenhardt, L., Coffey, C., Carlin, J. B., Swift, W., Moore, E., & Patton, G. C. (2010). Out- comes of occasional cannabis use in adolescence: 10-year follow-up study in Victoria, Australia.The British Journal of Psychiatry, 196(4), 290–295.http://dx.doi.org/10.

1192/bjp.bp.108.056952.

Donovan, J. E. (1996). Problem-behavior theory and the explanation of adolescent mari- juana use.Journal of Drug Issues,26(2), 379–404 Retrieved fromhttp://psycnet.apa.

org/psycinfo/1996-05103-004 Ellickson, P. L., Tucker, J. S., & Klein, D. J. (2001).Sex differences in predictors of adolescent smoking cessation.Health Psychology,20,186–195.

Ellickson, P. L., McCaffrey, D. F., Ghosh-Dastidar, B., & Longshore, D. L. (2003). New inroads in preventing adolescent drug use: Results from a large-scale trial of project ALERT inmiddle schools.American Journal of Public Health,93(11), 1830–1836.http://dx.doi.

org/10.2105/AJPH.93.11.1830.

Engberg, J., & Morral, A. R. (2006). Reducing substance use improves adolescents' school attendance.Addiction,101(12), 1741–1751.http://dx.doi.org/10.1111/j.1360-0443.

2006.01544.x.

Ensminger, M. E., & Slusarcick, A. L. (1992). Paths to high school graduation or dropout: A longitudinal study of aﬁrst-grade cohort.Sociology of Education,95–113.http://dx.

doi.org/10.2307/2112677.

Ewing, B. A., Tucker, J. S., Miles, J. N., Shih, R. A., Kulesza, M., Pedersen, E. R., & D'Amico, E. J.

(2015). Early substance use and subsequent DUI in adolescents.Pediatrics,136(5), 868–875.http://dx.doi.org/10.1542/peds.2015-1143.

Fabelo, T., Thompson, M. D., Plotkin, M., Carmichael, D., Marchbanks, M. P., & Booth, E. A.

(2011).Breaking schools' rules: A statewide study of how school discipline relates to stu- dents' success and juvenile justice involvement.New York, NY: Council of State Govern- ments Justice Center Retrieved fromhttp://knowledgecenter.csg.org/kc/content/ breaking-schools-rules-statewide-study Finn, J. D., & Rock, D. A. (1997). Academic success among students at risk for school fail- ure.Journal of Applied Psychology,82(2), 221–234 Retrieved fromhttp://www.viriya.

net/jabref/Academic_success_among_students_at_risk_for_school_failure.pdf Flannery, D. J. (1997).School violence: Risk, preventive intervention, and policy. Urban diversity series no. 109New York: ERIC Clearinghouse on Urban Education Retrieved fromhttp://ﬁles.eric.ed.gov/fulltext/ED416272.pdf Flannery, K. B., Fenning, P., Kato, M. M., & McIntosh, K. (2014).Effects of school-wide pos- itive behavioral interventions and supports andﬁ delity of implementation on prob- lem behavior in high schools.School Psychology Quarterly,29(2), 111.

Fuligni, A. J., Tseng, V., & Lam, M. (1999). Attitudes toward family obligations among American adolescents with Asian, Latin American, and European backgrounds.Child Development,70(4), 1030–1044 Retrieved fromhttp://www.jstor.org/stable/pdf/ 1132260.pdf Gregory, A., Skiba, R. J., & Noguera, P. A. (2010). The achievement gap and the discipline gap two sides of the same coin?Educational Researcher,39(1), 59–68.http://dx.doi.

org/10.3102/0013189X09357621.

