EXPLANATORY STYLE, PERCEIVED COMMUNITY DISORGANIZATION, AND DEPRESSIVE SYMPTOMS IN URBAN CHILDREN. Alisha Pollastri1, Ellen O'Donnell1, Esteban Cardemil1, 1Clark University, Worcester, MA United States
Despite considerable advances in our understanding of depression prevention in children (e.g., Beardslee & Gladstone, 2001), the majority of research has focused on Caucasian, middle-class populations. This is especially unfortunate given that the only depression prevention study that was conducted with children from underserved populations showed promising, yet intriguing, results (Cardemil, Reivich, & Seligman, 2002). In particular, there remain considerable gaps in our understanding of the relationship between cognitive and contextual factors specific to children from underserved populations. Further understanding of this relationship could aid in the development of effective depression prevention programs for these children.
Our research begins to explore this issue by examining the development of depressive symptoms in a longitudinal study of 306 low income, urban fifth and sixth graders. The sample was racially/ethnically diverse: only 25% of the sample was Caucasian. The cognitive variable examined was explanatory style, and the contextual variable was the perception of community disorganization (e.g., unsafe neighborhoods or high levels of crime or drugs). We hypothesized that 1) children´s explanatory style would be related to depressive symptoms and to their perception of community disorganization; 2) the relationship between community disorganization and depression would be mediated by explanatory style; and 3) explanatory style would predict the change in depressive symptoms and perceptions of community disorganization over time.
Analyses of Time 1 data supported our first hypothesis. Depressive symptoms were positively correlated with perceptions of community disorganization (r=0.37, p<0.01), depressive symptoms were negatively correlated with explanatory style (r=-0.53, p<0.01), and explanatory style was negatively correlated with perceptions of community disorganization (r=-0.21, p<0.01). Further analyses indicated that, contrary to our second hypothesis, the relationship between community disorganization and depression was not mediated by explanatory style. However, a series of regression analyses provided support for our third hypothesis. Explanatory style was a significant predictor of the change in depressive symptoms from Time 1 to Time 2 (F[1,219]=7.38, p<0.01) and also a significant predictor of the change in perceptions of community disorganization from Time 1 to Time 2 (F[1,164]=7.60, p<0.01).
These findings indicate that the perception of contextual factors like community disorganization in urban populations can be influenced by cognitive factors in children. Implications for future exploratory and prevention research will be discussed.
FRIENDSHIP, INTERACTION FREQUENCY AND SHARED GROUP MEMBERSHIP AS DETERMINANTS OF PEER SIMILARITY AND INFLUENCE. Scott Gest1, Kelly Rulison1, Janet Welsh1, 1Pennsylvania State University, University Park, PA United States
We examined the significance of three types of classroom social ties that emphasize different peer influence processes. Mutual friendships highlight the role of mutual-liking, frequent dyadic interactions emphasize social learning processes, and peer groups suggest indirect paths of influence. We expected that classmates sharing multiple processes of influence (friendship, frequent interaction, and shared group membership) would be the most salient. 412 students attending Grades 3 to 5 in a semi-rural community in the northeastern U.S participated in a 1-year study. Teacher ratings and student surveys were obtained in October and May of one school year and October of the following school year (86% with data at T3). Teachers rated academic skills, aggression and prosocial behavior. Students provided self-reported friendships, reports of peer groups in the classroom, and peer-nominated social preference, aggression and academic skills. Cairns´ SCM procedure was applied to reports of peer groups to identify frequently-nominated dyads (via binomial z-tests) and groups (via factor analysis). Children typically had 3 or 4 mutual friends (M = 3.75, SD = 2.1), 3 peer-nominated dyadic interaction partners (M = 3.2, SD = 2.1) and 5 or 6 members of their peer group (M = 5.4, SD = 2.4). Peer affiliations were highly selective: children had no social ties with 50% of same-sex classmates and had all three types of social ties with only 20% of same-sex classmates. Particular patterns of social ties (e.g., friends but not in same group) were stable across the school year, indicating that non-overlap across methods was not attributable to measurement error. Peer environment scores, based on the average score of a child´s peers, were calculated separately for subsets of peers identified by each method (friendship, dyadic interaction, peer group) and for subsets of peers with particular patterns of social ties (e.g., friends but not in same group). Results indicated that peer similarity (i.e., similarity between individuals and their peers) was almost entirely restricted to the 20% subset of same-sex peers identified by all 3 methods. Regression analyses predicting T3 individual behavior from T1 individual behavior and T1 peer environment scores revealed the same pattern: influence was only evident from the subset of peers identified by all 3 methods. Results converged on the conclusion that in terms of similarity and influence, the most salient classroom relationships were those characterized by mutual friendship, frequent interaction and shared group membership. Implications for conceptualizing and measuring peer influence processes will be discussed.
