Date created: | Last Updated:

Category: Project

Description: This is open-source R code that contains functions I have written to perform Bayesian analyses, including my own versions of a robust t-test, two-group ordinal-variable t-test (based on an ordered probit model), comparison of proportions, power analysis for simple designs, diagnostic efficiency/interrater agreement coefficients, Bayesian bootstrapped correlations (including robust correlations, polyserial correlations, and things not covered in the BBcor package), and calculation of several useful indices such as the probability of direction (the proportion of the posterior distribution on the same side of the point null as the median) and the Region of Practical Equivalence (ROPE) Bayes Factor, which summarizes the evidence regarding whether a parameter value is more likely within or outside of a pre-identified interval null hypothesis (i.e., the ROPE). This code is largely designed for my own use, but I hope to share it here to increase the reproducibility of my published research and allow me to cite this method in my publications. However, I have found these functions to be exceptionally useful for many statistical analyses, and I encourage other researchers to utilize them in their own work if they would like (and if you do, please cite this page). Documentation may be somewhat limited (one day I hope to get some enterprising grad student to help me turn this into a real R package), so do not hesitate to contact me by email (zachary.j.williams@vanderbilt.edu) should you have any issues running this code.