Gregory, A., Bell, J., & Pollock, M. (2014a).How educators can eradicate disparities in school discipline: A brieﬁng paper on school-based interventions.Bloomington, IN: The Equity Project at Indiana University Retrieved fromhttp://rtpcollaborative.indiana.edu/ brieﬁng-papers/ Gregory, A., Clawson, K., Davis, A., & Gerewitz, J. (2014b). The promise of restorative prac- tices to transform teacher-student relationships and achieve equity in school disci- pline.Journal of Educational and Psychological Consultation,1–29.http://dx.doi.org/ 10.1080/10474412.2014.929950.

Hemphill, S. A., Toumbourou, J. W., Herrenkohl, T. I., McMorris, B. J., & Catalano, R. F.

(2006). The effect of school suspensions and arrests on subsequent adolescent antiso- cial behavior in Australia and the United States.Journal of Adolescent Health,39(5), 736–744.http://dx.doi.org/10.1016/j.jadohealth.2006.05.010.

Hemphill, S. A., Plenty, S. M., Herrenkohl, T. I., Toumbourou, J. W., & Catalano, R. F. (2014).

Student and school factors associated with school suspension: A multilevel analysis of students in Victoria, Australia and Washington State, United States.Children and Youth Services Review,36, 187–194.http://dx.doi.org/10.1016/j.childyouth.2013.11.

022.

Horner, R. H., Sugai, G., & Anderson, C. M. (2010).Examining the evidence base for school- wide positive behavior support.Focus on Exceptional Children,42(8), 1–14 (Retrieved from, http://citeseerx.ist.psu.edu/viewdoc/download?doi = 10.1.1.458.

1803&rep = rep1&type = pdf).

Hox, J. (1998).Multilevel modeling: When and why. In I. Balderjahn, R. Mathar, & M.

Schader (Eds.),Classiﬁcation, data analysis, and data highways(pp. 147–154). New York: Springer Verlag.

Ingels, S. J., Pratt, D. J., Herget, D. R., Burns, L. J., Dever, J. A., Ottem, R., ... Leinwand, S.

(2011).High School Longitudinal Study of 2009 (HSLS:09). Base-year dataﬁle documen- tation (NCES 2011-328). U.S. Department of Education.Washington, DC: National Cen- ter for Education Statistics Retrieved fromhttp://nces.ed.gov/pubsearch Jeynes, W. H. (2002). The relationship between the consumption of various drugs by ad- olescents and their academic achievement.The American Journal of Drug and Alcohol Abuse,28(1), 15–35.http://dx.doi.org/10.1081/ADA-120001279.

Krezmien, M. P. (2007).Understanding disproportionate suspensions of minority students and students with disabilities: A multilevel approach.(Unpublished doctoral disserta- tion) College Park, MD: University of Maryland.

Krezmien, M. P., Leone, P. E., & Achilles, G. M. (2006). Suspension, race, and disability:

Analysis of statewide practices and reporting.Journal of Emotional and Behavioral Disorders,14(4), 217–226.http://dx.doi.org/10.1177/10634266060140040501.

Krezmien, M. P., Leone, P. E., Zablocki, M. S., & Wells, C. S. (2010). Juvenile court referrals and the public schools: Nature and extent of the practice inﬁve states.Journal of Contemporary Criminal Justice,26(3), 273–293.http://dx.doi.org/10.1177/ 1043986210368642.

Krezmien, M. P., Leone, P. E., & Wilson, M. G. (2014).Marginalized students, school exclu- sion, and the school-to-prison pipeline. In W. T. ChurchII, D. W. Springer, & A. R.

Roberts (Eds.),Juvenile justice sourcebook(pp. 267–288). New York: Ox University Press.

Kupchik, A., & Monahan, T. (2006). The new American school: Preparation for post-indus- trial discipline.British Journal of Sociology of Education,27(5), 617–631.http://dx.doi.

org/10.1080/01425690600958816.

