Linear Approach to Threshold with Ergodic Rate for Reaction Times

Implements the Linear Approach to Threshold with Ergodic Rate (LATER) model, which predicts distributions of reaction times and summarises them with as little as two free parameters. Allows for easy visualisation and comparison of datasets, along with fitting of datasets using the LATER model.

LATERmodel

The LATERmodel R package is an open-source implementation of Roger Carpenter’s Linear Approach to Threshold with Ergodic Rate (LATER) model (R. Carpenter (1981), Noorani and Carpenter (2016)). This package enables the easy visualisation of reaction time data in LATER’s signature reciprobit space, as well as estimating parameters to fit the model to datasets, comparing raw datasets, comparing fits (i.e., is the dataset better explained by a shift or a swivel?), adding an early component, etc.

This package also includes two canonical datasets digitised from R. H. S. Carpenter and Williams (1995) and Reddi, Asrress, and Carpenter (2003).

Installation

You can install LATERmodel with:

install.packages("LATERmodel")

Example

Load digitised data from Figure 1 in Carpenter and Williams (1995):

library(LATERmodel)

data(carpenter_williams_1995)

Extract data corresponding only to participant a (Figure 1.a):

raw_data <- subset(carpenter_williams_1995, participant == "a")

The data analysis functions within this package require the raw data to first undergo pre-processing using the prepare_data function. We pass our raw_data variable as the argument to the raw_data parameter of prepare_data to perform such pre-processing:

data <- prepare_data(raw_data = raw_data)

Fit each condition individually (no shared parameters between them), and include an early component for all of them:

data_fit <- individual_later_fit(data, with_early_component = TRUE)

Visualise the raw data and the best individual fits:

reciprobit_plot(data, data_fit)

For a more detailed analysis using the LATERmodel R package, see this article.