Influence functions continued. A framework for estimating standard errors in reweighting, matching, and regression adjustment

Jann, Ben (2020). Influence functions continued. A framework for estimating standard errors in reweighting, matching, and regression adjustment (University of Bern Social Sciences Working Papers 35). Bern: University of Bern, Department of Social Sciences

[img]
Preview
Text (revised version: April 9, 2020)
jann-2020-IF.pdf - Published Version
Available under License BORIS Standard License.

Download (965kB) | Preview
[img]
Preview
Text (original version: March 28, 2020)
jann-2020-IF.pdf - Published Version
Available under License BORIS Standard License.

Download (968kB) | Preview

In Jann (2019) I provided some reflections on influence functions for linear regression (with an application to regression adjustment). Based on an analogy to variance estimation in the generalized method of moments (GMM), I extend the discussion in this paper to maximum-likelihood models such as logistic regression and then provide influence functions for a variety of treatment effect estimators such as inverse-probability weighting (IPW), regression adjustment (RA), inverse-probability weighted regression adjustment (IPWRA), exact matching (EM), Mahalanobis distance matching (MD), and entropy balancing (EB). The goal of this exercise is to provide a framework for standard error estimation in all these estimators.

Item Type:

Working Paper

Division/Institute:

03 Faculty of Business, Economics and Social Sciences > Social Sciences > Institute of Sociology

UniBE Contributor:

Jann, Ben

Subjects:

300 Social sciences, sociology & anthropology

Series:

University of Bern Social Sciences Working Papers

Publisher:

University of Bern, Department of Social Sciences

Language:

English

Submitter:

Ben Jann

Date Deposited:

07 May 2020 15:05

Last Modified:

07 May 2020 15:05

JEL Classification:

C01, C12, C13, C21, C25, C31, C83, C87

BORIS DOI:

10.7892/boris.142529

URI:

https://boris.unibe.ch/id/eprint/142529

Actions (login required)

Edit item Edit item
Provide Feedback