Losen, D. J., & Martinez, T. E. (2013).Out of school & off track: The overuse of suspensions in American middle and high schools.Los Angeles, CA: The UCLA Center for Civil Rights Remedies at The Civil Rights Project Retrieved fromhttp://civilrightsproject.ucla.

edu/resources/projects/center-for-civil-rights-remedies/school-to-prison-folder/ federal-reports/out-of-school-and-off-track-the-overuse-of-suspensions-in- american-middle-and-high-schools 110M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111 Losen, D. J., & Skiba, R. (2010).Suspended education: Urban middle schools in crisis.Mont- gomery, AL: Southern Poverty Law Center Retrieved fromhttps://escholarship.org/ uc/item/8fh0s5dv Lynskey, M. T., Coffey, C., Degenhardt, L., Carlin, J. B., & Patton, G. (2003). A longitudinal study of the effects of adolescent cannabis use on high school completion.

Addiction,98(5), 685–692.http://dx.doi.org/10.1046/j.1360-0443.2003.00356.x.

Maas, C. J., & Hox, J. J. (2004). Robustness issues in multilevel regression analysis.Statistica Neerlandica,58(2), 127–137 Retrieved fromhttp://joophox.net/publist/sn04.pdf Maas, C. J., & Hox, J. J. (2005). Sufﬁcient sample sizes for multilevel modeling.

Methodology,1(3), 86–92 Retrieved fromhttp://www.joophox.net/publist/ methodology05.pdf Martins, S. S., & Alexandre, P. K. (2009). The association of ecstasy use and academic achievement among adolescents in two US national surveys.Addictive Behaviors, 34(1), 9–16.http://dx.doi.org/10.1016/j.addbeh.2008.07.022.

McCarthy, J. D., & Hoge, D. R. (1987). The social construction of school punishment: Racial disadvantage out of universalistic process.Social Forces,65(4), 1101–1120.http://dx.

doi.org/10.1093/sf/65.4.1101.

Miles, J. N., Shih, R. A., Tucker, J. S., Zhou, A., & D'Amico, E. J. (2012). Assessing measure- ment invariance of familism and parental respect across race/ethnicity in adolescents.

BMC Medical Research Methodology,12(1), 61 Retrieved fromhttp://www.

biomedcentral.com/1471-2288/12/61 Monahan, K. C., VanDerhei, S., Bechtold, J., & Cauffman, E. (2014). From the school yard to the squad car: School discipline, truancy, and arrest.Journal of Youth and Adolescence, 43(7), 1110–1122.http://dx.doi.org/10.1007/s10964-014-0103-1.

Morris, M. W. (2012).Race, gender, and the“School to Prison Pipeline”: Expanding our dis- cussion to include black girls.

Morris, E. W., & Perry, B. L. (2016).The punishment gap: School suspension and racial dis- parities in achievement.Social Problems,63(1), 68–86.

Morrison, G. M., & Skiba, R. (2001). Predicting violence from school misbehavior: Prom- ises and perils.Psychology in the Schools,38(2), 173–184.http://dx.doi.org/10.1002/ pits.1008 .

Morrison, G. M., Anthony, S., Storino, M., & Dillon, C. (2001). An examination of the disci- plinary histories and the individual and educational characteristics of students who participate in an in-school suspension program.Education and Treatment of Children,24(3), 276–293 Retrieved fromhttp://www.jstor.org/stable/42899662 Peguero, A. A., & Shekarkhar, Z. (2011). Latino/a student misbehavior and school punish- ment.Hispanic Journal of Behavioral Sciences,33(1), 54–70.http://dx.doi.org/10.1177/ 0739986310388021.

Petras, H., Masyn, K. E., Buckley, J. A., Ialongo, N. S., & Kellam, S. (2011). Who is most at risk for school removal? A multilevel discrete-time survival analysis of individual-and context-level inﬂuences.Journal of Educational Psychology,103(1), 223.http://dx.

doi.org/10.1037/a0021545.