THE IMPACT OF PARENTAL DEPRESSION ON PARENTS´ VIEWS OF CHILD PSYCHOPATHOLOGY. Erin Kelley1, 1George Washington University, Washington, DC United States
Research indicates that depressed mothers tend to over-estimate the incidence and severity of their children´s internalizing and externalizing problems in comparison with their children´s self-reports (Renouf & Kovacs, 1994) and the reports of others (Fergusson, Horwood, Gretton, & Shannon, 1985). This discrepancy has been termed negative attribution bias, and is consistent with Beck´s (1979) cognitive-behavioral conceptualization of depression as causing perceptual distortions and emphasizing the roles of negative cognitions and judgments. While findings have indicated that paternal psychopathology is a risk factor for child externalizing disorders (Connell & Goodman, 2001), and that maternal and paternal depression can interact with or have an additive effect in predicting youth psychopathology (Brennan, Hammen, Katz & Le Broque, 2002), there is a dearth of research investigating the link between paternal psychopathology and fathers´ perceptions of their children´s behavior.
The current study examines discrepancies between child internalizing and externalizing symptoms reported by mothers, fathers, children, and teachers to determine whether negative attribution bias is evident. Data utilized are drawn from a larger study of child coping in the context of parental job loss. Sampling was conducted from the population of southern Maryland families who had recently sought unemployment insurance. The sample is ethnically diverse (49% African American, 43% Caucasian). The children (N=203; 46% male) range in age from 9 to 14 years. Current analyses utilized the following questionnaire data: parent and teacher responses on the Child Behavior Checklist, parent responses on the Center for Epidemiological Studies Depression Scale, and children´s responses on the Baltimore Conduct Problems and Delinquency Scale, the Revised Children´s Manifest Anxiety Scale, and either the child or adolescent version of the Reynold´s Depression Scale.
Preliminary results indicate that parental depression predicts negative attribution bias relative to teacher report for child internalizing and externalizing behaviors, and that statistically significant differences exist between mean levels of discrepancy for depressed and non-depressed fathers. Additional analyses will determine whether negative attribution bias manifests when comparing parent perceptions to child self-report data and whether parent and child gender interact in predicting level of negative attribution bias. Future directions for research will be considered, including the potential impact of parental negative attribution bias on parenting behaviors, child coping, and frequency of child internalizing and externalizing behaviors.
POSITIVE EVENTS: MODERATING THE RELATIONSHIP BETWEEN STRESS AND SYMPTOMS IN CHILDREN. Cynthia Rohrbeck1, Tim Ayers2, Kathy Wilcox Doyle3, Irwin Sandler2, 1George Washington University, Washington, DC United States; 2Arizona State University, Tempe, AZ United States; 3Psychological Counseling Services, Ltd., Scottsdale, AZ United States
This study extends the life events research by examining the effects of positive events and experiences, as well as negative events, on child adaptation to stress. Prior research studies have found inverse relationships between negative life events and health and mental health outcomes. Research studies have also found direct, though positive, relationships between positive life events and adjustment and health outcomes. In addition to such direct effects for life events, there is evidence from the child and adult literature that positive events may also have a stress buffering effect, such that report of more positive events will weaken the negative event - negative outcome relationship. For example, Doyle, et al.(2003) found that positive life events buffered the effects of stress on outcomes for children in families experiencing parental divorce and parental repartnering. The current study examines whether the same mitigating effect is found for families experiencing a different family event – parental job loss. Analyses will consider both the respondant and the type of stress (contextual/interview versus self-report checklist).
