Abstract
Causal effects are usually estimated under the assumption of no interference between individuals. This assumption means that the potential outcomes for one individual are unaffected by the treatments received by other individuals. In many situations, this is not reasonable to assume. Moreover, not taking interference into account could result in misleading conclusions about the effect of a treatment. For two-stage observational studies, where treatment assigment is randomized in the first stage but not in the second stage, we propose IPW estimators of direct and indirect causal effects as defined by Hudgens and Halloran (J Am Stat Assoc 103(482):832–842, 2008) for two-stage randomized studies. We illustrate the use of these estimators in an evaluation study of an implementation of Triple P (a parenting support program) within preschools in Uppsala, Sweden.
Similar content being viewed by others
References
Aronow PM (2012) A general method for detecting interference between units in randomized experiments. Sociol Methods Res 41(1):3–16
Aronow PM, Samii C (2012) Estimating average causal effects under general interference. Working paper
Axberg U, Johansson Hanse J, Broberg A (2008) Parents description of conduct problems in their children-a test of the eyberg child behavior inventory (ECBI) in a swedish sample aged 3–10. Scand J Psychol 49(6):497–505
Bowers J, Fredrickson MM, Panagopoulos C (2013) Reasoning about interference between units: a general framework. Political Anal 21(1):97–124
Cox D (1958) Planning of experiments, volume 208. Wiley, New York
Halloran M, Struchiner C (1991) Study designs for dependent happenings. Epidemiology 2(5):331–338
Halloran ME (2012) The minicommunity design to assess indirect effects of vaccination. Epidemiol Methods 1(1):82–105
Heckman J, Lochner L, Taber C (1999) Human capital formation and general equilibrium treatment effects: a study of tax and tuition policy. Fiscal Stud 20(1):25–40
Hirano K, Imbens G, Ridder G (2003) Efficient estimation of average treatment effects using the estimated propensity score. Econometrica 71(4):1161–1189
Holland P (1986) Statistics and causal inference. J Am Stat Assoc 81(396):945–960
Hong G (2010) Marginal mean weighting through stratification: adjustment for selection bias in multilevel data. J Educ Behav Stat 35(5):499–531
Hong G, Raudenbush SW (2006) Evaluating kindergarten retention policy. J Am Stat Assoc 101(475):901–910
Hong G, Raudenbush SW (2008) Causal inference for time-varying instructional treatments. J Educ Behav Stat 33(3):333–362
Hudgens M, Halloran M (2008) Toward causal inference with interference. J Am Stat Assoc 103(482):832–842
Ichino N, Schündeln M (2012) Deterring or displacing electoral irregularities? Spillover effects of observers in a randomized field experiment in Ghana. J Politics 74(1):292–307
Kang JD, Schafer JL (2007) Demystifying double robustness: a comparison of alternative strategies for estimating a population mean from incomplete data. Stat Sci 22(4):523–539
Lunceford JK, Davidian M (2004) Stratification and weighting via the propensity score in estimation of causal treatment effects: a comparative study. Stat Med 23(19):2937–2960
Manski CF (2013) Identification of treatment response with social interactions. Econ J 16(1):S1–S23
Neyman J, Iwaszkiewicz K (1935) Statistical problems in agricultural experimentation. Suppl J R Stat Soc 2(2):107–180
Rahmqvist J, Wells MB, Sarkadi A (2013) Conscious parenting: a qualitative study on swedish parents’ motives to participate in a parenting program. J Child Fam Stud. doi:10.1007/s10826-013-9750-1
Rosenbaum PR (2007) Interference between units in randomized experiments. J Am Stat Assoc 102(477):191–200
Rosenbaum PR, Rubin DB (1983) The central role of the propensity score in observational studies for causal effects. Biometrika 70(1):41–55
Rubin DB (1974) Estimating causal effects of treatments in randomized and nonrandomized studies. J Educ Psychol 66:688–701
Rubin DB (1980) Discussion of randomization analysis of experimental data: the Fisher randomization test by D. Basu. J Am Stat Assoc 75(371):591–593
Rubin DB (2010) Reflections stimulated by the comments of Shadish (2010) and West and Thoemmes (2010). Psychol Methods 15(1):38–46
Sampaio F, Feldman I, Sarkadi A (2012) A cost-effectiveness analysis of the Triple P program in Uppsala municipality, Sweden. Eur J Pub Health 22(S2):101–101
Schafer JL, Kang J et al (2008) Average causal effects from nonrandomized studies: a practical guide and simulated example. Psychol Methods 13(4):279–313
Sobel M (2006) What do randomized studies of housing mobility demonstrate? J Am Stat Assoc 101(476):1398–1407
Tchetgen Tchetgen EJ, VanderWeele TJ (2012) On causal inference in the presence of interference. Stat Methods Med Res 21(1):55–75
Verbitsky-Savitz N, Raudenbush SW (2012) Causal inference under interference in spatial settings: a case study evaluating community policing program in Chicago. Epidemiol Methods 1(1):106–130
Acknowledgments
The authors wish to thank the research group in social pediatrics lead by Anna Sarkadi, Uppsala University for providing the Triple P data set, a study funded by the Swedish National Institute of Public Health, Grant \(\#\)HFÅ 2008/214. The authors also acknowledge The Institute for Evaluation of Labour Market and Education Policy for partial financial support.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Appendix: Unbiasedness of the estimators \(\widehat{\textit{CE}}_{i}^{D*}, \widehat{\textit{CE}}^{D*}, \widehat{\textit{CE}}^{I*}\) and \(\widehat{\textit{CE}}^{T*}\)
Appendix: Unbiasedness of the estimators \(\widehat{\textit{CE}}_{i}^{D*}, \widehat{\textit{CE}}^{D*}, \widehat{\textit{CE}}^{I*}\) and \(\widehat{\textit{CE}}^{T*}\)
In the following, we show that the expected value of \(\widehat{Y}_{i}^{*}(1;\psi )\) conditional on \(S_{i}=1\) equals \(\overline{Y}_{i}(1;\psi )\) and that the expected value of \(\widehat{Y}^{*}(z;\psi )\) equals \(\overline{Y}(z;\psi )\). From this and the definitions of \(\overline{\textit{CE}}_{i}^{D}\), \(\overline{\textit{CE}}^{D}\), \(\overline{\textit{CE}}^{I}\) and \(\overline{\textit{CE}}^{T},\) unbiasedness of \(\widehat{\textit{CE}}_{i}^{D*}\), \(\widehat{\textit{CE}}^{D*}\), \(\widehat{\textit{CE}}^{I*}\) and \(\widehat{\textit{CE}}^{T*}\)follow. We define \(R_{k}^{n}\) as the subset of all possible treatment assignment vectors \(\varvec{\omega },\) with exactly \(k\) of the \(n\) individuals receiving treatment \(z=1\).
In the same manner, it can be shown that \(E(\widehat{Y}_{i}^{*}(0;\psi )\mid S_{i}=1)=\overline{Y}_{i}(0;\psi )\),
Rights and permissions
About this article
Cite this article
Lundin, M., Karlsson, M. Estimation of causal effects in observational studies with interference between units. Stat Methods Appl 23, 417–433 (2014). https://doi.org/10.1007/s10260-014-0257-8
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10260-014-0257-8