create_thresholds: Create score thresholds

In taylordunn/gasr: Simulate and Analyze Goal Attainment Scaling Data

Create score thresholds

Description

Creates a set of thresholds which can be used to discretize latent continuous goal scores into "observed" discrete scores.

Usage

create_thresholds(n_levels = 5, score_dist = c("norm", "unif"), centre = 0)

Arguments

n_levels	The number of levels to use. Defaults to the traditional 5 levels.
score_dist	The distribution by which scores should be approximated. The default "norm" will return thresholds which approximate a standard normal distribution. Choosing "unif" will return thresholds which approximate a uniform distribution.
centre	The centre of the normal distribution from which the thresholds are taken.

Details

This function follows the approach introduced by Urach et al. 2019 to get attainment levels from a cumulative standard normal distribution. The score_dist argument will alter the thresholds to return thresholds that will give an approximately normal ("norm") or uniform ("unif") discrete score distribution (assuming a standard normal continuous score distribution). The distribution centre can be shifted to adjust the "difficulty" of goals.

Value

A numeric vector of thresholds with n_levels + 1 values. The first value will be -Inf, and the last Inf, so that extremely small or large values can still be discretized.

References

\insertRef

Urach2019gasr

Examples

create_thresholds()
create_thresholds(n_levels = 3)

# Make goals slightly easier by shifting the thresholds left
create_thresholds(centre = -0.2)
# Or harder by shifting right
create_thresholds(centre = 0.2)

taylordunn/gasr documentation built on April 5, 2022, 1:37 a.m.