The Finite Sample Performance of Estimators for Mediation Analysis Under Sequential Conditional Independence

Abstract

Using a comprehensive simulation study based on empirical data, this article investigates the finite sample properties of different classes of parametric and semiparametric estimators of (natural) direct and indirect causal effects used in mediation analysis under sequential conditional independence assumptions. The estimators are based on regression, inverse probability weighting, and combinations thereof. Our simulation design uses a large population of Swiss jobseekers and considers variations of several features of the datagenerating process (DGP) and the implementation of the estimators that are of practical relevance. We find that no estimator performs uniformly best (in terms of root mean squared error) in all simulations. Overall, so-called "g-computation" dominates. However, differences between estimators are often (but not always) minor in the various setups and the relative performance of the methods often (but not always) varies with the features of the DGP.

The Long-Run Effect of Childhood Poverty and The Mediating Role of Education

“…(See Huber et al . () for an analysis of the finite sample properties of different classes of parametric and semiparametric estimators under sequential conditional independence assumptions and Linden and Karlson () for a comparison of the methods applied to the JOBS II data set.) These methods cover semiparametric and non‐parametric estimation procedures (Imai et al ., ; Pearl, , b; Hafeman and Schwartz, ), matching based on the propensity score (Hill et al ., ), weighting procedures (Peterson et al ., ; VanderWeele, ; Hong, ), the principal stratification approach (Frangakis and Rubin, ; Jo, ; Jo et al ., ) and the g ‐computation‐based algorithm (Robins and Greenland, ).…”

Does homeownership affect education outcomes?

“…The statistical tools and techniques required to establish such inferences are available and can be used in developing robust housing policies [6]. Some evidence indicates that important mediators through which housing tenure affects education include stability of tenure and extent of crowding within a household.…”

“…While statistical techniques to identify the direct effect of a treatment such as homeownership net of the indirect effect through mediators are available, they require controlling for the endogeneity of the treatment and mediating variables. And that generally means that additional assumptions are needed about the independence of the potential outcomes and treatment and the independence of potential outcomes and mediators [6].…”

The finite sample performance of semi- and non-parametric estimators for treatment effects and policy evaluation

“…The outcome variable Y in the simulations is defined as the cumulative months an individual 10 Our final sample size differs slightly from theirs because we exclude, following Huber, Lechner, and Mellace (2014a), individuals who were registered in the Italian-speaking part of Switzerland in order to reduce the number of language interaction terms to be included in the model. We also deleted 102 individuals who registered with the employment office before 2003.…”

“…As for Huber, Lechner, and Mellace (2014a), the data used in our EMCS include individuals who registered at Swiss regional employment offices anytime during the year 2003. Detailed jobseeker characteristics are available from the unemployment insurance system and social security records, including gender, mother tongue, qualification, information on registration and deregistration of unemployment, employment history, participation in active labor market programs, and an employability rating by the caseworker in the employment office.…”