Randomization does not justify logistic regression
From MaRDI portal
Abstract: The logit model is often used to analyze experimental data. However, randomization does not justify the model, so the usual estimators can be inconsistent. A consistent estimator is proposed. Neyman's non-parametric setup is used as a benchmark. In this setup, each subject has two potential responses, one if treated and the other if untreated; only one of the two responses can be observed. Beside the mathematics, there are simulation results, a brief review of the literature, and some recommendations for practice.
Recommendations
- Logit effect values and efficiency of significance tests when a noncollapsible variable is omitted
- On regression adjustments to experimental data
- COMPLEMENTARY NONPARAMETRIC ANALYSIS OF COVARIANCE FOR LOGISTIC REGRESSION IN A RANDOMIZED CLINICAL TRIAL SETTING
- Experimental Randomization and the Validity of Normal-Theory Inference
- Bias of the regression estimator for experiments using clustered random assignment
Cites work
- scientific article; zbMATH DE number 3148413 (Why is no real title available?)
- scientific article; zbMATH DE number 3984351 (Why is no real title available?)
- scientific article; zbMATH DE number 720687 (Why is no real title available?)
- scientific article; zbMATH DE number 3213198 (Why is no real title available?)
- Alternative Models for the Analysis of Variance
- Association, causation, and marginal structural models.
- Causal Parameters and Policy Analysis in Economics: A Twentieth Century Retrospective*
- Concavity of the Log Likelihood
- Covariance adjustment in radomized experiments and observational studies
- On Multinomial Coefficients
- On regression adjustments in experiments with several treatments
- On regression adjustments to experimental data
- Probability Inequalities for Sums of Bounded Random Variables
- Some Surprising Results about Covariate Adjustment in Logistic Regression Models
- Statistical Models: Theory and Practice
- Statistics and Causal Inference
- The Effect of Pooling Across Strata in Perfectly Balanced Studies
- Using Regression Models to Analyze Randomized Trials: Asymptotically Valid Hypothesis Tests Despite Incorrectly Specified Models
Cited in
(16)- Randomization-based causal inference from split-plot designs
- Randomization-based test for censored outcomes: a new look at the logrank test
- No star is good news: a unified look at rerandomization based on \(p\)-values from covariate balance tests
- On regression adjustments in experiments with several treatments
- Using standard tools from finite population sampling to improve causal inference for complex experiments
- Randomization in the design of experiments
- Logit effect values and efficiency of significance tests when a noncollapsible variable is omitted
- Combining multiple observational data sources to estimate causal effects
- Lasso adjustments of treatment effect estimates in randomized experiments
- The Generalized Oaxaca-Blinder Estimator
- Correct and logical causal inference for binary and time‐to‐event outcomes in randomized controlled trials
- Using Regression Models to Analyze Randomized Trials: Asymptotically Valid Hypothesis Tests Despite Incorrectly Specified Models
- Transportability of model-based estimands in evidence synthesis
- Toward Better Practice of Covariate Adjustment in Analyzing Randomized Clinical Trials
- Randomization-based Joint Central Limit Theorem and Efficient Covariate Adjustment in Randomized Block 2 K Factorial Experiments
- Robust variance estimation for covariate-adjusted unconditional treatment effect in randomized clinical trials with binary outcomes
This page was built for publication: Randomization does not justify logistic regression
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q900462)
