Data Analysis Using Regression and Multilevel/Hierarchical Models

Chapter 12

radon_m0

Model to predict radon measurements in houses in Minnesota from page 259 of Data Analysis Using Regression and Multilevel/Hierarchical Models.

Dataset: radon

radon_m0_nc

Model to predict radon measurements in houses in Minnesota from page 259 of Data Analysis Using Regression and Multilevel/Hierarchical Models with non-centered parameterization.

Dataset: radon

radon_m1_nc

Model to predict radon measurements in houses in Minnesota from page 259 of Data Analysis Using Regression and Multilevel/Hierarchical Models with non-centered parameterization and floor-level predictor.

Dataset: radon

radon_m2_nc

Model to predict radon measurements in houses in Minnesota from page 266 of Data Analysis Using Regression and Multilevel/Hierarchical Models with non-centered parameterization, floor-level predictor, and county-level soil uranium predictor.

Dataset: radon

Chapter 13

latin_square_with_predictors

Non-centered variant of the ANOVA model for latin square experiment from equation 13.10 on page 292 of Data Analysis Using Regression and Multilevel/Hierarchical Models. This is an expanded version of latin_square_without_predictors, including linear predictors for rows, columns, or treatments.

Dataset: latin_square

Chapter 14

m1

Model for CBS poll results with state-level effects from page 302 of “Data Analysis Using Regression and Multilevel/Hierarchical Models”.

Dataset: election88

m2-nc

Model for CBS poll results with non-nested hierarchical effects from based on page 381 of “Data Analysis Using Regression and Multilevel/Hierarchical Models”. The original BUGS code is available at http://www.stat.columbia.edu/~gelman/arm/examples/election88/election88.M2.bug. The model has been modified as follows:

• a_age, a_edu, and a_region are non-centered.

• The hierarchical prior on a_state given regional parameters has been moved to the logits.

• Features black, female, black-female interaction, and v_prev_full have been de-meaned.

Dataset: election88

m2

Model for CBS poll results with non-nested hierarchical effects from page 381 of “Data Analysis Using Regression and Multilevel/Hierarchical Models”. The BUGS code is available at http://www.stat.columbia.edu/~gelman/arm/examples/election88/election88.M2.bug.

Dataset: election88

Chapter 22

latin_square_without_predictors

Non-centered variant of the ANOVA model for latin square experiment from equation 22.6 on page 497 of Data Analysis Using Regression and Multilevel/Hierarchical Models. This is a simplified version of latin_square_with_predictors, not including linear predictors for rows, columns, or treatments.

Dataset: latin_square

pilots-1-nc

Two-way ANOVA fron page 495 based on http://www.stat.columbia.edu/~gelman/arm/examples/pilots/pilots.1.bug.

Dataset: pilots

pilots-1

Two-way ANOVA fron page 495 based on http://www.stat.columbia.edu/~gelman/arm/examples/pilots/pilots.1.bug.

Dataset: pilots