Raffaele Mendez, L. M., & Knoff, H. M. (2003). Who gets suspended from school and why:

A demographic analysis of schools and disciplinary infractions in a large school dis- trict.Education and Treatment of Children,26(1), 30–51 Retrieved fromhttp://www.

jstor.org/stable/42900535 Rausch, M. K., & Skiba, R. (2004). Unplanned outcomes: Suspensions and expulsions in In- diana.Education policy briefs. 2, . Bloomington, IN: Center for Evaluation and Educa- tion Policy Number 2, Summer 2004. Retrieved fromhttp://ﬁles.eric.ed.gov/fulltext/ ED488917.pdf Rios, V. M. (2011).Punished: Policing the lives of Black and Latino boys.NYU Press.

Rocque, M. (2010). Ofﬁce discipline and student behavior: Does race matter?American Journal of Education,116(4), 557–581.http://dx.doi.org/10.1086/653629.

Rosenthal, B. S. (1998).Non-school correlates of dropout: An integrative review of the lit- erature.Children and Youth Services Review,20(5), 413–433.

Rumbaut, R. G. (2005).Children of immigrants and their achievement: The roles of family, acculturation, social class, gender, ethnicity, and school context. Addressing the achieve- ment gap: Theory informing practice, 23–59.

Rushton, J. L., Forcier, M., & Schectman, R. M. (2002). Epidemiology of depressive symp- toms in the National Longitudinal Study of Adolescent Health.Journal of the American Academy of Child and Adolescent Psychiatry,41(2), 199–205.http://dx.doi.

org/10.1097/00004583-200202000-00014.

Scott, T. M., & Barrett, S. B. (2004). Using staff and student time engaged in disciplinary procedures to evaluate the impact of school-wide PBS.Journal of Positive Behavior Interventions ,6(1), 21–27.http://dx.doi.org/10.1177/10983007040060010401.

Shedd, C. (2015).Unequal City: Race, schools, and perceptions of injustice.Russell Sage Foundation.

Skiba, R. J., & Rausch, M. K. (2006).Zero tolerance, suspension, and expulsion: Questions of equity and effectiveness. In C. M. Evertson, & C. S. Weinstein (Eds.),Handbook ofclassroom management: Research, practice, and contemporary issues (pp. 1063–1089). Mahwah, New Jersey: Lawrence Erlbaum Associates, Publishers.

Skiba, R. J., Peterson, R. L., & Williams, T. (1997). Ofﬁce referrals and suspension: Disciplin- ary intervention in middle schools.Education and Treatment of Children,20(3), 295–315 Retrieved fromhttp://www.jstor.org/stable/42900491 Skiba, R. J., Michael, R. S., Nardo, A. C., & Peterson, R. L. (2002). The color of discipline:

Sources of racial and gender disproportionality in school punishment.The Urban Review,34(4), 317–342.http://dx.doi.org/10.1023/A:1021320817372.

Skiba, R. J., Horner, R. H., Chung, C. G., Karega Rausch, M., May, S. L., & Tobin, T. (2011).

Race is not neutral: A national investigation of African American and Latino disproportionality in school discipline.School Psychology Review,40(1), 85 Retrieved fromhttp://youthjusticenc.org/download/education-justice/disparities/Race%20is% 20not%20Neutral.pdf Skiba, R. J., Chung, C. G., Trachok, M., Baker, T. L., Sheya, A., & Hughes, R. L. (2014). Parsing disciplinary disproportionality contributions of infraction, student, and school char- acteristics to out-of-school suspension and expulsion.American Educational Research Journal,51(4), 640–670.http://dx.doi.org/10.3102/0002831214541670.

Stewart, A. L., Ware, J. E., Sherbourne, C. D., & Wells, K. B. (1992).Psychological distress/ well-being and cognitive functioning measures. In A. Stewart, & J. E. Ware (Eds.), Measuring functioning and well-being: The medical outcomes study approach (pp. 102–142). Durham, NC: Duke University Press.

