Attention:
Graduate students, faculty and postdocs

Initial data analysis, Thursday, April 22, 1-4pm

Statistical model building for prediction in the biological and social sciences inevitably must take into account properties of the collected data that cannot easily be anticipated. Indeed, as this workshop will demonstrate, overlooked unexpected data properties, unsystematic exploratory data analysis and/or lack of transparent reporting may threaten the validity and reproducibility of prediction models. The utility of an initial data analysis plan will be demonstrated in three different data collections: from a diagnostic prediction study of bacteremia with 50 laboratory variables, from a longitudinal, cross-national survey of health and retirement to analyze grip strength, and from a randomized trial in trauma patients. Data summaries and data visualizations are shown in an electronic lab notebook (R markdown), but no prior knowledge of statistical software is required. Registration: https://cstat.msu.edu/event/initial-data-analysis

A 5 minute video on this topic: "Only fools rush in"
https://mediaspace.msu.edu/id/1_1n07g3j0