|
Activity Number:
|
580
- Methodological Developments and Implications for Social Scientists
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, July 31, 2019 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Social Statistics Section
|
|
Abstract #307133
|
Presentation
|
|
Title:
|
Did You Conduct a Sensitivity Analysis? a New Weighting-Based Approach for Evaluations of the Average Treatment Effect for the Treated
|
|
Author(s):
|
Guanglei Hong* and Fan Yang and Xu Qin
|
|
Companies:
|
University of Chicago and University of Colorado Denver and University of Pittsburgh
|
|
Keywords:
|
Causal inference;
confounding;
identification assumption;
propensity score;
selection bias;
sensitivity parameter
|
|
Abstract:
|
In quasi-experimental research, a sensitivity analysis helps determine whether a causal conclusion could be easily reversed in the presence of hidden bias. A new approach to sensitivity analysis extends and supplements propensity score weighting methods for identifying the average treatment effect for the treated (ATT). In its essence, the discrepancy between a new weight that adjusts for the omitted confounders and an initial weight that omits them captures the role of the confounders. This new weighting-based SA strategy is intuitive and demonstrates important advantages. (1) It does not rely on additional simplifying assumptions. (2) The number of weighting-based sensitivity parameters is small and their forms never change with data generation functions. (3) It is unconstrained by the measurement scales of the omitted confounders. (4) It can conveniently assess the aggregate bias associated with multiple omitted confounders. And (5) a graphical display of the sensitivity parameter values provides a holistic view of the dominant potential bias. An application to the well-known LaLonde data lays out the implementation procedure and illustrates its broad utility.
|
Authors who are presenting talks have a * after their name.
