PEER GROUPS: CHANNELS FOR FRIENDSHIPS AND RISK BEHAVIORS?. Michele Mouttapa1, Jennifer Unger2, Steven Sussman2, Thomas Valente2, 1Johns Hopkins University, Columbia, MD United States; 2University of Southern California, Alhambra, CA United States
Adolescents may engage in health-risk behaviors in an effort to develop a social identity among their peers. Previous research suggests that self-reports of identification with stereotypical high-risk peer groups (e.g., “gangsters”, “rappers”) are associated with health risk behaviors including violence, smoking, and substance use. Little is known about whether self-reported peer groups correspond with actual friendship networks. In the present study we examined whether identification with high-risk peer groups was associated with: (1) health-related variables (e.g., smoking, violence), (2) social network characteristics (e.g., number of friendship nominations received and reciprocated), and (3) friendships with other high-risk friends.
Participants were 2,822 Southern Californian 7th grade girls and boys who were in the second year of a longitudinal school-based experimental trial of smoking prevention strategies. Approximately half of the sample was Latino and nearly one quarter was Asian. Participants completed a survey that included items about health-risk behaviors, the peer groups they identified with (e.g., “jocks”, “gangsters”), and friendship nominations from a list of their classmates.
High-risk students were those who identified with one or more of the following groups: “rockers”, “skaters”, “gangsters”, and “immigrants.” High-risk students were over twice as likely to have ever smoked relative to other students, and were more likely to be aggressive victims in the previous year. High-risk students also reported markedly higher rates of hostility and depressive symptoms, lower rates of parental monitoring, communication with parents, school bonding, and lower grades. Overall, high-risk students did not differ from other students in the number of friendship nominations that they received and reciprocated. However those who identified with “jocks” and “popular students” received the largest number of friendship nominations. High-risk students had a slightly higher percentage of high-risk friends (40%) than did other students (32%).
The findings suggest that self-identification with specific peer groups are strongly associated with health-risk behaviors. Despite their lower levels of psychosocial functioning, poorer parent-child relationships, and school performance, high-risk students are not isolated by their classmates. Furthermore actual friendship groups are relatively heterogeneous in regards to peer group self-identification. Hence, many students who do not identify with high-risk groups may still be exposed to the behaviors and attitudes of their high-risk friends. Implications are discussed.
EXPERIENCES OF PHYSICAL AGGRESSION AMONG UNIVERSITY STUDENTS. Paul Tremblay1, Kathryn Graham1, Jennifer Jelley1, 1Centre for Addiction and Mental Health, London, Ontario Canada
Students´ experiences of physical aggression were investigated as part of a large study on alcohol-related aggression. Random samples of 1500 full-time undergraduate students at six Canadian universities (total N = 9000) were invited by email to participate in a web-based aggression study. A total of 2647 participants (63.6% female) reported their experiences of physical aggression in the past 12 months at a (1) bar, nightclub or pub, (2) home or residence, (3) party or social event, and (4) other locations. Participants were asked about their role in the incidents, the gender of the opponent(s), whether alcohol was involved, and the severity of aggression used by them and the opponent(s). We found that 24.6% of all participants (30.2% males and 21.5% females) experienced at least one episode of aggression. In terms of location, 12.7% of students (16.4% males and 10.6% females) experienced one episode of aggression at a bar/nightclub/pub, 10.5% at a home/residence (11.3% male and 10.0% female), 4.8% at a party/social event (7.9% males and 3.0% females), and 4.4% at other locations (7.2% males and 2.8% females). Male participants often reported that their aggressive experiences occurred with one or more males (95.4% at a bar/nightclub/pub, 80.4 % at a home/residence, and 95.8% at a party or social event) whereas female participants reported only slightly more same sex than opposite sex opponents at a bar/nightclub/pub (53.4%) and only a minority of aggressive experiences with same sex opponents at a party/social event (39.6%) and at a home/residence (29.4%). Male participants were more likely than female participants to have engaged in mutual or one-sided aggression as the aggressor at a bar (males = 63.8%; females = 49.4%) at a home/residence (males = 82.5%; females = 61.5%), and at a party or social event (males = 68.1%, females = 46.9%). Severity of participants´ aggression was associated with their alcohol consumption at the time of the incident when the aggressive experience occurred at a bar, party or social event but not when it occurred at a home/residence. We report on the influence of additional participant characteristics including trait aggression, perceived likelihood of aggressive retaliation in hypothetical conflict situations, and patterns of alcohol consumption.
