Confounding

Confounding as a Bias in Effect Estimation

A classic discussion of confounding in which explicit reference was made to ‘confounded effects’ is in Chapter 10 of John Stuart Mill's 1843 edition of A System of Logic, Ratiocinative and Inductive. In Chapter 3, Mill lays out the primary issues and acknowledges Francis Bacon as a forerunner in dealing with them. In Chapter 10, Mill lists a requirement for an experiment intended to determine causal relations: ‘None of the circumstances [of the experiment] that we do know shall have effects susceptible of being confounded [italics added] with those of the agents whose properties we wish to study.’

In Mill's time, the word ‘experiment’ referred to an observation in which some circumstances were under the control of the observer, as it still is used in ordinary English, rather than to the notion of a comparative trial. Nonetheless, Mill's requirement suggests that a comparison is to be made between the outcome of our ‘experiment’ (which is, essentially, an uncontrolled trial) and what we would expect the outcome to be if the agents we wish to study had been absent. If the outcome is not as one would expect in the absence of the study agents, then Mill's requirement ensures that the unexpected outcome was not brought about by extraneous ‘circumstances’ (factors). If, however, these circumstances do bring about the unexpected outcome and that outcome is mistakenly attributed to effects of the study agents, then the [Page 228]mistake is one of confounding (or confusion) of the extraneous effects with the agent effects.

Much of the modern literature follows the same informal conceptualization given by Mill. Terminology is now more specific, with ‘treatment’ used to refer to an agent administered by the investigator and ‘exposure’ often used to denote an unmanipulated agent. The chief development beyond Mill is that the expectation for the outcome in the absence of the study exposure is now almost always explicitly derived from observation of a control group that is untreated or unexposed. Confounding typically occurs when natural or social forces or personal preferences affect whether a person ends up in the treated or control group, and these forces or preferences also affect the outcome variable. While such confounding is common in observational studies, it can also occur in randomized experiments when there are systematic improprieties in treatment allocation, administration, and compliance. A further and somewhat controversial point is that confounding (as per Mill's original definition) can also occur in perfectly randomized trials due to random differences between comparison groups; this problem will be discussed further below.

...