pair.generator produces sets of pairs for use in pairwise comparisons projects. This package is designed for use in educational assessment, where students' performances are judged against one another, two at a time. The end result of this process is a scale in which performances are located on an ability continuum. pair.generator conveniently forms pairs and provides diagnostics for the design stage.
Install from GitHub by cloning the package, or install using remotes.
# load remotes packages to install
library(remotes)
remotes:install_github("KenBredemeyer/pair.generator")
Generate pairs with pairs_generate(). This function returns a data.frame which includes the variables left and right, for
presenting performances to judges on the left and right sides. The package has some built in data sets, including data_standard,
for experimenting with the functions.
# load package
library(pair.generator)
# set rng seed for reproducability
set.seed(1)
# generate pairs, with each performance included 10 times on average
pairs <- pairs_generate(data_standard, av_inclusions = 10, inclusion_tolerance = 1)
# plot the number of times each performance is included in the set of pairs
pairs_plot_inclusions(pairs)
As with any R package, get a full list of exported fuctions with lsf.str(), and access help with ?.
# see all exported functions in pair.generator
lsf.str("package:pair.generator")
# access help on chain(), for chaining pairs.
?chain