What question did this study set out to answer?

The research aims to compare logistic regression and propensity score methods in low event conditions with multiple confounders.

July 25, 2003American Journal of EpidemiologyOpen Access

Comparison of Logistic Regression versus Propensity Score When the Number of Events Is Low and There Are Multiple Confounders

View Full Paper

Ask AI

Bookmark

The research aims to compare logistic regression and propensity score methods in low event conditions with multiple confounders.
Used Monte Carlo simulations to perform 252,480 trials.
Simulated a cohort study focusing on bias, precision, coverage probability, and power.
Examined performance related to the ratio of events to confounders.
Propensity scores were less biased and more precise with seven or fewer events per confounder.
Logistic regression showed increased coverage probability with higher events per confounder.
Overall, propensity score methods exhibited more empirical power than logistic regression.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Randomized trial compares logistic regression and propensity score methods for dealing with low event counts and confounders, suggesting practical applications.

Cite This Study

M. Soledad Cepeda (2003) studied this question.

synapsesocial.com/papers/6a0d293cd7cdc72b8665672e https://doi.org/10.1093/aje/kwg115

View Full Paper

Ask AI

Bookmark

Implication

Randomized trial compares logistic regression and propensity score methods for dealing with low event counts and confounders, suggesting practical applications.

Comparison of Logistic Regression versus Propensity Score When the Number of Events Is Low and There Are Multiple Confounders

Key Points

Discussion

Implication

Cite This Study

Comparison of Logistic Regression versus Propensity Score When the Number of Events Is Low and There Are Multiple Confounders

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider