Adjustment with Three Continuous Variables

TL;DR

This paper evaluates various statistical adjustment techniques for controlling confounding variables in data analysis, emphasizing the importance of selecting methods based on data specifics and illustrating how causal graphs can guide this choice.

Contribution

It demonstrates the use of simulations to compare adjustment techniques and introduces causal graphs as a tool for selecting appropriate methods.

Findings

No single adjustment technique is universally best.

Technique effectiveness depends on data-generating process.

Causal graphs can effectively guide adjustment method selection.

Abstract

Spurious association between X and Y may be due to a confounding variable W. Statisticians may adjust for W using a variety of techniques. This paper presents the results of simulations conducted to assess the performance of those techniques under various, elementary, data-generating processes. The results indicate that no technique is best overall and that specific techniques should be selected based on the particulars of the data-generating process. Here we show how causal graphs can guide the selection or design of techniques for statistical adjustment. R programs are provided for researchers interested in generalization.