PROTECTIVE EFFECTS ON SMOKING AND DRINKING OF ALCOHOL AVAILABILITY AND CONTROLS: FINDINGS OF A NATIONAL MULTILEVEL STUDY. Elissa Weitzman1, Ying-Yeh Chen2, 1Harvard University, Boston, MA United States; 2Harvard University, Taipei City Pschiatric Institute, Boston, MA United States
AIMS. To determine whether community patterns of alcohol availability and controls would affect individual risks for smoking and binge drinking among a national sample of youth in college and to assess differential effects of alcohol availability and controls on outcomes for youth with varying perceptions about alcohol´s availability and high and low sensitivities to its control. METHODS. Data were drawn from the 2001 Harvard School of Public Health College Alcohol Study (CAS), an anonymous self-report mailed survey of n=10,924 students at 120 nationally representative colleges. Outcomes included past 30 day smoking and binge drinking. Predictors included individual and contextual level measures of alcohol availability, policy enforcement and controls, with contextual measures estimated as the percentage of students at each college who thought alcohol was “very easy” to obtain, school drinking policies were “strongly enforced,” and it was “very likely” that parents would be notified if a student were caught illegally trying to obtain alcohol in the community. A hierarchical multilevel modeling strategy was used to simultaneously estimate the odds of smoking and binge drinking for each measure of alcohol availability and controls, considering main effects of individual perceptions on two simultaneously estimated outcomes, then the main effect plus the contextual measure built from it, then a final model controlling for school binge prevalence. MLwin multilevel software with Marginal Quasilikelihood first-order estimation procedures were used adjusting for demographic, social and college confounders plus response rate. Cross-level interactions of individual perceptions about availability, enforcement and controls and contextual measures of the same in multilevel multivariate models were tested. RESULTS. Contextual measures of the alcohol environment independently predicted individual risks of smoking and binge drinking, over and above effects of individual perceptions about these factors, compositional characteristics of students, college characteristics and binge drinking rates. Contextual effects on the adjusted odds (95% CIs) for smoking and binge drinking were as follows for measures of: easy availability 4.04 (2.00, 8.15) & 4.01 (2.18, 7.39); strong enforcement 0.34 (0.18, 0.63) & 0.19 (0.11, 0.33); and parental notification 0.07 (0.02, 0.30) & 0.07 (0.02, 0.23) all p-values <.001. Significant cross-level interactions were observed for both outcomes. CONCLUSIONS. In college communities, where tobacco and alcohol use co-occurrence is high there is strong evidence that aspects of the drinking world also shape smoking risks. Implications for prevention theory and practice are discussed.
PREVALENCE AND COMORBIDITY OF MOOD, ANXIETY, PERSONALITY, AND ALCOHOL USE DISORDERS AMONG RURAL MEN AND WOMEN IN EMERGING ADULTHOOD: FINDINGS FROM PROJECT FAMILY. W. Alex Mason1, Rick Kosterman1, Kevin P. Haggerty1, J. David Hawkins1, Cleve Redmond2, Richard L. Spoth2, 1University of Washington, Seattle, WA United States; 2Iowa State University, Ames, IA United States
Emerging adulthood is a peak period of risk for the occurrence and co-occurrence of mood, anxiety, personality, and alcohol use disorders (Newman et al., 1996). Several epidemiological studies have estimated the prevalence of these disorders based on Diagnostic and Statistical Manual criteria (American Psychiatric Association, 1994), primarily among urban and suburban samples (Mason et al., 2004; Newman et al., 1996). Findings from these samples indicate that mental illness is widespread and that the personal and social costs of psychiatric disorders are high (Kessler et al., 2003). Less is known about mental illness among young men and women in rural settings. Knowing which psychiatric disorders and combinations of disorders are most prevalent in rural areas could help guide planning for implementation of preventive interventions that target this type of special population. Using a modified version of the Diagnostic Interview Schedule (DIS; Robins et al., 1981), this study examined the prevalence and comorbidity of major depression, generalized anxiety, conduct disorder, antisocial personality disorder, alcohol abuse, and alcohol dependence in a sample of rural men and women (average age = 21.6 years). The data were collected as part of Project Family, a longitudinal efficacy trial of the Iowa Strengthening Families and Preparing for the Drug Free Years programs. The study, which began in 1993, is now in its 11th year, and includes 7 waves of multi-rater, multi-method data collected from 667 target participants and their families. Analyses were restricted to individuals in the control condition of the randomized, controlled trial who were originally recruited from 11 schools in a Midwestern state. Overall, 27% of control participants met criteria for a disorder within the year prior to the diagnostic assessment that occurred at age 21. The most prevalent disorders were alcohol abuse (21%) and major depression (9.3%); these two disorders tended to co-occur. Current analyses are examining adolescent psychosocial factors, such as parent-child relationship quality and drug use refusal skills, that prospectively predict young adult mental disorders. Findings from this collection of analyses will increase our understanding of the nature and scope of mental illness among young men and women in rural settings, and will have implications for understanding the etiology and prevention of psychiatric disorders.
