Sources of Bias
Previous chapters have alluded to various types of bias that can occur in comparative studies. These biases result from imperfections in the study design or implementation. In the current chapter, we attempt a more comprehensive summary of the various sources of bias. We have defined a source of bias to be any aspect of a comparative study that results in a discrepancy between the empirical effect generated by the study and the causal effect of interest.
We will consider the real-world conditions that can affect a study and potentially give rise to bias. These sources of bias have been described informally in many epidemiological and social science publications, often in the context of particular studies or areas of research. Our purpose here is to describe and organize these sources under the umbrella of causal theory developed throughout this book. In doing so, we are not attempting an exhaustive compendium or proposing a checklist that can be used to validate a study’s methodology in a mechanical fashion. Rather, we attempt to provide a fairly high-level overview of the landscape of bias, focusing primary attention on the most important sources that arise in practice.
Our ability to interpret an observed effect in causal terms depends on being able to rule out competing theories. In practice, this entails making a convincing case that various possible sources of bias are unlikely to account for the observed effect. This idea was elaborated in Section ...