rtmpt

R package for fitting Response-Time extended Multinomial Processing Tree (RT-MPT) models by Klauer and Kellen (2018)

Description

The model class RT-MPT incorporate response frequencies and response latencies. This enables the estimation of

process-completion times: the time a process takes
encoding plus motor execution times (aka non-decision times): the time for encoding a stimuli plus executing the response
process-probabilities: the probability with which a process occures (as in traditional MPTs).

In total we have for each process one probability parameter and two process completion times (for each outcome of the process one).

rtmpt has two sub-model classes: the exponential RT-MPT uses a Metropolis-within-Gibbs sampler and builds on the C++ code by Klauer and Kellen (2018) with some modifications. The diffusion RT-MPT uses a Hamiltonian-Within-Gibbs sampler and builds on the C++ code by Klauer, Hartmann, and Meyer-Grant (submitted manuscript).

In the exponential RT-MPT (short ertmpt) model class it is possible to

set process probabilities to constants
set two or more probabilities equal
set the process completion time of one process outcome to zero
set the process completion time of multiple processes with the same outcome (+ or -) equal

In the diffusion RT-MPT (short drtmpt) model class it is possible to

set one or more of the diffusion parameters (threshold, drift rate, and relative starting point) for a process to constants
set one or more of the diffusion parameters (threshold, drift rate, and relative starting point) for multiple processes equal

For more information about the functionalities check the help files or the vignette of the package.

Installation

CRAN

An installation via CRAN is now possible. Write in the command line in R:

    install.packages("rtmpt")

Alternative instructions

Linux

Before installing the R package rtmpt you need to install the latest version of the GNU Scientific Library (GSL):

Download the latest *tar.gz file from GSL
Unzip the file
Assuming you downloaded version 2.5 and unziped the file to ~/Downloads/gsl-2.5, open terminal and change directory via $ cd ~/Downloads/gsl-2.5
Execute the following lines:
```
./configure
make
sudo make install
```

For some Linux distributions it is necessary to run sudo ldconfig -v in the terminal before installing rtmpt.

Installing rtmpt is possible either

with the R package devtools and the command devtools::install_github("RaphaelHartmann/rtmpt@master") in R or
with the following steps:
1. Download all the source files from the rtmpt package above
2. In the terminal change directory via $ cd ~/Downloads (assuming the rtmpt source folder is in /Downloads)
3. Execute:
```
R CMD build rtmpt
R CMD INSTALL rtmpt_<XYZ>.tar.gz
```
where XYZ stands for the version number.

MacOS

Before installing the R package rtmpt from source you need to do the following:

If not already installed, install Homebrew
Install GSL via Homebrew using terminal and command brew install gsl
Install Xcode from the App Store and then install the Command Line Tools using xcode-select --install

If you use MacOS Mojave, make sure to also run the following command:

open /Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg

Also execute brew install gcc in your terminal.
Check via ls /usr/local/bin/ what endings the gcc/g++ executables have (in my case -9)
If ~/.R/Makevars does not already exist, create it with mkdir ~/.R/ and touch ~/.R/Makevars
Use open ~/.R/Makevars to write the following in the Makevars file (assuming your gcc/g++ executables have the ending -9):
```
CC=/usr/local/bin/gcc-9
CXX=/usr/local/bin/g++-9
CXX11=/usr/local/bin/g++-9
```
and save.

Installing the rtmpt package:

This is possible either

with the R package devtools and the command devtools::install_github("RaphaelHartmann/rtmpt@master") in R or
with the following steps:
1. Download all the source files from the rtmpt package above
2. In the terminal change directory via $ cd ~/Downloads (assuming the rtmpt source folder is in /Downloads)
3. Execute:
```
R CMD build rtmpt
R CMD INSTALL rtmpt_<XYZ>.tar.gz
```
where XYZ stands for the version number.

Citation

If you want to cite this package write

Hartmann, R., Johannsen, L., & Klauer, K. C. (2020). rtmpt: An R package for fitting response-time extended multinomial processing tree models. Behavior Research Methods, 52(3), 1313–1338. doi: 10.3758/s13428-019-01318-x

or use the BibTeX code:

@article{Hartmann_2020,
	doi = {10.3758/s13428-019-01318-x},
	url = {https://doi.org/10.3758%2Fs13428-019-01318-x},
	year = 2020,
	month = {may},
	publisher = {Springer Science and Business Media {LLC}},
	volume = {52},
	number = {3},
	pages = {1313--1338},
	author = {Raphael Hartmann and Lea Johannsen and Karl Christoph Klauer},
	title = {rtmpt: An R package for fitting response-time extended multinomial processing tree models},
	journal = {Behavior Research Methods}
}

References

Klauer, K. C., & Kellen, D. (2018). RT-MPTs: Process models for response-time distributions based on multinomial processing trees with applications to recognition memory. Journal of Mathematical Psychology, 82, 111-130. 111–130. doi: 10.1016/j.jmp.2017.12.003

Hartmann, R., & Klauer, K. C. (2020). Extending RT-MPTs to enable equal process times. Journal of Mathematical Psychology, 96, 102340. doi: 10.1016/j.jmp.2020.102340

Galassi, M., Davies, J., Theiler, J., Gough, B., Jungman, G., Alken, P., . . . Ulerich, R. (2018). GNU scientific library [Computer software]. Retrieved from http://www.gnu.org/software/gsl/