Suh, S., Suh, J., & Houston, I. (2007). Predictors of categorical at-risk high school dropouts.

Journal of Counseling and Development,85(2), 196–203.http://dx.doi.org/10.1002/j.

1556-6678.2007.tb00463.x.

Sullivan, A. L., Klingbeil, D. A., & Van Norman, E. R. (2013). Beyond behavior: Multilevel analysis of the inﬂuence of sociodemographics and school characteristics on students' risk of suspension.School Psychology Review,42(1), 99 Retrieved fromhttp://search.

proquest.com/openview/fb33a4f43b32623eba2fb1675922b7d4/1?pq-origsite= gscholar Sutherland, I., & Shepherd, J. P. (2001). Social dimensions of adolescent substance use.

Addiction,96(3), 445–458.http://dx.doi.org/10.1046/j.1360-0443.2001.9634458.x (Retrieved from DOI:).

Theriot, M. T., Craun, S. W., & Dupper, D. R. (2010). Multilevel evaluation of factors predicting school exclusion among middle and high school students.Children and Youth Services Review,32(1), 13–19.http://dx.doi.org/10.1016/j.childyouth.2009.06.

009.

Tucker, J. S., Ellickson, P. L., & Klein, D. J. (2003a).Predictors of the transition to regular smoking during adolescence and young adulthood.Journal of Adolescent Health, 32(4), 314–324.

Tucker, J. S., Orlando, M., & Ellickson, P. L. (2003b).Patterns and correlates of binge drink- ing trajectories from early adolescence to young adulthood.Health Psychology,22(1), 79–87.

U.S. Department of Education (2014). Civil rights data collection; data snapshot: School discipline. Issue brief number 1 (March, 2014). Retrieved fromhttps://www2.ed.

gov/about/ofﬁces/list/ocr/docs/crdc-discipline-snapshot.pdf Unger, J. B., Ritt-Olson, L. T., Huang, T., Hoffman, B. R., & Palmer, P. (2002). Cultural values and substance use in a multiethnic sample of California adolescents.Addiction Research and Theory,10(3), 257–279.http://dx.doi.org/10.1080/16066350211869.

Valenzuela, A., & Dornbusch, S. M. (1994). Familism and social capital in the academic achievement of Mexican origin and Anglo adolescents.Social Science Quarterly, 75(1), 18–36 Retrieved fromhttp://psycnet.apa.org/psycinfo/1995-15273-001 Vincent, C. G., Sprague, J. R., & Tobin, T. J. (2012). Exclusionary discipline practices across students' racial/ethnic backgrounds and disability status: Findings from the Paciﬁc Northwest.Education and Treatment of Children,35(4), 585–601.http://dx.doi.org/ 10.1353/etc.2012.0025.

Wallace, J. M., Jr., Goodkind, S., Wallace, C. M., & Bachman, J. G. (2008). Racial, ethnic, and gender differences in school discipline among US high school students: 1991–2005.

The Negro Educational Review,59(1–2), 47 Retrieved fromhttp://www.ncbi.nlm.nih.

gov/pmc/articles/PMC2678799/ Wehlage, G. G., & Rutter, R. A. (1986). Dropping out: How much do schools contribute to the problem?Teachers College Record,87(3), 374–393 Retrieved fromhttp://eric.ed.

gov/?id=ED275799 Welch, K., & Payne, A. A. (2010). Racial threat and punitive school discipline.Social Problems,57(1), 25– 48.http://dx.doi.org/10.1525/sp.2010.57.1.25.

WestEd (2008). California healthy kids survey. Available at:http://chks.wested.org Wu, S. C., Pink, W., Crain, R., & Moles, O. (1982). Student suspension: A critical reappraisal.

The Urban Review,14(4), 245–303.http://dx.doi.org/10.1007/BF02171974.111 M.L. Mizel et al. / Children and Youth Services Review 70 (2016) 102–111