|
A New Strategy for Eliminating Selection Bias in Non-experimental Evaluations
|
| Presenter(s):
|
| Laura Peck, Arizona State University, laura.peck@asu.edu
|
| Furio Camillo, University of Bologna, furio.camillo@unibo.it
|
| Ida D’Attoma, University of Bologna, ida.dattoma2@unibo.it
|
| Abstract:
This paper presents a creative and practical approach to dealing with the problem of selection bias. Taking an algorithmic approach and capitalizing on the known treatment-associated variance in the X matrix, we propose a data transformation that allows estimating unbiased treatment effects. The approach does not call for modeling the data, based on underlying theories or assumptions about the selection process, but instead it calls for using the existing variability within the data and letting the data speak. We illustrate with an application of the method to Italian Job Centers.
|
|
The Truncation-by-Death Problem: What to do in an Experimental Evaluation When the Outcome is Not Always Defined
|
| Presenter(s):
|
| Sheena McConnell, Mathematica Policy Research, smcconnell@mathematica-mpr.com
|
| Elizabeth Stuart, Johns Hopkins University, estuart@jhsph.edu
|
| Barbara Devaney, Mathematica Policy Research, bdevaney@mathematica-mpr.com
|
| Abstract:
While experiments are viewed as the gold standard for evaluation, some of their benefits may be lost when, as is common, outcomes are not defined for some sample members. In evaluations of marriage interventions, for example, a key outcome—relationship quality—is undefined when a couple splits up. This paper shows how treatment-control differences in mean outcomes can be misleading when outcomes are not defined for everyone and discusses ways to identify the seriousness of the problem. Potential solutions to the problem are described, including approaches that rely on simple treatment-control differences-in-means as well as more complex modeling approaches.
|
| |