Plot Bootstrap Distribution of Lineup Bias

Plot Bootstrap Distribution of Lineup Bias

Usage

plot_lineup_bias_distribution(
  boot_values,
  target_pos = NULL,
  bins = NULL,
  binwidth = NULL
)

Arguments

boot_values

Numeric vector of bootstrap estimates, as returned by lineup_bias_boot_dist.

target_pos

Optional target position label for title.

bins

Number of histogram bins. Defaults to NULL, which selects the bin width automatically (see Details).

binwidth

Histogram bin width. Defaults to NULL, which selects the bin width automatically (see Details). Takes precedence over bins.

Value

A ggplot object.

Details

Bootstrap estimates of a lineup proportion are counts divided by the number of lineup choices, so they fall on a discrete lattice with spacing 1/n rather than on a continuum. Binning such values at an arbitrary resolution produces a misleading histogram: bins that contain no attainable value are empty regardless of the data, and bins that happen to straddle two attainable values are twice as tall as their neighbours. When neither bins nor binwidth is supplied, the lattice spacing is recovered (from the "n" attribute when present, otherwise from the values themselves) and used as the bin width, so each bar covers exactly one attainable proportion. Estimates that are effectively continuous fall back to 30 bins.

Examples

vec <- round(runif(200, 1, 6))
dist <- lineup_bias_boot_dist(vec, target_pos = 3, k = 6, R = 500)
plot_lineup_bias_distribution(dist, target_pos = 3)