Marginal and Conditional Confounding Using Logits.

Kristian Bernt Karlson,Frank Popham,Anders Holm

doi:10.1177/0049124121995548

Abstract

This article presents two ways of quantifying confounding using logistic response models for binary outcomes. Drawing on the distinction between marginal and conditional odds ratios in statistics, we define two corresponding measures of confounding (marginal and conditional) that can be recovered from a simple standardization approach. We investigate when marginal and conditional confounding may differ, outline why the method by Karlson, Holm, and Breen recovers conditional confounding under a "no interaction"-assumption, and suggest that researchers may measure marginal confounding by using inverse probability weighting. We provide two empirical examples that illustrate our standardization approach.

Highlights

This article presents two ways of quantifying confounding using logistic response models for binary outcomes
This article argues that sociologists should distinguish between marginal and conditional odds ratios when they analyze the confounding or mediating role of control variables for associations of interest
We show that confounding comes in both marginal and conditional forms, we explain how their interpretation differs, and we demonstrate that the KHB method recovers a particular type of conditional confounding under a no-interaction assumption

Summary

Introduction

This article presents two ways of quantifying confounding using logistic response models for binary outcomes. Drawing on the distinction between marginal and conditional odds ratios in statistics, we define two corresponding measures of confounding (marginal and conditional) that can be recovered from a simple standardization approach. We investigate when marginal and conditional confounding may differ, outline why the method by Karlson, Holm, and Breen recovers conditional confounding under a “no interaction”-assumption, and suggest that researchers may measure marginal confounding by using inverse probability weighting. We provide two empirical examples that illustrate our standardization approach

Methods

Findings

Conclusion