A measure of positive events was created for this study by selecting items that measured positive experiences in family, school, and peers domains. Items were drawn primarily from the Children´s Uplifts Scale (Kanner et al, 1987), and the Life Events Checklist (Jackson & Warren, 2000). The final measure consisted of 18 items, e.g., “Your teacher told you that you did a good job” and “You made it onto an athletic team”. The response format measured how frequently the event had occurred in the last year. The positive life events measure was completed by both a parent and the child during the first wave of data collection in a longitudinal study examining the effects of stress on child outcomes in families experiencing parental job loss. Stress was measured by both a life events checklist and by a contextual measure/interview of stressors within the last year. Participants included 203 families sampled from unemployment databases in the state of Maryland. Approximately half were single parent families and ethnicity was approximately half Caucasian and half African American). Children in those families ranged from 8 to 15 years of age, and included 88 males, and 115 females. Assuming that our buffering hypothesis is supported, this study will add another potential target of stress and coping interventions, that is, to increase the number of positive life experiences and events for children and families experiencing stress.
PEER SOCIAL NETWORKS AND CIGARETTE SMOKING IN MIDDLE CHILDHOOD. Patricia Aloise-Young1, Randall Swaim1, 1Colorado State University, Fort Collins, CO United States
In this study, the development of tobacco-using peer groups is analyzed through the use of network analysis. Despite the widely recognized importance of peer influence for adolescent tobacco use, it has rarely been examined with network analysis. However, Ennett & Bauman (1993, 1994) used social network analysis to uncover evidence for both influence and selection in cigarette smoking among 8th and 9th graders. The present study builds upon Ennett and Bauman by examining the roles of perceived harm and school attitudes/achievement in cigarette onset, in addition to exposure to smoking peers in the peer network.
Participants were 161 children enrolled in an ethnically diverse elementary school. The children were in the 4th, 5th and 6th grade in the first year of this three-year longitudinal project. During each year of the project, students were administered two surveys. The first asked about cigarette and alcohol use and a variety of risk factors, including liking for school (4 questions), school achievement (2 questions) and perceived harm (2 questions). The second survey asked children to nominate five of their friends and to indicate how often and where they spend time together.
An adjacency matrix was created showing which students listed each other as friends at time 1. The matrix was analyzed using UCINET (Borgatti, Everett, & Freeman, 1992). Cliques were identified and distances between members of the peer network were computed.
At time 1, there were 10 children who indicated that they had ever smoked. Eight of these children were represented in the social network. The remaining two children were not nominated by any of their peers. There were 74 cliques (ranging from 3 to 6 members) identified, including 11 containing one or more smokers. We computed the distance between each of the nonsmokers and the 8 smokers in the social network (e.g., if John nominated Sue and Sue nominated a smoker then John´s distance = 2 and Sue´s = 1). Of the 151 nonsmokers, 93 were connected to at least 1 smoker.
We conducted a logistic regression with onset of smoking (at time 2 or time 3) as the dependent variable. Predictors in the model were: average distance to smokers, perceived harm, and school attachment (school liking and grades combined). The results revealed that the relation between smoking onset and the average distance to smokers in the social network was negative and marginally significant, p = .08. However, the relation between smoking onset and school attitudes was much stronger p = .009. These results suggest that in middle childhood peer influence may not play as great a role in cigarette smoking as it does in adolescence. Analyses of time 2 and time 3 data are forthcoming and should provide further insight.
