Estimation of causal effects in observational studies with interference between units

Lundin, Mathias; Karlsson, Maria

doi:10.1007/s10260-014-0257-8

Cited by 6 publications

(5 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Early work on relaxing assumption (i) considered blocks of non-overlapping units, where assumptions (i) and (ii) held across blocks, but not necessarily within blocks. This setting is known as partial interference (Sobel, 2006;Hong and Raudenbush, 2006;Hudgens and Halloran, 2008;Liu and Hudgens, 2014;Lundin and Karlsson, 2014;Ferracci et al, 2014).…”

Section: Introductionmentioning

confidence: 99%

Auto-G-Computation of Causal Effects on a Network

Tchetgen

Fulcher

Shpitser

2020

Journal of the American Statistical Association

View full text Add to dashboard Cite

Methods for inferring average causal effects have traditionally relied on two key assumptions:(i) the intervention received by one unit cannot causally influence the outcome of another; and (ii) units can be organized into non-overlapping groups such that outcomes of units in separate groups are independent. In this paper, we develop new statistical methods for causal inference based on a single realization of a network of connected units for which neither assumption (i) nor (ii) holds. The proposed approach allows both for arbitrary forms of interference, whereby the outcome of a unit may depend on interventions received by other units with whom a network path through connected units exists; and long range dependence, whereby outcomes for any two units likewise connected by a path in the network may be dependent. Under network versions of consistency and no unobserved confounding, inference is made tractable by an assumption that the networks outcome, treatment and covariate vectors are a single realization of a certain chain graph model. This assumption allows inferences about various network causal effects via the auto-g-computation algorithm, a network generalization of Robins' well-known g-computation algorithm previously described for causal inference under assumptions (i) and (ii).

show abstract

Section: Introductionmentioning

confidence: 99%

Auto-G-Computation of Causal Effects on a Network

Tchetgen

Fulcher

Shpitser

2020

Journal of the American Statistical Association

View full text Add to dashboard Cite

show abstract

“…Tchetgen and VanderWeele (2012) also use group-level propensity scores to propose inverse probability weighted (IPW) estimators of the direct, indirect, total and overall causal effects in observational studies in the absence of two-stage randomization. Lundin and Karlsson (2014) develop similar IPW estimators of direct, indirect, and total effects under interference where treatment assignment is randomized only at the first stage, while in some groups, all units are untreated. Perez-Heydrich, Hudgens, Halloran, Clemens, Ali, andEmch (2014), andBecker-Dreps (2016) consider the asymptotic properties of different IPW estimators in the presence of partial interference.…”

Section: Partially Randomized and Observational Studiesmentioning

confidence: 99%

Policy Evaluation Using Causal Inference Methods

Fougère

Jacquemet

2020

SSRN Journal

View full text Add to dashboard Cite

“…These IPW estimators can be viewed as a generalization of the usual IPW estimator of the causal effect of a treatment in the absence of interference. Lundin and Karlsson (2014) proposed similar IPW estimators of direct, indirect, and total effects under interference where treatment assignment is randomized only at the first stage, and in some groups, all individuals remain untreated (Figure 1). They were motivated by the Positive Parenting Program, an Australian parenting support program tested in pre-schools in Uppsala, Sweden.…”

Section: Partial Interferencementioning

confidence: 99%

Dependent Happenings: a Recent Methodological Review

Halloran

Hudgens

2016

Curr Epidemiol Rep

View full text Add to dashboard Cite

One hundred years ago Sir Ronald Ross published his treatise on a general Theory of Happenings. Dependent happenings are those in which the frequency depends on the number already affected. When there is dependency of events, interventions can have different types of effects. Interventions such as vaccination can have direct protective effects for the person receiving the treatment, as well as indirect/spillover effects for others in the population. Causal inference is a framework for carefully defining the causal effect of a treatment, exposure, or policy, and then determining conditions under which such effects can be estimated from the observed data. We consider here scenarios in which the potential outcomes of an individual can depend on the treatment of other individuals in the population, known as causal inference with interference. Much of the research so far has assumed the population is divided into groups or clusters, and individuals can interfere with others within their clusters but not across clusters. Recent developments have assumed more general forms of interference. We review some of the different types of effects that have been defined for dependent happenings, particularly using the methods of causal inference with interference. Many of the methods are applicable across disciplines, such as infectious diseases, social sciences, and economics.

show abstract

Estimation of causal effects in observational studies with interference between units

Cited by 6 publications

References 26 publications

Auto-G-Computation of Causal Effects on a Network

Auto-G-Computation of Causal Effects on a Network

Policy Evaluation Using Causal Inference Methods

Dependent Happenings: a Recent Methodological Review

Contact Info

Product

Resources

About