Skip to contents

Getting started with actdata

Installation

You can install the development version of actdata from GitHub with:

# install.packages("devtools")
devtools::install_github("ahcombs/actdata")

Available data

The primary purpose of actdata is to be a repository for publicly available affect control theory data sets in a way that makes them easy to incorporate into a reproducible, transparent R-based analysis workflow.

Two types of data are available.

  • Dictionaries provide empirically measured evaluation, potency, and activity ratings for terms. Two types of dictionary data are available in this package. The first is summary statistics–the means and in some cases standard deviation and covariance for term ratings provided by respondents. The second type of data, made publicly available for the first time through this package, are respondent-level ratings. See the dictionaries help page for more information on both types of dictionary data.
  • Equations provide coefficients that have been estimated for various equations that can be used to calculate affective responses to situations. See the equations help page for more information, and run eqn_info() to see metadata for all available equation sets.

Dictionaries have been standardized to make it easy to compare terms across them. See the standardization details page for more information on these processes. There is also a worked example of the kind of cross-dictionary, over-time analysis that this standardization makes relatively straightforward.

Synergy with analysis programs

Often, researchers who use affect control theory data sets do so to simulate interaction. Several other tools are available to run these simulations. actdata provides functions that make it easy to access data with the package, modify it in whatever way necessary, and convert it to formats that are easy to import to other programs that can be used for analysis.

actdata is designed to work alongside three simulation softwares: Interact Java, inteRact, and BayesACT and its R wrapper bayesactR. See the data export help page for more information.