Causal diagrams for interference

By Elizabeth L. Ogburn, Tyler J. VanderWeele. 2014

169 citations recorded by the source corpus.

The term “interference” has been used to describe any setting in which one subject’s exposure may affect another subject’s outcome. We use causal diagrams to distinguish among three causal mechanisms that give rise to interference. The first causal mechanism by which interference can operate is a direct causal effect of one individual’s treatment on another individual’s outcome; we call this direct interference. Interference by contagion is present when one individual’s outcome may affect the outcomes of other individuals with whom he comes into contact. Then giving treatment to the first individual could have an indirect effect on others through the treated individual’s outcome. The third pathway by which interference may operate is allocational interference. Treatment in this case allocates individuals to groups; through interactions within a group, individuals may affect one another’s outcomes in any number of ways. In many settings, more than one type of interference will be present simultaneously. The causal effects of interest differ according to which types of interference are present, as do the conditions under which causal effects are identifiable. Using causal diagrams for interference, we describe these differences, give criteria for the identification of important causal effects, and discuss applications to infectious diseases.

Subfields: Artificial Intelligence, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions (2018, background): “As such, causal methods can be used to study interference [97, 87] and heterogeneous effects [45].”
Identification and Estimation of Spillover Effects in Randomized Experiments (2017, methodology): “Using the terminology of Ogburn and VanderWeele (2014), this phenomenon is known as direct interference.”
A Survey on Causal Inference (2020, background): “outcomes could be inferred using treatment assignments and confounder representations. The dependence in data often leads to interference because some subjects’ treatments can affect others’ outcomes [54, 88]. This difficulty can impede the identification of causal parameters of interest. Extensive ”
Handbook of Big Data (2016, methodology): “To deal with the problem of dependent observations in the estimation of peer effects, Ogburn and VanderWeele (2014b) proposed analyzing a subsample of alter-ego pairs that are mutually independent conditional on carefully chosen observations on intermediate nodes in the network. Testing and estimati”
Making Decisions that Reduce Discriminatory Impacts (2019, unknown): “Because interventions applied to one individual i often affect other individual j, we consider a generalization of SCMs called interference models (Sobel, 2006; E. L. Ogburn, 2014).”
Estimating peer effects in networks with peer encouragement designs (2016, background): “(20) addresses the negative consequences for identification of endogeneous timing of the mediating behaviors.”
Generalizing causal inferences from randomized trials: counterfactual and graphical identification (2019, background): “All three consistency conditions entail assumptions of no interference between units [34]; that is to say, that the counterfactual outcomes S, Z, and Y r=1,s=1,z of each individual are not affected by the trial engagement and treatment status of any other individual.”