Data Management, Measurement and Statistical

点击翻译按钮获取中文摘要

The Data Management, Measurement, and Statistics core (DMMS) will focus on four aims: (1) to provide state-of-the-art data management and statistical resources and support across projects; (2) to develop common, psychometrical-ty sound measures to use across projects, notably measures of smoking behaviors and outcomes; (3) to implement cross-project, integrated analyses; and (4) to develop statistical techniques and approaches to address key issues related to the analysis of adolescent smoking data. ( 1) SUPPORT. A computerized database will be maintained containing all longitudinal reseamh data obtained from EMA forms, structured interviews, adolescent and parent questionnaires, schools, and physiological measurements. Rigorous methods for data entry, editing, and updating will be implemented to ensure that the data are clean, consistent, and secure. Detailed documentation of the database will be developed by this core. Statistical support and collaboration for all projects will be provided. (2) MEASUREMENT. The program project aims to increase understanding of the longitudinal patterns of adolescent smoking and the emotional and social contexts in which these occur. A DMMS core aim is to conceptualize and develop measures that will: (a) characterize longitudinal smoking patterns and (b) be tractable and interpretable in the analysis of the contexts in which these patterns develop. (3) INTEGRATION. A key feature of the program project is that data will be available from the main study cohort (seven waves) and the three projects. In the DMMS, development of statistical models that allow integrated analyses of data from many sources will be a major priority. This v, ill be accomplished using a variety of advanced statistical approaches. The goal is to obtain a more complete picture of adolescent smoking than could be obtained from analyzing the data separately. (4) DEVELOPMENT. Methodological research within the DMMS core will seek to advance development of statistical approaches for the analysis of smoking data. These efforts will be focused around the following primary issues: the conceptualization/combination of multiple indices of"dependence," individual heterogeneity in smoking development across time, heterogeneity in the influence of smoking predictors, and characterizing stages of smoking dependence.