Resources for Causal Reasoning in Health Services Research

Attributable proportion

a proportion of cases that would not have occurred had the population not been exposed, relative to observed cases in the total population exposed (i.e. counterfactual risk when exposed is compared with observed risk)

Back door criterion

identifies variables blocking all biasing paths that would result in association between X and Y; a set of variables S relative to variables X and Y such that no variable from S is a descendent of X and S blocks all backdoors paths from X to Y

Causal diagram

a directed acyclic graph in which dependencies among connected nodes are linked to the observational data by the Markov assumption, so that the network of dependencies defines the set of the conditional probabilities of variables on their parents; the product of these conditional probabilities is said to be the factorization of the joint probability function of the variables

Causal inference

the thought process to conclude whether a relation of cause to effect exists.


the property of being related in causal fashion


the production of an effect by its cause

Conditional independence

equivalence of the joint distribution of two variables to the product of distributions of each variable, given value of a third variable