Two postestimation commands for assessing confounding effects in epidemiological studies
Zhiqiang Wang
School of Medicine and School of Population Health
University of Queensland
Brisbane, Queensland, Australia
[email protected]
|
Abstract. Confounding is a major issue in observational epidemiological studies. This
paper describes two postestimation commands for assessing confounding
effects. One command (confall) displays and plots all possible effect
estimates against one of p-value, Akaike information criterion, or
Bayesian information criterion. This computing-intensive procedure allows
researchers to inspect the variability of the effect estimates from various
possible models. Another command (chest) uses a stepwise approach to
identify variables that have substantially changed the effect estimate. Both
commands can be used after most common estimation commands in
epidemiological studies, such as logistic regression, conditional logistic
regression, Poisson regression, linear regression, and Cox proportional
hazards models.
View all articles by this author:
Zhiqiang Wang
View all articles with these keywords:
confall, confgr, chest, epidemiological methods, confounding, all possible effects, change in estimate
Download citation: BibTeX RIS
Download citation and abstract: BibTeX RIS
|