HIGH-RISK SEXUAL BEHAVIORS AND USES OF DRUGS AMONG MINORITY COLLEGE STUDENTS. Yan Wang1, Dorothy Browne1, Patty Clubb1, Fernando Wagner1, 1Morgan State University, Baltimore, MD United States
Introduction: College students often engage in risk behaviors that increase their rates of morbidity and mortality. For example, they often experiment with drugs (O'Malley PM & Johnston LD, 2002). Additionally, a great majority of college students are sexually active, and most do not consistently protect themselves or their partners against pregnancy or the transmission of sexually transmitted diseases (STD) (CDC, 1997). Until recently, few researchers have examined the high-risk sexual behaviors and uses of drugs among African-American students at Historically Black Colleges and Universities (HBCUs). Methodology: Totally 10,132 African American freshmen of 35 HBCUs participated in a self-administered paper-and-pencil survey. Analyses were restricted to 6052 (60%) unmarried freshmen with sex experiences, but without same-gender sex experiences. Conditional logistic regression models were performed, while matching on universities attended and family income and adjusting for other demographic covariates. Results: Nearly half (47%) of the sample had more than 4 different lifetime sex partners. About 41% did not always use condoms during sexual activities. An estimated 21% had ever been pregnant or gotten someone pregnant. About 9% had a history of STD infection. The majority (74%) had used at least one of the following drugs during their lifetime: tobacco (31%), alcohol (71%), marijuana (38%) or other illicit drugs (4%). About one third ever used only one drug (30%); 23% only two drugs, and 21% three or more drugs. Users of one drug, two drugs, three or more drugs were separately 1.4 times (OR=1.4; 95% CI=1.0, 2.0), 1.8 times (OR=1.8, 95% CI=1.3, 2.5) and 3.0 times (OR=3.0; 95% CI=2.1, 4.2) more likely to have STD infection history compared to non-users. Similarly they were separately 1.4 times (OR=1.4; 95% CI=1.2, 1.7), 2.5 times (OR=2.5; 95% CI=2.1, 3.1) and 4.2 times (OR=4.2; 95% CI=3.4, 5.1) more likely to have more than 4 lifetime sex partners. Drug users were also more likely to have been pregnant or get someone pregnant. Although users of one drug were not significantly less likely, users of two drugs and three or more drugs were 0.7 times (OR=0.7; 95% CI= 0.6, 0.8) and 0.5 times (OR=0.5; 95% CI=0.4, 0,6), respectively, less likely to use condoms compared to non-users. These results held consistent across different genders. Comments: Risky sex behavior prevention may be combined with prevention of drug uses among minority college students. Acknowledgements: CHDS grant 5P60MD000217-03 and NIDA grant 5U24DA012390-05.
