MBESS: An R Package


What is MBESS?

MBESS (pronounced em-bes) is an Open Source R Package that provides various advanced and nonstandard methods that are not easily implemented elsewhere, especially with regards to effect sizes, confidence intervals for effect sizes, and sample size planning. For those not familiar with R, R is a powerful program for statistical computing and graphics and is Freely available for (essentially) all operating systems. Originally, MBESS stood for “Methods for the Behavioral, Educational, and Social Sciences,” but at this point MBESS contains methods applicable and used in a wide variety of fields and is an orphan acronym, in the sense that what was an acronym is now literally its name.

The Goal of MBESS-General

The goal of MBESS is to provide with an R package that contains useful functions for nonstandard techniques when designing studies and analyzing data.

The Goal of MBESS-Specific

The long term goal of MBESS is for it to contain a relatively complete set of functions to compute a wide variety of unstandardized and standardized effect sizes, confidence intervals for those effect sizes, and to plan sample size from the power analytic and accuracy in parameter estimation approaches.

Strengths of MBESS

MBESS has a suite of functions for a variety of related topics: effect sizes, confidence intervals for effect sizes (including standardized effect sizes), sample size planning (from the accuracy in parameter estimation [AIPE], power analytic, equivalence, and minimum-risk sequential perspectives), mediation analysis, various properties of distributions, and a variety of utility functions.


Contributions can be as simple as submitting ideas for functions that you would like to see included in the MBESS package (feel free to email me with any ideas), by pointing out bugs or typos (such contributions will be noted in the Acknowledgements file), or by submitting functions that can be included in the package (of course, the author of the function will be given in the help-files for the contributed functions).

MBESS has had several contributors of functions or helped as I prepared functions to accomplish some task (e.g., in a collaborative project). I am grateful for their help and I list these contributors here:

Bhargab Chattopadhyay
- Worked collaboratively on the minimum risk for point estimation function [mr.smd() & mr.cv()]

Keke (Brian) Lai
-Provided several contributed functions based on collaborative projects, former author (of Version 3), former maintainer (versions 2 & 3).

Mark Lachowicz
- Provided UpsES() function and upsilon().

Kem Phillips
-Provided equivalence functions.

Kris Preacher
-Consulted on mediation functions.

Sunthud Pornprasertmanit
-Several contributed functions based on collaborative projects. 

Joe Rausch
-Provided help and discussion on some early functions used in MBESS.

Leann J. Terry (now Leann J. Terry Diederich)
-Helped with ss.aipe.reliability() function.

Po-Ju Wu
-Helped with vit() function.

Jan Hartman
-Provided bug fix for ss.aipe.reg.coef()

Uri Simonsohn, Jeremy Biesanz, Jochem König
-Suggested that conf.limits.nct() should treat zero noncentrality in accord with any other value of the noncentrailty paramether (and not revert to the central t-distribution when the noncentrailty value was zero).

Wing Ho Man

-Suggested updated for the example used in the mediation() function documentation.

Samantha Anderson

-Provided bug reports and suggested fixes for ci.sc() and sc.anova().

Guy Prochilo

-Asked a question which led to a fix for ss.aipe.sm() regarding negative standardized means.

Giulio Costantini

-Pointed out a mismatch between the behavior of ci.reliability() and the manual.

Keith Widaman

-Pointed out issues with the Holzinger and Swineford dataset.

Licensing Information
The MBESS package is Open Source software that follows the guidelines set forth by the Free Software Foundation's GNU General Public License. Thus, the MBESS software (i.e., code) is freely available to anyone and can be freely modified by anyone (although, if the MBESS package is modified its name must be changed).

Getting and Using MBESS
The MBESS source code, Mac and Windows binaries, and help files are available on the Comprehensive R Archival Network's MBESS page. The easiest way to get MBESS is actually to use install.packages(“MBESS”) at the R command prompt.

Getting Help with MBESS
The MBESS Help files are available here.

History of MBESS
The first public release version (0.0.1) of MBESS was May 6, 2006. MBESS was peer reviewed in Kelley (2007a, Behavior Research Methods) and Kelley (2007b, Journal of Statistical Software). Version 1.0.0 was released in December, 2007, Version 2.0.0 was released in November, 2008, and Version 3.0.0 was released in May, 2010. The new version of MBESS (Version 4.0.0) was released in 2016. MBESS 3.0.0 – 3.3.3 included, my (former) graduate student Keke (Brian) Lai (now at UC Merced) as coauthor. Keke Lai was the maintainer of the package from Version 2.0.0 – 3.3.3 but is no longer.

Citing MBESS

Please cite the articles in which MBESS was peer reviewed as well as the version of MBESS used within your research. Additionally, if you use a function developed in an article, please also cite that article.

Kelley, K. (2017). MBESS (Version 4.0.0 and higher) [computer software and manual]. 
Accessible from http://cran.r-project.org.

Kelley, K. (2007b). Confidence intervals for standardized effect sizes: Theory, application, and implementation. Journal of Statistical Software, 20(8), 1-24.

Kelley, K. (2007a). Methods for the Behavioral, Educational, and Social Sciences: An R Package. Behavior Research Methods, 39, 979–984.