Core C -- Statistics Core

摘要-统计核心 该中心的总体目标是为所有研究项目提供方法论和统计支持 U19和附属AD/ADRD U01，并最大限度地促进丰富的MIDUS纵向数据的使用 有可能被调查人员发现。核心活动鼓励开发创新的、综合性的跨项目 研究，建立在MIDUS共同的强大跨学科科学家团队和实质性目标的基础上 项目。核心服务将补充和加强公众使用MIDUS数据，以便它提供 为更大的科学界增加了最大的价值。核心的核心是提供多个 讲习班，邀请专家、项目调查人员和公开数据的用户参加，以审查 采用复杂方法和共享措施的综合途径，旨在充分利用 WAVE、多队列MIDUS调查数据及其新的遗传、生物标记物和神经成像数据。具体来说， 核心将(1)促进最大限度地利用复杂的多面MIDUS纵向数据。统计核心 将举办讲习班，重点讨论四个主题的相关方法和数据分析问题： (A)利用纵向数据处理研究目标的先进统计方法； 分析新的遗传、生物标记物和神经科学数据；(C)各项目共同面临的分析挑战 (例如，缺少数据技术，如多重归罪)；以及(D)实质性和概念性挑战 这统一了项目(例如，概念化、分析和模拟年龄、时期和队列效应)。 选择特定主题是一个持续和反复的过程，旨在满足 MIDUS调查人员和用户社区。(2)促进科学交叉研究。统计核心将促进 并指导与其他主要纵向研究的交叉研究比较。核心领导者将展示 如何进行多项研究比较，并就在这种情况下部署MIDUS的最佳方式向数据用户提供建议 综合分析。我们将借用Meta分析的工具，在多研究框架内使用MIDUS。 我们还将使用随机效应荟萃分析等方法并向数据用户提供建议，这些方法允许 考虑研究水平的变量来解释研究结果之间的差异。(三)提供分析指导 在分析历史语境效应时。统计核心将提供说明性指导和统计 就数据分析进行咨询，利用MIDUS设计提供的机会来审查长期 大衰退的影响，以及最近新冠肺炎大流行对身体的影响， MIDUS参与者的心理、经济和社会福利。其目的是协助调查人员 通过分析比较大衰退和新冠肺炎大流行前后的结果，两者 提供自然的人与人之间变化的实验设计。

ABSTRACT – STATISTICS CORE The Core’s overarching goal is to provide methodological and statistical support for all research projects in the U19 and Adjunct AD/ADRD U01, and facilitate the use of the rich MIDUS longitudinal data to the fullest extent possible by investigators. Core activities encourage the development of innovative, integrative cross-project studies, building on the strong interdisciplinary team of scientists and substantive aims shared by MIDUS projects. Core services will complement and enhance the public use MIDUS data so that it provides the greatest value-added to the larger scientific community. The centerpiece of the Core is the provision of multiple workshops, involving invited experts, project investigators, and users of the publicly available data, to examine integrative pathways using complex methods and shared measures, targeted toward fully exploiting the multi- wave, multi-cohort MIDUS survey data, and its novel genetic, biomarker and neuroimaging data. Specifically, the Core will (1) Promote maximal use of complex multi-faceted MIDUS longitudinal data. The Statistics Core will organize workshops focused on relevant methodological and data analytic issues addressing four themes: (a) advanced statistical methods that exploit the longitudinal data to address study aims; (b) the use and analysis of new genetic, biomarker, and neuroscience data; (c) analytic challenges shared across the projects (e.g., missing data techniques such as multiple imputation); and (d) substantive and conceptual challenges that unify the projects (e.g., conceptualization, analysis, and modeling age-period-and cohort effects). Selection of specific topics is an ongoing and iterative process, designed to meet the evolving needs of the MIDUS investigators and user community. (2) Facilitate cross-study science. The Statistics Core will facilitate and guide cross-study comparisons with other major longitudinal studies. The Core leaders will demonstrate how to carry out multi-study comparisons and advise data users on the best ways to deploy MIDUS in such integrative analyses. We will borrow the tools of meta-analysis to use MIDUS within multi-study frameworks. We will also use and advise data users on methodologies such as random effect meta-analysis, which permit the consideration of study-level variables to explain differences among findings. (3) Provide analytic guidance in analysis of historical context effects. The Stats Core will provide illustrative guidance and statistical consultation on data analyses that exploit opportunities afforded by the MIDUS design to examine long-term impacts of the Great Recession, and more recently, impacts of the COVID-19 pandemic on the physical, psychological, economic, and social well-being of MIDUS participants. The objective is to assist investigators with analyses that compare outcomes before and after the Great Recession and the COVID-19 pandemic, both offering natural experimental designs of within-person change.