As I understand it, matching is one way to identify causality in observational studies. By matching observations that are "similar" and comparing ones that did or did not receive treatment, you can consider this as a quasi-experiment of sorts.
What is overmatching? What kind of bias does it introduce? I have mostly seen matching from an economics perspective, but have recently seen some papers on epidemiology suggesting that "overmatching" can result in bias. I find it hard to understand the terminology of the papers and would greatly appreciate if someone could help explain some of the main concepts. Below is an article that references the idea:
Over-matching can cause bias. BMJ. 2002 August 10;
325(7359)
Best Answer
From Modern Epidemiology 3rd Edition by Rothman, Greenland and Lash:
The answer from AndyW is about the second form of overmatching. Briefly, here's how they all work:
1: In order to be a confounder, one of the criteria is that the covariate be associated with both the outcome and the exposure. If it's only associated with one of them, its not a confounder, and all you've succeeded in doing is widening your confidence interval.
2: This is partially discussed by AndyW. Matching on an intermediate factor will bias your estimate, as will matching on something affected by both the exposure and outcome. This is essentially controlling on a collider, and any technique that does so will bias your estimate.
3: This is more of a study design problem. Extensively matching on variables that you needn't match on for reasons 1 & 2 can cause you to reject easily obtained controls (friends, family, nearby social network, etc.) in favor of far harder to obtain controls that can be matched on the unnecessary set of covariates. That costs money - money that could have been spent on more subjects, better exposure or disease ascertainment, etc., for no appreciable gain in bias or precision, and indeed having threatened both.