Adolescent Peer Social Network Dynamics and Problem Behavior

描述（由申请人提供）：生态学理论认为，青少年早期的同伴群体经常形成，鼓励问题行为，导致更紧密，倾向于越轨的同伴群体，等等，在一个自我强化的循环中。我们称这个过程为有毒同伴传染。最好将其概念化为非线性动态系统（NDS）。我们建议发展NDS模型来解释中学同伴社会生态中问题行为（反社会行为、攻击、物质使用）的发生和发展，特别是检验有毒同伴传染作为一种解释假设。这项研究将在21所六年级到八年级的中学进行，这些中学已经参与了NIDA资助的积极行为支持（PBS）项目的随机分组试验。一个主要目标是调查学校之间毒性同伴传染的差异，特别是PBS项目和对照学校之间的差异。该项目需要收集多个场合的社会关系（社交网络）和相关行为数据，并观察它们之间随时间的相互依赖模式。每个学年将进行四次评估，连续两次进入六年级，直到21所参与学校的八年级结束；因此，每个队列有12个观察结果。学校级别的模型将使用SIENA创建，SIENA是一种最近开发的统计方法，专门用于估计不断发展的社会网络和共同发展的行为模型。另外，如果各学校的效果差异较大，将开发两级模型。为了实现这些雄心勃勃的目标，有必要对SIENA计算机软件进行几次主要升级，通过将其移植到多处理器计算环境来提高其估计性能，并增强其包括交互效果、全信息多层模型和跨层交互的能力。为此，我们正在与开发SIENA的研究人员合作，为项目目标进行必要的升级。该项目的结果将包括：(a)估计有毒同伴传染在不同中学样本中的流行程度；(b)最可能和最不可能发生这种情况的情况；(c)公共服务计划作为一种阻止这种行为的方法的有效性；(d)对处理这一问题的同伴指导干预的建议。公共卫生相关性：本项目将检验一项已经资助的学校层面积极行为干预对早期青少年问题行为和同伴群体动态之间的长期相互依赖关系的群体随机试验的影响。这个过程，从社会生态学和发展理论发展而来，是有毒的同伴传染。使用SIENA（专门为模拟此类过程而开发的计算机程序），将从纵向（4次/年，6至8年级）数据中统计估计隶属关系-行为相互依赖的非线性动态模型。结果将解释同伴关系在问题行为发展中的作用，积极行为支持对这种联系的影响，并为未来的干预提供新的途径。

DESCRIPTION (provided by applicant): Ecological theory suggests that early adolescent peer groups regularly form which encourage problem behavior, leading to even closer, deviance-prone peer groups, and so on in a self-reinforcing cycle. We call this process toxic peer contagion. It is best conceptualized as a Nonlinear Dynamic System (NDS). We propose to develop NDS models that explain the onset to and growth of problem behaviors (antisocial behavior, aggression, substance use) in middle school peer social ecologies, and particularly to examine toxic peer contagion as an explanatory hypothesis. The study will be conducted in 21 sixth- to eighth-grade middle schools involved in an already underway, NIDA- funded group-randomized trial of the Positive Behavior Support (PBS) program. A major goal is to investigate variation in toxic peer contagion differences between schools, especially between PBS program and control schools. The project requires collecting social relationship (social network) and relevant behavioral data on multiple occasions, and observing the pattern of interdependence between them over time. Assessments will be conducted four times per school year, for two consecutive entering sixth-grade cohorts, through the end of their eighth-grade year in each of the 21 participating schools; thus, 12 observations per cohort. School-level models will be created using SIENA, a recently developed statistical methodology specifically designed to estimate models of evolving social networks and co-evolving behavior. Also, two-level models will be developed when effects vary significantly across schools. In order to accomplish these ambitious goals with this large, diverse sample of schools and students, it is necessary to perform several major upgrades to the SIENA computer software, improving its estimation performance by porting it to multiprocessor computing environments, and enhancing its ability to include interaction effects, full- information multilevel models, and cross-level interactions. For this purpose, we are partnering with the researchers who developed SIENA to make the upgrades necessary for project goals. The results of this project will include: (a) estimates of the prevalence of toxic peer contagion in a varigated sample of middle schools; (b) conditions under which it is most and least likely to occur; (c) the effectiveness of PBS as a way to discourage it; and (d) recommendations for peer-directed interventions for dealing with this problem. PUBLIC HEALTH RELEVANCE: This project will examine the effect of a group-randomized trial of an already funded school-level positive behavior intervention on the over-time interdependence between early adolescent problem behavior and peer group dynamics. This process, developed from theories of social ecologies and development, is toxic peer contagion. Nonlinear dynamic models of affiliation-behavior interdependence will be statistically estimated from longitudinal (4x/yr, 6th to 8th grade) data using SIENA, a computer program specifically developed to model such processes. Results will explain the role of peer affiliations in development of problem behavior, the effect of positive behavior support on this linkage, and suggest new avenues for future interventions.