IDENTIFYING OPTIMAL RISK AND PROTECTIVE FACTORS TO PREDICT ADOLESCENT SUBSTANCE USE. Kelly R. Breeden1, Nicoletta Lomuto2, P. Allison Minugh2, 1DATACORP, Cheyenne, WY United States; 2DATACORP, Providence, RI United States
This poster explores the relationship between risk factors, protective factors, and past 30-day substance use. The central question is to ascertain if there are certain risk and protective factors that are stronger predictors for adolescent substance use. The data are drawn from a census of approximately 20,000 students in grades 6, 8, 10, and 12 in a large, sparsely populated western state. The questionnaire measured student scores on 34 different risk and protective factor scales and on past 30-day use of 12 different substances. A cut point was used to transform the risk scales into dichotomous variables indicating whether a youth was at risk or not at risk. Protective factor scales were also dichotomized using cut points. By summing the dichotomous risk factor variables, we created a composite variable that represents the total number of risk factors for each student. An analogous composite variable was created for the total number of protective factors. The poster focuses on two aspects of the relationship between risk, protection, and past 30-day substance use. The first aspect pertains to the disparity in overall risk and protection between users and non-users. How wide are the disparities? Does the size of the disparity vary by substance? These questions are investigated by comparing the average number of risk factors among users of the substance with the average number among non-users. A parallel comparison was performed using the average number of protective factors. Charts display these disparities for both risk and protection by substance.The second aspect concerns the relationship between particular risk and protective factors and substance use. Which risk and protective factors best predict the use of each substance? To shed light on this question, the odds ratios for students at risk versus those not at risk with respect to using each substance were calculated. Similar odds ratios were calculated for those students that were protected versus those that were not. A summary of these calculations is presented.
CONCURRENT 4, CULTURAL SENSITIVIY, Grouped papers
SCHOOL BASED INTERVENTIONS AND HIGH RISK MINORITY YOUTH
Chair: Esteban Cardemil
DOES RACE MATTER? TRANSLATING A SCHOOL-BASED INTERVENTION FOR LOW-INCOME MINORITY YOUTHS FROM ONE RACIAL/ETHNIC GROUP TO ANOTHER. Inna Altschul1, Daphna Oyserman1, 1University of Michigan, Ann Arbor, MI United States
Once an evidence-based intervention has been established and evaluated, the challenge of translating it to the field remains because each implementation introduces new parameters. Each new implementation of an intervention requires thinking through the implications of these new parameters for the underlying theory of change within the intervention. Practitioners and researchers must ask (1) how do the new population and/or context affect the mechanisms through which the intervention is hypothesized to work? (2) What changes must be made in the implementation of the intervention to account for the new population and/or context? And (3) how will the evaluation strategy need to be adjusted in order to take these changes into account? A key example of this process involves translation of interventions across racial-ethnic groups.
This paper focuses on the translation of an evidence-based preventative intervention, School-to-Jobs (STJ), from use with African American youths to a new population, Mexican American youths. School-to-Jobs (STJ) is a brief (12 session), in-school intervention that aims to improve academic achievement and prevent school dropout among low-income African American youths living in high-poverty contexts. The program specifically focuses on the difficult transition between middle school and high school when youths are particularly vulnerable to detaching from school, thus the program was designed as a complement to school curricula in the last year of middle school. STJ has been evaluated using a randomized controlled trial at three middle schools in Detroit with a predominantly African American population (n=261). Students randomly assigned to STJ had fewer unexcused absences, higher passing rates on standardized tests, and were significantly less likely to be referred to summer school than students assigned to the control group, both at the end of 8th grade and through 9th grade.
In this presentation, we will share findings from a pilot project to test the effectiveness of STJ with low-income Mexican American youths. Analyses of differences in cultural focus suggest that African American youth and Mexican American youth may differ in relative focus on attaining successes (academic possible selves) vs. utilizing academics to avoid life failures (feared off-track possible selves). Thus, we posit that among Mexican American youth, the active ingredient of the STJ intervention is more likely to focus on school as a way to avoid becoming off-track, failing, and disappointing parents and family. In addition, we will discuss the practicalities of adapting the STJ intervention for this new population and changes in the evaluation strategy employed to assess STJ for Mexican American youth.
UNDERSTANDING MEDIATING FACTORS IN THE DEVELOPMENT OF CULTURALLY SENSITIVE SCHOOL BASED PREVENTIVE INTERVENTIONS FOR LATINO IMMIGRANT STUDENTS.. Elvia Y. Valencia1, Valerie Johnson1, Robert J. Pandina1, Marco Zarate2, Mohammed R. Taheriazar3, 1Rutgers, The State University of New Jersey, Center of Alcohol Studies, Piscataway, NJ United States; 2North Carolina Society of Hispanic Professionals, Apex, NC United States; 3University of North Carolina at Chapel Hill, Chapel Hill, NC United States
Latino immigrant adolescents face unique challenges in adapting to a new country, culture, language and educational system. These unique challenges might impact each child differently depending upon their gender, country of orignin, length of tenure in the United States, degree of parental involvement and perceived parental support, as well as other mediating factors. It is important to better understand the risk and protective factors associated with these adolescents school-engagement and how these protections might affect their academic aspirations. In order to develop more culturally sensitive programs that meet the unique needs of this population, it is desirable to enhance not only their levels of academic achievement but their levels of psychological functioning and school engagement.
