Epi Night School Lesson 13: Bias in Epidemiology

Avoiding the wrong conclusions, even if you tried to do everything right.

Welcome back to Epidemiology Night School! In our last lesson, we untangled the web of confounding factors, those sneaky variables that make relationships between exposure and outcome seem stronger or weaker than they really are. Today, we’re diving into another common challenge in epidemiology: bias. Bias are the systematic errors that can distort study results and lead us astray.

Bias isn’t just about bad intentions (though those can happen too!). More often, it’s about mistakes in study design, data collection, or analysis that create misleading results. Let’s explore how bias creeps into research, the different types of bias, and how epidemiologists work to minimize it.

I guess I don’t lose weight because of my bias for tacos.

What is Bias?

In epidemiology, bias refers to a systematic error that leads to incorrect estimates of the association between an exposure and an outcome. Unlike random errors (which even out if you collect enough data), bias skews your results in a particular direction, and no amount of data can fix it.

Imagine trying to weigh yourself on a scale that’s off by 5 pounds. No matter how many times you step on, the error doesn’t go away—it’s built into the system. That’s bias.