profileR

JOSS manuscript:

Title: Profile Analysis of Multivariate Data in R

Type: Package

Overview

profileR can be used for estimating profile analytic models. This includes the multivariate methods and data visualization tools to implement profile analysis and cross-validation techniques described in previous studies, including Davison and Davenport (2002), Bulut (2013), and Bulut, Davison, and Rodriguez (2017). Some of the principal functions in profileR include:

1. Statistical methods:

• Profile analysis for one sample using Hotelling's T-squared statistic -- paos()
• Profile analysis by groups to assess parallelism, equality, and flatness -- pbg()
• Profile analysis via multidimensional scaling (a.k.a., PAMS) -- pams()
• Criterion-related profile analysis -- cpa()
• Pattern and level reliability for profiles -- pr()

1. Visualizations:

• Profile plots for a set of multivariate scores (see profileplot())
• Profile plots returned from profile analysis by groups (see pbg())
• Plots for showing pattern and levels effects in criterion-related profile analysis (see cpa())

Citing profileR

To cite profileR in your work, please use the following APA-style citation:

Desjardins, C. D., & Bulut, O. (2020). profileR: An R package for profile analysis. Journal of Open Source Software, 5(47), 1941, doi: 10.21105/joss.01941

Installing profileR

Most recent version (and early versions) are installable from CRAN by using:

R install.packages("profileR")

and the developmental version on GitHub can be installed by using:

R Sys.setenv(R_REMOTES_NO_ERRORS_FROM_WARNINGS=TRUE) # in case your R version is older devtools::install_github(repo = "cddesja/profileR") # without the vignette

If LaTeX is available, profileR can be installed with its vignette by using:

R Sys.setenv(R_REMOTES_NO_ERRORS_FROM_WARNINGS=TRUE) # in case your R version is older devtools::install_github(repo = "cddesja/profileR", build_vignettes = TRUE) # with the vignette

Verifying and testing profileR

To verify the installation and test that the examples work as intended, please visit our wiki.

Using profileR

To learn how to conduct profile analysis using profileR, you can see our paper available on PsyArxiv. This paper walks through some of the principal methods available in profileR. In the paper, a brief theory behind each method is presented, followed by a working example demonstrating how to use these methods in profileR. To see this paper as a vignette in the package:

R vignette("profiler-vignette", package = "profileR")

To cite this vignette in your work, you can use the following APA-style citations:

Bulut, O., & Desjardins, C. D. (2020). Profile analysis of multivariate data: A brief introduction to the profileR package. Retrieved from psyarxiv.com/sgy8m. doi: 10.31234/osf.io/sgy8m

Updates in profileR

Changes since v0.3

• Added vignette to test installation
• Added profileR vignette
• Added gtheory functions (temporarily removed)
• Added the spouse data in the package.
• Added ef and writing data.

Changes since v0.2

• Added wprifm, which performs a within-person random intercept factor model to obtain a score profile.
• EXPERIMENTAL: Added moderated profile analysis. This function is untested and based on unpublished methodology.
• Changed cp to cpa to avoid confusion with cp in Linux and Mac environments.
• Changed pc to pcv.

Contributing and Contact Information

We are open to suggestions for improvements in profileR. Therefore, we invite all users to post a question or to provide us with (either positive or negative) feedback on the functions available in profileR. When filing an issue, the most important thing is to include a minimal reproducible example so that we can quickly verify the problem, and then figure out how to fix it. There are two things you may need to include to make your example reproducible: data and code. If you are using additional packages, we would need that information as well.

In addition, we welcome users who would like to contribute to the profileR package by adding new functions or co-developing some functions with us. You can let us know which function(s) you want to develop or which of the existing function(s) you want to improve. Please note that profileR is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

To get further help regarding the functions available in profileR or inquries regarding contributions, please email us:

• Okan Bulut (okanbulut84@gmail.com)
• Chris Desjardins (cddesjardins@gmail.com)