INTERVENTION EFFECTS ON PARENT AFFECTIVE QUALITY AND ADOLESCENT PROBLEM SOLVING: A TEST AND REPLICATION. Catherine Lillehoj1, Richard Spoth1, Max Guyll1, 1Partnerships in Prevention Science Institute at Iowa State University, Ames, IA United States
This poster will report two studies that test a model created by combining theoretical predictions regarding how an intervention and parent-child affective quality are both expected to influence adolescent problem solving. The Family Competency Training Outcome Model (FCTOM; e.g., Spoth, Redmond, & Shin, 1998) predicts that a family-focused preventive intervention will positively affect both parent affective quality and adolescent problem solving. A consensus exists among parenting theorists that warm, supportive and responsive parenting behaviors are associated with increased child skills, such as problem solving (Peterson & Hann, 1999). Thus, the current study tested a mediation model that included direct effects from the intervention to both parent affective quality and adolescent problem solving, as well as a direct effect of parent affective quality on adolescent problem solving. The primary hypothesis of this investigation is that parent affective quality will mediate a portion of the intervention's effect on adolescent problem solving, as evidenced by a significant indirect effect of the intervention on adolescent problem solving. Analyses were based on longitudinal data from two studies. The two samples consist of 445 families (232 girls, 213 boys) and 372 families (183 girls, 189 boys). Results of structural equation modeling analyses provided no support for the hypothesized relationship, in that in neither sample did the intervention evidence an indirect effect on adolescent problem solving via parent affective quality. Findings did provide support for intervention effects; the intervention influenced both fathers' affective quality and adolescent boys´ problem solving skill in the larger sample. In the smaller sample, fathers' affective quality predicted girls´ problem solving, thereby providing a measure of support for the idea that supportive parenting increases child skills. The discussion focuses on adolescent gender differences related to the impact of parent affective quality on adolescent problem solving and response to the family-focused intervention.
CHILDREN´S COPING AS A MEDIATOR OF CUMULATIVE CONTEXTUAL THREAT & SYMPTOMS. Tim Ayers1, Karin Back1, Irwin Sandler1, 1ASU Prevention Research Center, Tempe, AZ United States
Numerous studies have found that active coping relates to better adaptation in children and adolescents, both concurrently and prospectively and that avoidant coping typically relates to higher maladjustment and poorer adaptation. However, two important issues have been suggested by authors that deserve further clarification regarding the direct and mediational relationships between these variables, stressful events and children´s adaptation. First, some authors have made an important distinction between avoidant-focused strategies and a related form of coping referred to as distraction-focused strategies. Indeed, some investigators have suggested that these distraction forms of coping may have different relationships to outcomes relative to avoidant strategies. A couple of studies have found that distraction strategies are actually prospective predictors of lower anxiety or other internalizing symptoms. Second, authors have noted that the mediational role of active, avoidant, and distraction-focused coping between stress and adaptation may differ as a function of whether the events are controllable or uncontrollable by the child. For example, in controllable situations, active coping efforts appear to be adaptive; yet, in uncontrollable situations persistent efforts to use active strategies may exacerbate the symptoms for the child, whereas avoidant or distraction strategies of coping in these situations may lead to better adaptation.
This poster will present findings from a longitudinal study of the direct and mediational roles that active, avoidant and distraction forms of coping have between both a checklist and a contextualized measure of stress and children´s internalizing and externalizing symptomatology. Findings about how controllability of the event moderates this relationship will also be presented. Participant were 203 families in which a parent recently experienced a parental job loss. Children ranged in age from 8-15 years old. Events that occurred in the previous twelve months were assessed and will be objectively rated as to their controllability. Results of this study will be discussed in terms of their implications for designing coping interventions for children and families.
