Bayesian Methods for Management and Business

Bayesian Methods for Management and Business: Pragmatic Solutions for Real Problems

Bayesian statistical methods for data analysis have surged in popularity in recent years. Bayesian statistics have been used in many disciplines to reveal new insights and understand difficult problems. However, to date they have been infrequently used for most kinds of business and management problems.

In this book, Bayesian methods are used to examine real-world data arising in business areas such as strategy, international business, accounting and information systems. A practical problem-solving approach is used and the reader is shown how to use freely-available software tools to answer business questions using data arising in the business environment.

Pathways through the Book

There are several pathways through the book so that different kinds of readers with different backgrounds and interests can approach the material differently. Here are some of these pathways:

For advanced MBA students with a solid grounding in the regression model, the pathway might include Chapters 1, 2, 5, 6, and 7 with the other chapters being optional or for future reading.

For Masters of Science in Analytics students, the above path could be followed and then continued by the earlier parts of Chapters 8 and 9.

For those who want a solid grounding in the foundations of Bayesian inference, Chapters 3 and 4 should be added. A traditional approach would place foundations early and the book follows this approach. However, do not feel compelled to place them early in your reading of the book -- different learning styles approach material differently.

For those with a highly practical orientation, the In Practice sections throughout the book focus on immediately applying Bayesian statistics to real business data.

Downloads

Click here to download a ZIP file containing all of the code used in the book. The freely-available WinBUGS and R software applications are needed to run the code.
2023 Update: I currently recommend MultiBUGS 2.0 (available at https://www.multibugs.org/) for use with the WinBUGS code in the book. MultiBUGS supports parallel computation (much faster processing) using the distribute workers per chain option. It also provides effective sample size (ESS) computations by default (see page 174 in the book), removing the need for separate analyses. There are also other improvements such as more informative error messages, additional examples, new information criterion options, new distributions in the ReliaBUGS examples, and LaTeX export to name a few. I've been personally using MultiBUGS and MultiBUGS 2.0 for several years and generally prefer it to WinBUGS (for an advanced exception see here.)
Please note: if you are trying to use the files with OpenBUGS and get a file-load error, this is probably due to OpenBUGS not accepting WinBUGS graphics. What you can do instead is load the file with WinBUGS, then copy/paste all of the code and data (without the graphics) to OpenBUGS. The program should then hopefully run in OpenBUGS. I have not tested the code with OpenBUGS but I offer this as a work-around to an issue that was reported to me. I tested this work-around with one such file and the file ran in OpenBUGS.

Errata

Click here for the current errata for Bayesian Methods for Management and Business. Spotted a typo? Please send me an email at edhahn@salisbury.edu. Your contribution will be acknowledged in the errata listing.

Other Resources for Bayesian Management Research

More information on Bayesian methods in management is available at the Bayesian Management Research page.