splithalf {splithalf}

Internal consistency of task measures via a permutation split-half reliability approach

Description

This function calculates split half reliability estimates via a permutation approach for a wide range of tasks. Most of the user inputs relate to the variables in the dataset splithalf needs to read in order to estimate reliability. Currently supports response time and accuracy outcomes, for several scoring methods: average, difference, difference of difference scores, and a DPrime development. The (unofficial) version name is "This function gives me the power to fight like a crow"

Usage

splithalf(
  data,
  outcome = "RT",
  score = "difference",
  conditionlist = FALSE,
  halftype = "random",
  permutations = 5000,
  var.RT = "latency",
  var.ACC = "accuracy",
  var.condition = FALSE,
  var.participant = "subject",
  var.compare = "congruency",
  compare1 = "Congruent",
  compare2 = "Incongruent",
  average = "mean",
  plot = FALSE,
  round.to = 2,
  check = TRUE
)

Arguments

data

specifies the raw dataset to be processed

outcome

indicates the type of data to be processed, e.g. "RT" or "accuracy"

score

indicates how the outcome score is calculated, e.g. most commonly the difference score between two trial types. Can be "average", "difference", "difference_of_difference", and "DPrime"

conditionlist

sets conditions/blocks to be processed

halftype

specifies the split method; "oddeven", "halfs", or "random"

permutations

specifies the number of random splits to run - 5000 is good

var.RT

specifies the RT variable name in data

var.ACC

specific the accuracy variable name in data

var.condition

specifies the condition variable name in data - if not specified then splithalf will treat all trials as one condition

var.participant

specifies the subject variable name in data

var.compare

specifies the variable that is used to calculate difference scores (e.g. including congruent and incongruent trials)

compare1

specifies the first trial type to be compared (e.g. congruent trials)

compare2

specifies the second trial type to be compared (e.g. incongruent trials)

average

use "mean" or "median" to calculate average scores?

plot

logical value giving the option to visualise the estimates in a raincloud plot. defaults to FALSE

round.to

sets the number of decimals to round the estimates to defaults to 2

check

runs several checks of the data to detect participants/conditions/trialtypes with too few trials to run splithalf

Value

Returns a data frame containing permutation based split-half reliability estimates

splithalf is the raw estimate of the bias index

spearmanbrown is the spearman-brown corrected estimate of the bias index

Warning: If there are missing data (e.g one condition data missing for one participant) output will include details of the missing data and return a dataframe containing the NA data. Warnings will be displayed in the console.

Examples

## Not run: 
## see online documentation for full examples
https://github.com/sdparsons/splithalf
## example simulated data
n_participants = 60 ## sample size
n_trials = 80
n_blocks = 2
sim_data <- data.frame(participant_number = rep(1:n_participants,
                       each = n_blocks * n_trials),
                       trial_number = rep(1:n_trials,
                       times = n_blocks * n_participants),
                       block_name = rep(c("A","B"),
                       each = n_trials,
                       length.out = n_participants * n_trials * n_blocks),
                       trial_type = rep(c("congruent","incongruent"),
                       length.out = n_participants * n_trials * n_blocks),
                       RT = rnorm(n_participants * n_trials * n_blocks,
                       500,
                       200),
                       ACC = 1)

## example run of splithalf on a difference score
splithalf(data = sim_data,
          outcome = "RT",
          score = "difference",
          conditionlist = c("A", "B"),
          halftype = "random",
          permutations = 5000,
          var.RT = "RT",
          var.condition = "block_name",
          var.participant = "participant_number",
          var.compare = "trial_type",
          compare1 = "congruent",
          compare2 = "incongruent",
          average = "mean",
          plot = TRUE)

## example run of splithalf on an average score
splithalf(data = sim_data,
          outcome = "RT",
          score = "average",
          conditionlist = c("A", "B"),
          halftype = "random",
          permutations = 5000,
          var.RT = "RT",
          var.condition = "block_name",
          var.participant = "participant_number",
          average = "mean")

## example run of splithalf on a difference of differences score
splithalf(data = sim_data,
          outcome = "RT",
          score = "difference_of_difference",
          conditionlist = c("A", "B"),
          halftype = "random",
          permutations = 5000,
          var.RT = "RT",
          var.condition = "block_name",
          var.participant = "participant_number",
          var.compare = "trial_type",
          compare1 = "congruent",
          compare2 = "incongruent",
          average = "mean")


## End(Not run)