As part of this research study, 275 students participating in the North Carolina Society of Hispanic Professionals (NCSHP) 2004 Education Summit completed an open ended questionnaire addressing questions of their school engagement, perceived parental involvement, academic goals, and perceptions about why students might drop out of school. Open ended measures were categorized using Grounded Theory and Constant Comparisons methods (Strauss and Corbin 1990) from the data obtained in the 2003 NCSHP Educational Summit. These categories were used to create a coding template for quantitative analysis of the 2004 Summit measures.
Subjects were 53% female and 43% male, junior high and high school students, the highest proportion (68%) being of Mexican decent. Analyses revealed that 69% have been in the US for fewer than 5 years.
Questions addressed in this study include: How does acculturation level, country of origin, age, and gender influence perceptions including educational needs, reasons for school drop out, and perceived parental support. How do age, gender, acculturation level, and perceived parental involvement affect school engagement and academic aspirations? The associations among these factors are explored and discussed. In adition, implications these findings hold for planning prevention activities for Latino junior high and high school students will be discussed.
Strauss, A. and J. Corbin (1990). Basics of qualitative research: Grounded theory procedures and techniques. Newbury Park, CA, Sage Publications.
EARLY SCHOOL ADJUSTMENT AMONG HIGH RISK MINORITY YOUTH. Laura Zionts1, 1Kent State University, Kent, OH United States
African American students are identified for special education services in the category of “emotional disturbance” at a higher rate than expected (Harry, 1994). It appears that students who are labeled as “emotionally disturbed” in the special education system are identified for services later than peers within other disability categories. However, there is evidence to suggest that emotional and behavioral disorders can be identified early in a child´s behavioral trajectory (Hinshaw, et al., 1993; Walker, Colvin, & Ramsey, 1995) and intervened upon early in a child´s development with reasonable success and cost (Forness, Kavale, MacMillan, Asarnow, & Duncan, 1996; Knitzer, 1996).
Researchers hypothesize that the influence of teacher expectations for normative behavior may contribute to the situation (Horowitz, Bility, Plichta, Leaf, & Haynes, 1998; McLaughlin & Talbert, 1992). It has been suggested that teachers must implement culturally competent approaches to working with diverse students (Harry, 1991; Isaacs-Shockley, Cross, Bazron, Dennis, & Benjamin, 1996). Culturally appropriate approaches can create a situation in which primary prevention efforts are more effective (Comer, 1996; Zins, Coyne, & Ponti, 1988).
This pre-intervention study examined the quality of the student teacher relationship in terms of its potential contribution to the early school adjustment of African American students in grades k-3 in three high poverty, low-performing urban schools. Teachers from 21 classrooms completed the Child Behavior Checklist (Achenbach, 1991) and the Strengths and Difficulties Questionnaire (Goodman, 1997) for all students in their classrooms (N=330). Students were identified by researchers for continued participation in the study by scores in the clinical range for externalizing behaviors on the CBCL-TRF (.65 or higher) or by scores between 9-12 on the prosocial scale of the SDQ. This second phase of data collection included teacher completion of the Student-Teacher Relationship Scale-Short form (Pianta, 1997), reports of reading and math achievement, and student completion of the Psychological Sense of School Membership (Goodenow, 1993) and the Harter Self-Concept Scale (Harter, 1984). Hierarchal Linear Modeling (HLM) was used to disentangle the effects of student, teacher and school level variables to include SES, race (teacher and student), and normative level of aggression on the positive school adjustment of young African American students attending high poverty, urban schools. Comparing students who were rated high in prosocial behaviors with students who demonstrated high levels of externalizing behaviors may assist researchers in elucidating school and teacher-related factors that impact the school trajectories of high risk youth.