Confounding obscures our view, effect modification is part of reality

In 1983, one of us (AK) attended a course given by one of the world's most prominent scholars in epidemiological methodology, Olli Miettinen [1], and heard him say that in the face of the Lord there is no confounding, but there is effect modification. This may well be the clearest available definition of the difference between two concepts which are still too often confused and not appropriately distinguished, theoretically, and in data-analytic practice dealing with ‘covariables.’

Source: Journal of Clinical Epidemiology - September 28, 2019 Category: Epidemiology Authors: J. Andr é Knottnerus, Peter Tugwell Tags: Editorial Source Type: research