Confounding and Precision Variables in Linear Regression

Variables can serve different roles in a regression model beyond being the primary explanatory variable of interest. In this lecture we introduce confounders and precision variables.

For confounding, we discuss the general concept before defining two approaches to evaluate for potential confounders: the classical and operational criteria. An example is provided for evaluating a potential confounder.

Precision variables are discussed at the end, with their potential role at changing the variance estimates of our coefficients discussed.

A video for the Biostatistical Methods I (BIOS 6611) course in the Department of Biostatistics and Informatics at the University of Colorado-Anschutz Medical Campus taught by Dr. Alex Kaizer. Slides and additional material available at https://www.alexkaizer.com/bios_6611.

Table of Contents:

00:00 - Intro Song
00:20 - Welcome
00:37 - Confounding
03:00 - Classical Criteria for Confounding
04:03 - Operational Criterion for Confounding
05:35 - Classical and Operational Connection
06:29 - Positive Confounding
07:23 - Negative Confounding
08:10 - Accounting for Confounding
09:11 - Confounding Example
14:32 - Precision Variables