Outcome-free Design of Observational Studies: Peer Influence on Smoking
View/ Open
347.pdf (1.081Mb)
Access Status
Full text of the requested work is not available in DASH at this time ("restricted access"). For more information on restricted deposits, see our FAQ.Author
Langenskioeld, S.
Note: Order does not necessarily reflect citation order of authors.
Published Version
https://doi.org/10.2307/27917241Metadata
Show full item recordCitation
Langenskiöld, S. and Donald Rubin. 2008. “Outcome-Free Design of Observational Studies: Peer Influence on Smoking.” Annales d’Économie et de Statistique (91/92): 107. doi:10.2307/27917241.Abstract
For estimating causal effects of treatments, randomized experiments are appropriately considered the gold standard, although they are often infeasible for a variety of reasons. Nevertheless, nonrandomized studies can and should be designed to approximate randomized experiments by using only background information to create subgroups of similar treated and control units, where "similar" here refers to their distributions of background variables. This activity should be conducted without access to any outcome data to assure the objectivity of the design. In many situations, these goals can be accomplished using propensity score methods, as illustrated here in the context of a study on whether nonsmoking Harvard freshmen are influenced by their smoking peers. In that study, propensity score methods were used to create matched groups of treated units (rooming with at least one smoker) and control units (rooming with only non-smokers) who are at least as similar with respect to their distributions of observed background characteristics as if they had been randomized, thereby approximating a randomized experiment with respect to the observed covariates.Citable link to this page
http://nrs.harvard.edu/urn-3:HUL.InstRepos:34325448
Collections
- FAS Scholarly Articles [18292]
Contact administrator regarding this item (to report mistakes or request changes)