NEIGHBORHOOD, FAMILY, AND PEER RISK FACTORS AND EARLY-STARTING ANTISOCIAL TRAJECTORIES. Erin Ingoldsby1, Daniel S. Shaw2, 1University of Utah, Salt Lake City, UT United States; 2University of Pittsburgh, Pittsburgh, PA United States
It has been argued that there are distinct patterns in the development of antisocial behavior (AB) that vary according to age of onset, gender, and other correlates (e.g., ethnicity), have different antecedents (e.g., coercive parenting, deviant peer influence), and relate to the types and severity of outcomes (e.g., overt versus covert delinquency, Loeber et al., 1998). Because children who engage in high rates of AB are at greatest risk for later delinquency and crime, it is crucial to identify key constellations of early risk factors. While family and peer factors have received attention, less research has examined neighborhood risk as a potential differential risk factor. The current study examined neighborhood context (i.e., disadvantage, low cohesion), deviant peer relationships, family dysfunction (i.e., family conflict, poor parental supervision, low parental involvement) and the growth of overt and covert AB in 218 low-income European American (EA) and African American (AA) boys followed from ages 2 to 10. Data were collected from mothers and sons using multiple methods at ages 2, 3.5, 5, 6, 8, and 10. Maternal ratings of five overt items (ages 2 to 10) and five covert items (ages 5 to 10) from the CBCL were submitted to semi-parametric group mixture modeling (Nagin, 1999). For both overt and covert behavior, four groups emerged, with most boys showing low rates of AB over time. However, approximately 8% of boys demonstrated a high/persistent pattern of overt, and 5% of boys showed increasing covert behavior over time, consistent with existing models of early-starting AB. Individuals were assigned to groups according to their posterior probability scores. ANCOVAs were computed to test the hypotheses that (1) boys in the early-starting covert problem group would be characterized by disadvantaged neighborhoods, exposure to deviant peers in the neighborhood, and low parental involvement and supervision, and (2) boys in the early-starting overt group would be characterized by high family conflict. Neighborhood disadvantage, exposure to neighborhood-based deviant peers, and low parental involvement were more consistently characteristic of boys with early-starting covert behavior than early-starting overt behavior. Family conflict and low parental supervision was significantly related to both overt and covert group membership, with higher conflict in the early-starting groups. Differences by ethnicity also emerged. Results suggest that neighborhood disadvantage and neighborhood peer deviancy might be more salient and discriminatory early risk factors for the development of covert problems.
AN ARTS-BASED APPROACH TO ENDING DATING VIOLENCE AMONG URBAN, MIDDLE SCHOOL YOUTH. Tania Araya1, Phyllis Sharps2, Jacqueline Campbell2, Heidi Lary2, Gene Shelley3, Kendall Cephas3, 1House of Ruth Maryland, Baltimore, MD United States; 2Johns Hopkins University, Baltimore, MD United States; 3Centers for Disease Control and Prevention (CDC), Atlanta, GA United States
Project Description Funded by the CDC, this five-year demonstration project was designed to evaluate a school-based program aimed at promoting healthy relationships and preventing intimate partner violence among predominantly African American youth in four Baltimore City middle schools. The overall purpose is to test the hypothesis that intervention schools will have significantly better scores regarding student attitudes and experiences of violence measures post-intervention than comparison schools, and likewise, that school personnel in the intervention schools will have better scores on the attitudes and knowledge survey post intervention than comparison schools. The project is a collaborative effort between the Baltimore City Public School System, the Johns Hopkins University School of Nursing, the George Washington University School of Public Health and several community-based agencies, including a domestic violence shelter. Intervention Design Open to all 7th graders, the intervention program includes four components: 1) Student participatory Afrocentric theatre project, a visual arts project, and a web page design project with dating violence prevention content, 2) Dating violence prevention student discussion groups 3) 4 curricular sessions for all 7th graders and 4) teacher and staff training. Results Our system change based evaluation strategy uses baseline data, process variables, and outcomes contrasting two intervention schools with two comparison schools in the fourth year, with the comparison schools becoming intervention schools in this fifth and final year. Collected baseline data includes: level of dating violence and violence related attitudes (7th grade student survey and student focus groups), observational data (student behavior, protocols and policies), teacher knowledge & attitudes (teacher focus groups and survey). Preliminary analysis demonstrates: 1) 12 – 30% of urban 7th graders report participation in dating violence (physical) and 68-75% exposed to violence at home, 2) 25 – 45% of teachers witnessed student violence within last week, 3) some student gender differences in attitudes but none in violent experiences (perpetration or victimization). Summary Findings will be described in terms of student and faculty attitudes and behaviors related to dating violence, challenges and strategies to working with large, complex institutions, and the strengths inherent in creating a collaboration between community-based agencies and educational institutions. Most importantly, we will discuss how these findings suggest ways that may be most useful in addressing and preventing dating violence among urban middle school youth for future social work practice.