RANDOM COEFFICIENT MODELS OF FAMILY SUBSTANCE USE

家庭物质使用的随机系数模型

• 批准号: 2677889
• 负责人: TERRY E DUNCAN
• 金额: $ 7.65万
• 依托单位: OREGON RESEARCH INSTITUTE
• 依托单位国家: 美国
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-09-01 至 2000-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/2677889
• 关键词: African American; Hispanic Americans; adolescence (12-20); behavior disorders; caucasian American; family structure /dynamics; gender difference; human data; mathematical model; model design /development; parent offspring interaction; psychological models; racial /ethnic difference; sibling relations; statistics /biometry; substance abuse epidemiology; substance abuse related disorder

DESCRIPTION: (Applicant's Abstract) The aim of the proposed small grant project is to examine the hierarchical nature of substance use, accounting for the interdependence of family data, using existing databases and sophisticated analytic techniques. Family level and individual level research questions will be addressed. Models will specify the interdependence of parents and adolescents (siblings and target adolescents) on patterns of drug use and deviant behaviors. The impact of family level variables and other contextual characteristics, on family substance use and deviant behavior will also be studied. Typical analyses of family data ignore the potential interdependence and hierarchical nature of the data. In this proposal, new analytical techniques suited to a hierarchical data structure will be employed. These techniques, labeled hierarchical or multilevel models, take into account correlated observations and observations from heterogeneous populations with varying parameter values that often arise in hierarchical data collections. A great deal of time and money has been expended to obtain excellent data sets relevant to an examination of non random clustering of substance use, but many of these data sets have not been analyzed using stage-of-the-art analytic techniques specifically designed for hierarchical data. The hierarchical nature of the data to be examined will allow us to construct, estimate, and test a variety of complex models concerning substance use at both the family and individual level. Tests of major theoretical relationships using different analytic techniques, and comparisons of findings across datasets is likely to yield important information regarding adolescent substance use and other related problem behaviors that might otherwise go unexplored in these datasets. Examination at multiple levels of the hierarchy is likely to further enhance our understanding of the role of families, and the larger social context, in the development of substance use and related deviant behaviors.

描述：（申请人摘要） 拟议的小额赠款项目的目的是审查等级制度， 物质使用的性质，家庭数据相互依存的原因， 使用现有的数据库和复杂的分析技术。 家庭 水平和个人水平的研究问题将得到解决。 模型 将具体说明父母和青少年（兄弟姐妹和 以青少年为对象）进行关于吸毒模式和异常行为的调查。 的 家庭一级变量和其他背景特征对 还将研究家庭药物使用和异常行为。 典型 对家庭数据的分析忽视了潜在的相互依赖性， 数据的层次性。 在这份报告中，新的分析 将采用适合于分层数据结构的技术。 这些 技术，标记为层次或多级模型，考虑到 相关观测值和来自异质群体的观测值， 在分层数据集合中经常出现的变化参数值。 大量的时间和金钱已经花费在获得优秀的数据 与物质使用非随机聚集检查相关的集合， 但是这些数据集中的许多都没有用最先进的技术进行分析， 专门为分层数据设计的分析技术。 的 要检查的数据的分层性质将允许我们构建， 估计，并测试各种复杂的模型，有关物质的使用， 家庭和个人层面。 主要理论测试 使用不同的分析技术的关系，并比较 跨数据集的发现可能会产生有关以下方面的重要信息 青少年物质使用和其他相关的问题行为， 否则在这些数据集中没有被探索过。 多层次考试 可能会进一步加强我们对角色的理解 家庭和更大的社会背景， 使用和相关的异常行为。