Join Andrew Olshan, for SERforum Live!

“Career and Professional Development”

JOIN SERforum – Live
December 5th, 12pm EST


Do you ever find yourself struggling to figure out a question about epidemiologic methods, or other topics in epidemiology, and don’t know who to ask? The SERforum allows for individuals to answer questions that come up in our daily work around substantive and methodological topics in epidemiology.
All topics may be viewed, but to read and post comments, SER membership is required. If you are member, login! Not a member, join us!

You need to log in to create posts and topics.

Noncollapsibility definition of confounding

Dear All,

To make it clear at the beginning: I prefer to use DAGs for thinking about confounding nowadays.

In his great paper entitled "Commonalities in the classical, collapsibility and counterfactual concepts of confounding" (S.C. Newman, J Clin Epi 2004;57:325-3299), Newman states:

"According to the collapsibility definition of confounding, F is a confounder of a given measure of effect if and only if both of the following conditions are satisfied: (1) the measure of effect is homogeneous across the strate of F, and (2) the crude and common stratum-specific values of the measure of effect are equal".

Newman quotes:

Kupper et al. 1981

Yanagawa, 1984

Boivin & Wacholder, 1985

From these papers, I cannot see why homogeneity across strata is a necessary factor for the definition of a confounder.

If according to this Definition, a factor F is not a confounder, what is then?

i) neither a confounder nor a effect measure modificator

ii) an effect measure modificator

Maybe, Newman actually meant:

Given that homogeneity across strata of F is present: if noncollapsibility is present, F is a confounder. If the analysis is collapsible, confounding is absent. Isn't that a better Definition here?

Best wishes

Andreas Stang