SBIR-STTR Award

EIFACS (Expanded Integrated Facility for Analysis of Complex Surveys) will provide epidemiologists at NIAAA and other researchers with the capability to perform valid multivariate statistical analyses on survey data involving complex sampling designs in an easy-to-use windowed PC environment. Variance estimation will utilize Taylor series linearization, Jackknife, and, possiblys bootstrap techniques. Essentially any practical sampling design can be accommodated (up to seven nest variables). Analysis procedures envisioned include Weighted Kappa, Fourfold Tables (including table stratification), discriminant analysis, survival analysis, cluster analysis, and structural equation modeling. The effort capitalizes on the existence of IFACS (Integrated Facility for Analysis of Complex Surveys), which has taken major steps in the ultimate direction planned for EIFACS. Phase I resurts will include algorithmic development of the various new or augmented procedures, along with their variance estimation methods; a Monte Carlo test facility (to verify validity/reliability of the procedures and variance estimation); and improved design of the windowed interface for user navigation through the available input, procedure execution, and output functions. Phase I output will provide a clear definition of all new and m( dified procedures at the algorithmic level for actual implementation into computer code during Phase II. Phase I will also provide a partial demonstration of a procedural augmentation to generate confidence in the ultimate success of the effort.Awardee's statement of the potential commercial applications of the research:The potential market for EIFACS includes epidemiologists, demographers, and other social scientists who analyze complex survey data and need to reach valid inferential conclusions. EIFACS is unique among all complex survey packages in that it provides a comprehensive missing value imputation capability, as well as other pre-analysis data management functions.National Institute on Alcohol Abuse and Alcoholism (NIAAA)

EIFACS will enable epidemiologists and other researchers to perform valid multivariate analyses on complex-sampling-design survey data within a visual user-interface PC environment. The survey data may be in character form or in special package-generated data sets (e.g., SAS) and either locally stored or at a connected workstation or mainframe (remote execution). Procedures include weighted kappa, fourfold tables (fft), stratified fft, discriminant analysis, polytomous logistic regression, structural equations modeling, and proportional hazards survival analysis. Variances are estimated by Taylor series linearization or resampling. All practical sampling design specifications are permitted. Additionally, an interface is provided to SUDAAN (another complex survey package) which enables EIFACS defined data to be visually set up for foreign execution. This effort will expand on the earlier developed IFACS system which embodies the same concept. The research plan includes requirements update, software functional specification, new and augmented procedures development (algorithms, code, alpha-test, Monte Carlo simulation), user-interface improvement, pre-analysis augmentation (transformation/recoding, missing value imputation, etc.), remote execution refinement, detesting, documentation, and final installation at NIAAA.National Institute on Alcohol and Alcoholism (NIAAA)