CORE--STATISTICS

点击翻译按钮获取中文摘要

This core will provide statistical advice and recommendations to all projects. It will facilitate the application of several novel statisticalm ethods for combining data across multiple domains, including recursive partitioning implemented as tree-structured regression- grade-of-membership(GoM)models, and person-centered analysis of longitudinal data. Statistical methods for the analysis of twin at a will be an important theme of this core. Strategies for integrating data collected on different time scales - e.g. daily diaries, panel data with a 7-year inter wave window, and retrospective data referring to irregularly spaced(in time) events - will alsobe a central feature of this core. Training work shops, targeted at investigators on all projects and new investigators with projects in the pilot core, will be given twice each year for five consecutive years, beginning in the second year of the project. The workshops will feature the novel statistical methods mentioned above and behavioral genetics models suited to the studyof twin data in the MIDUS population. Workshops on the novel statistical methods will be conducted by Dr. Singer and,for twin studies, by Drs. Robert Krueger and Ron Kessler.