Causes and mixed probabilities

Abstract

In Section 1 I examine the use of probabilistic data to establish causal conclusions in non‐experimental research. In Section 2 I show that the probabilities involved in such research are inhomogeneous ‘mixed’ probabilities. Section 3 then argues that such mixed probabilities are responsible for the way common causes screen off correlations between their joint effects. Section 4 concludes that mixed probabilities are therefore crucial for the nature of the causal relation itself.