Category: Cultural Diversity / Vulnerable Populations

PS5- #C82 - Ethnic Differences in a Social Ecological Model of Adolescent Problem Behavior in a National Sample

Friday, Nov 17
1:30 PM – 2:30 PM
Location: Indigo Ballroom CDGH, Level 2, Indigo Level

Keywords: Adolescents | Substance Abuse | Race / Ethnicity

Socioecological factors have been found to play significant roles in adolescent substance use, delinquency, and mental health outcomes. In general, socioeconomic status (SES) has covaried with race and ethnicity in the United States. Other factors include peer influence which is well established as a salient predictor of adolescent substance and delinquency. However, studies have also shown some ethnic differences suggesting perceived peer influence as more salient among Asian youth compared to Hispanic youth. Comparably, parental monitoring has long been regarded as a protective factor against adolescent problem behavior. Generally, parental monitoring has consistently correlated with lower levels of adolescent substance use behaviors and delinquency; this is true even when co-occurring mental health symptoms are evident. School engagement has been negatively associated with adolescent substance use and delinquency outcomes; as perceptions of school connectedness decreases, adolescent substance and delinquent behaviors increase. Targeting school engagement in different ethnic minority groups, such as Mexican American adolescents, has demonstrated reductions in delinquent behaviors. Similarly, adolescent engagement with social support networks including participation in extracurricular activities decreases risk of substance use, delinquency, suicidal ideation and depression. And although studies examining socioeconomic influences on adolescent substance use, delinquency, and mental health symptoms have yielded mixed results, low SES and early geographic relocation are specific variables that have been linked with poor adolescent mental health outcomes, particularly among African American and Latino youth. We hypothesized that (1) a social ecological structural equation model of problematic behavioral outcomes would comprise five latent variables: SES, Parental Monitoring, School Bonding, Social Engagement and Peer Behavior, (2) our socioecological model would best predict substance use outcomes; (3) peer behavior and social engagement would serve as mediators in the model while the other three variables would be intercorrelated; and (4) pathways in the model would vary by ethnic group. Data was used from the 2014 NHSDUH including 13,600 participants, ages 12 to 17. We tested an SEM model which yielded good fit overall (RMSEA=.022). The CFI (.823) and chi-square ( < .05) for this model were not appropriate measures of fit as the RMSEA was too small and the sample size too large for interpretation. We trimmed the model to achieve the best fit and in the final model, social engagement and peer behavior had significant mediating pathways from socioeconomic stability and parental monitoring, while school bonding had a significant direct path to problematic behavioral outcomes. We compared the six ethnic groups on the final SEM model using a critical ratio for differences approach. We used Bonferroni’s correction (105 comparisons) to set the critical difference at z= +/- 3.48. The comparisons showed variability by ethnicity on each pathway except for the relationship between parental monitoring and peer behavior, which showed no statistically significant variability across ethnic groups. 

David Stewart

Chief Psychologist
Cambridge Health Alliance-Harvard Medical School

Claudine Moise-Campbell

Doctoral Candidate
Seattle Pacific University
Tacoma, Washington

Ashley Estoup

Graduate Student
Seattle Pacific University
Seattle, Washington

Lindsay Moore

Graduate Student
Seattle Pacific University

Elizabeth Lehinger

Graduate Student
Seattle Pacific University
Seattle, Washington

Erin Underbrink

M.S.
Seattle Pacific University
Seattle, Washington