Lab 7: Communicating results from Bayesian analysis
This document mainly serves as an opportunity to show different ways of presenting and communicating Bayesian model results. Below, I reprint some examples of mostly visualizations, some of which you may have seen before in lecture. Many authors use similar types of figures to communicate results. The examples here are merely examples meant to provide some inspiration for your own work! There are several R packages with functions to create tables and figures similar to the ones below. In Scogin et al. (2019), we discuss some of them.
For background on the respective empirical models and context, please access the articles listed in this tutorial; you can find full citations at the end, and the articles are made available to you as part of this course.
Regression coefficient tables
The table above is from Lee and Murdie (2021).
The table above is from Blais, Guntermann, and Bodet (2017).
The table above is from Beazer and Woo (2016).
The table above is from Helgason and Mérola (2017).
The table above is from Cao and Ward (2017).
Regression coefficient plots
The figure above is from Cao and Ward (2017).
The figure above shows the posterior distribution of linear regression coefficients; from Lee and Murdie (2021).
The figure above is from Karreth (2018).
Postestimation quantities
The figure above shows the posterior distribution of predicted changes; from Helgason and Mérola (2017).
The figure above shows the posterior distribution of first differences after a logistic regression model; from Karreth (2018).
The figure above shows the posterior distribution of linear regression coefficients; from Karreth, Tir, and Gibler (2022).
The figure above shows the credible intervals of first differences after a series of logistic regression models; from my working paper, “Regional Trade Agreements and Concessions in GATT/WTO Trade Disputes”.
The figure above shows the posterior distribution of latent democracy scores based on an IRT model; from Treier and Jackman (2008).