Biostatistical Analysis of Epidemiologic Data I: Logistic Regression
June 12-16, 2017
8:30 a.m. – 12:00 p.m.
Course Number: 140.676.11
Course describes important biostatistical tools followed by a discussion of useful analytic strategies and how they work. A bit of of algebra, 10th grade algebra, and some elementary statistical/mathematical notation provide a clear and extensive chronicle of a number of ways statistics provide a path to decision making. The goal is to identify and understand statistical tools and how they are used to to distill messages from data. This toolbox approach begins with a brief review of statistical estimation and probability distributions. Also included is an introduction to bootstrap methods of statistical estimation. Then, confidence intervals are explored in detail. The analysis of two of the most common and important biostatistical/epidemiological tools, namely 2 by 2 tables and 2 by k tables, follows. The roll of a variety of issues such as confounding variables, interaction, bias and independence, key elements in many statistical applications, are an additional focus of these discussions. Weighted averages are discussed particularly in the context of combining tables and estimates.
Learning Objective: Upon successful completion of this course, students will be able to
- Understand and interpret confidence intervals in a variety of situations
- Understand the application of relative risk and odds ratio statistics particularly applied to 2 by 2 tables and 2 by k tables
- Take advantage of the combining of tables to provide simpler and more precise analyses
- Understand rigorously the biostatistical/epidemiological concepts, precision, accuracy,independence, bias and confounding, in the context of the analysis of tables
- To use extremely simple and comprehensive bootstrap estimation methods
Grading Options: Letter Grade or Pass/Fail
Course Materials: Provided in class