Facilitating Secondary Analysis and Archiving of MIDUS through DDI

描述（由申请人提供）：该提案响应RFA-AG-13-009（老龄化社会和行为数据集的二次分析和存档），其目标之一是“存档和传播数据集以实现二次分析”。我们的主要目标是将我们之前的工作扩展到MIDUS（美国中年），这是一项针对大约10,000名美国人的全国性纵向研究，旨在研究衰老作为一个综合的生物、心理和社会过程。我们将采用一种称为数据文档计划（DDI）的技术元数据标准，为所有MIDUS项目构建增强型电子代码本。先前的工作将DDI（2.1）应用于MIDUS代码本，这些代码本被交付给ICPSR，并用于创建可搜索的在线存储库。将所有MIDUS代码本升级到更高级的DDI（3.2）版本，将有助于识别和协调类似的MIDUS变量，并通过数据提取功能增强我们的在线存储库，该功能将实现前所未有的功能：能够获得定制的跨项目下载符合DDI的协调MIDUS数据。这样做将极大地促进公众高效和有效地使用包括MIDUS研究的大型纵向和多学科数据集。DDI基于一种名为可扩展标记语言（eXtensible Markup language， XML）的强大语言，这是一种基于web的数据发布标准，类似于HTML，但功能更强大。XML允许在web上发布的数据具有高度结构化和交互性。它的主要优点是它是机器可读的，并且采用专门设计用于描述数据的格式。最新版本的DDI（3.2）包含近900个新的和修订的标签，旨在提供更有效的纵向和其他复杂分层数据结构文档。该提案的具体目标是：(1)将DDI（3.2）元数据标准应用于所有MIDUS相关数据，升级M1和M2 DDI（2.1）码本，并从最近资助的数据收集工作中创建新的码本（包括Refresher队列（国家和密尔沃基特定样本），第三波原始MIDUS样本（M3），第二波来自日本版MIDUS，称为日本中年（MIDJA））；(2)将Aim #1产生的DDI材料整合到一个可搜索的在线存储库中，研究人员可以从中创建内部协调变量的定制MIDUS数据集。实现这些目标将产生一个DDI存储库，该存储库可以跨不同的MIDUS项目数据集执行变量搜索，识别相关变量及其完整的元数据，支持选定变量的自定义数据提取，并提供包含信息、过程和算法的自定义代码本。这些目标将在两年的时间内完成，按照顺序完成Refresher、M3和MIDJA的数据收集。

DESCRIPTION (provided by applicant): This proposal responds to RFA-AG-13-009 (Secondary Analyses and Archiving of Social and Behavioral Datasets in Aging), one goal of which is "archiving and dissemination of data sets to enable secondary analysis." The primary objective is to extend our prior work in applying new documentation technology to MIDUS (Midlife in the United States), a national longitudinal study of approximately 10,000 Americans designed to study aging as an integrated biopsychosocial process. We will employ a technological metadata standard called the Data Documentation Initiative (DDI) to construct enhanced electronic codebooks for all MIDUS projects. Prior work applied DDI (2.1) to MIDUS codebooks, which were delivered to ICPSR, as well as used to create a searchable online repository. Upgrading all MIDUS codebooks to a more advanced version of DDI (3.2) will facilitate identification and harmonization of similar MIDUS variables, as well as enhance our online repository with a data extract function that will accomplish something unprecedented: the ability to obtain customized cross-project downloads of harmonized MIDUS data that are DDI-compliant. So doing will greatly facilitate efficient and effective public use of the large longitudinal and multi-disciplinary datasets that comprise the MIDUS study. DDI is based on a powerful language called eXtensible Markup Language (XML), a web-based data-publishing standard that is similar to HTML, but more powerful. XML allows data published on the web to be highly structured and interactive. Its main advantage is that it is machine-readable and in a format specifically designed to describe data. The latest version of DDI (3.2) contains nearly 900 new and revised tags and is designed to provide more efficient documentation of longitudinal and other complex hierarchical data structures. The specific aims of this proposal are: (1) to apply the DDI (3.2) metadata standard to all MIDUS-related data, upgrading M1 and M2 DDI (2.1) codebooks, and creating new codebooks from recently funded data collection efforts (including a Refresher cohort (national and Milwaukee-specific samples), a third wave of the original MIDUS sample (M3), and a second wave from the Japanese version of MIDUS called Midlife in Japan (MIDJA)); and (2) to integrate the DDI materials produced by Aim #1 into a searchable online repository from which researchers can create customized MIDUS datasets of internally harmonized variables. Accomplishing these aims will produce a DDI repository that performs variable searches across different MIDUS project datasets, identifies related variables along with their full metadata, enables custom data extracts of selected variables, and provides custom codebooks that contain information, procedures, and algorithms for their harmonization. These goals will be accomplished over a two-year period, following the sequential schedule for completion of data collection for the Refresher, M3 and MIDJA.