Data and Code Supplement for 'A Diagnostic for Bias in Linear Mixed Model Estimators Induced by Dependence Between the Random Effects and the Corresponding Model Matrix'
Description
This dataset contains the R code and saved yearly model-fitting results supporting "A Diagnostic for Bias in Linear Mixed Model Estimators Induced by Dependence Between the Random Effects and the Corresponding Model Matrix" by Andrew Karl and Dale Zimmerman, published in the Journal of Statistical Planning and Inference 211 (2021), 107–118. The 2026 update preserves all original supplement files and adds code, inputs, fitted reference values and saved results for the additional conditional sports simulations described in the revised paper. For five unbalanced 2017 schedules, 100,000 replications per schedule estimate the residual-weight contribution added in the repaired Theorem 2, with Monte Carlo standard errors and a mean-zero control variate. The balanced WNBA contribution is checked algebraically. These calculations condition on the schedules and original fitted team effects and residual variances; they do not establish general negligibility of the extra term. The update also supplies portable R scripts for the original yearly-results plots and synthetic experiment, explicit random-sampling settings, saved-result verification, file descriptions and checksums. See README.txt and CHANGELOG.txt for scope, commands and changes.
Files
Steps to reproduce
Instructions are in the README file.