TreeBUGS: Hierarchical MPTs

TreeBUGS is an R package that facilitates using hierarchical multinomial processing tree (MPT) models that are often used in cognitive psychology (Erdfelder et al., 2009). Specifically, the Beta-MPT (Smith & Batchelder, 2010) and the latent-trait MPT model (Klauer, 2010) are implemented.

General Procedure of Using TreeBUGS

In the most simple user scenario, the following steps are required:

  1. Define path to existing MPT model file in .eqn format (cf. multiTree; Moshagen, 2010)
  2. Define path to data set with individual frequencies (.csv file: comma separated, rows=persons, columns=labeled categories)
  3. Call betaMPT or traitMPT (exact code in manual/vignette)
  4. Check convergence of MCMC chains
  5. Summarize and plot results using functions tailored to MPT models

These steps are explained in more detail in the package vignette, which can be opened in R by typing vignette("TreeBUGS"). Note that TreeBUGS requires a valid installation of the software JAGS.

To install the latest stable release of TreeBUGS from CRAN, run:

install.packages("TreeBUGS")

To install the latest developer version of TreeBUGS from the GitHub repository, run:

install.packages("devtools")
library(devtools)
install_github("denis-arnold/TreeBUGS", build_vignettes = TRUE)

 

Citation

If you use TreeBUGS, please cite the software as follows:

  • Heck*, D. W., Arnold*, N. R., & Arnold, D. (in press). TreeBUGS: An R package for hierarchical multinomial-processing-tree modeling. Behavior Research Methods. doi:10.3758/s13428-017-0869-7
    [BibTeX] [Abstract]

    Multinomial processing tree (MPT) models are a class of measurement models that allow explaining categorical data by a finite number of underlying cognitive processes. Traditionally, data are aggregated across participants and analyzed under the assumption of independently and identically distributed observations. Hierarchical Bayesian extensions of MPT models explicitly account for participant heterogeneity by assuming that the individual parameters follow a continuous hierarchical distribution. We provide an accessible introduction to hierarchical MPT modeling and present the user-friendly and comprehensive R package TreeBUGS, which implements the two most important hierarchical MPT approaches for participant heterogeneity − the beta-MPT (Smith & Batchelder, 2010) and the latent-trait MPT approach (Klauer, 2010). TreeBUGS reads standard MPT model files and obtains Markov chain Monte Carlo samples that approximate the posterior distribution. The functionality and output is tailored to the specific needs of MPT modelers and provides tests for the homogeneity of items and participants, individual and group parameters estimates, fit statistics, between-subject comparisons, as well as goodness-of-fit and summary plots. We also propose and implement novel statistical extensions to include continuous and discrete predictors (either as fixed or random effects) in the latent-trait MPT model.

    @article{heck2017treebugs2,
    title = {{{TreeBUGS}}: {{An R}} Package for Hierarchical Multinomial-Processing-Tree Modeling},
    doi = {10.3758/s13428-017-0869-7},
    abstract = {Multinomial processing tree (MPT) models are a class of measurement models that allow explaining categorical data by a finite number of underlying cognitive processes. Traditionally, data are aggregated across participants and analyzed under the assumption of independently and identically distributed observations. Hierarchical Bayesian extensions of MPT models explicitly account for participant heterogeneity by assuming that the individual parameters follow a continuous hierarchical distribution. We provide an accessible introduction to hierarchical MPT modeling and present the user-friendly and comprehensive R package TreeBUGS, which implements the two most important hierarchical MPT approaches for participant heterogeneity − the beta-MPT (Smith \& Batchelder, 2010) and the latent-trait MPT approach (Klauer, 2010). TreeBUGS reads standard MPT model files and obtains Markov chain Monte Carlo samples that approximate the posterior distribution. The functionality and output is tailored to the specific needs of MPT modelers and provides tests for the homogeneity of items and participants, individual and group parameters estimates, fit statistics, between-subject comparisons, as well as goodness-of-fit and summary plots. We also propose and implement novel statistical extensions to include continuous and discrete predictors (either as fixed or random effects) in the latent-trait MPT model.},
    journaltitle = {Behavior Research Methods},
    author = {Heck*, Daniel W and Arnold*, Nina R and Arnold, Denis},
    date = {2017},
    keywords = {heckshared},
    pubstate = {inpress}
    }

(* Equal contributions of the first two authors.)

 

References

  • Erdfelder, E., Auer, T.-S., Hilbig, B. E., Assfalg, A., Moshagen, M., & Nadarevic, L. (2009). Multinomial processing tree models: A review of the literature. Journal of Psychology, 217, 108–124. http://doi.org/10.1027/0044-3409.217.3.108

  • Klauer, K. C. (2010). Hierarchical multinomial processing tree models: A latent-trait approach. Psychometrika, 75, 70–98. http://doi.org/10.1007/s11336-009-9141-0

  • Matzke, D., Dolan, C. V., Batchelder, W. H., & Wagenmakers, E.-J. (2015). Bayesian estimation of multinomial processing tree models with heterogeneity in participants and items. Psychometrika, 80, 205–235. http://doi.org/10.1007/s11336-013-9374-9

  • Moshagen, M. (2010). multiTree: A computer program for the analysis of multinomial processing tree models. Behavior Research Methods, 42, 42–54. http://doi.org/10.3758/BRM.42.1.42

  • Smith, J. B., & Batchelder, W. H. (2010). Beta-MPT: Multinomial processing tree models for addressing individual differences. Journal of Mathematical Psychology, 54, 167–183. http://doi.org/10.1016/j.jmp.2009.06.007