Understanding the Causal Logic of Confounds
York Hagmayer, Björn Meder, Michael R. Waldmann · eScholarship (California Digital Library) · 2006
The detection of causal relations is often complicated by confounding variables.Handbooks on methodology therefore suggest experimental manipulations of the independent variable combined with randomization as the principal method of dealing with this problem.Recently, progress has been made within the literature on causal Bayes nets on the proper analysis of confounds with non-experimental data (Pearl, 2000).The present paper summarizes the causal analysis of two basic types of confounding: common-cause and causal-chain confounding.Two experiments are reported showing that participants understand the causal logic of these two types of confounding.