Core C: Data Management and Analysis

Project Summary ? Core C The principal objective of the Data Management and Analysis Core (Core C) is to provide project investigators a centralized resource for quantitative expertise. Statistical and methodological issues will be addressed at all stages: from the initial design of laboratory assays and observational studies to the maintenance of data quality, to the analysis of complex data. In support of this objective, Specific Aims of Core C are: Aim 1. Collaborate with project investigators to align study design and analysis with research questions. Core C will assist researchers at the design stage by: formulating biological, immunological, epidemiological, and sociological questions as testable statistical hypotheses; devising efficient study designs and matching these with appropriate statistical models; determining sample sizes necessary to ensure high power while controlling Type I error. Survey development will utilize appropriate methods (e.g., cognitive interviewing or pilot testing) to ensure that they are psychometrically appropriate. Core personnel will ensure that: the chosen designs and surveys avoid biases and decrease measurement errors; the selected primary outcome measures will answer specific research questions; and the statistical analysis plans will use data to answer the research questions under realistic and transparent assumptions. Aim 2. Provide services and support for data collection and management to ensure: integrity, security and accessibility; processing and quality control; sharing across projects; and creating a centralized repository of all survey instruments, analysis procedures and results. To promote data integration across projects, Core C will create a centralized repository of all data that will be secure and accessible to investigators. Quantitative data will be cleaned, merged and de-identified. Missing and outlying data reports will be provided to the investigators. Developed surveys will be content validated and further validated by psychometric analyses as appropriate. Qualitative data will follow a rigorous transcription and coding process. To ensure high data integrity and availability for statistical modeling, Core C will implement rigorous quality control and processing procedures, using approaches such as normalization, filtering, and visualization. Aim 3. Provide services and support for data analysis and interpretation phases of all projects. Core C services will include: (i) formal hypothesis tests for laboratory, epidemiological, and qualitative data that ensure strong conclusions, (ii) exploratory analyses that lead to new or refined hypotheses, (iii) statistical modeling and sensitivity analyses of complex data, and (iv) visual displays that clarify conclusions. In accordance with critical principles of rigor and reproducibility in science, Core C biostatisticians will work closely with investigators to develop comprehensive analysis plans that distinguish between formal testing and exploratory analyses. Qualitative data (e.g., cognitive interviews) will be analyzed following a standardized coding process.