plot_empirical_icc2 {irt}

Plot Empirical Item Characteristic Curve

Description

plot_emprical_icc plots empirical item characteristic curve. Examinees will be put into bins based on their total raw scores and the proportion of examinees who correctly answered an item for each bin will be plotted.

Usage

plot_empirical_icc2(
  resp,
  item,
  bins = 10,
  binwidth = NULL,
  ip = NULL,
  theta = NULL,
  title = "",
  suppress_plot = FALSE,
  x_axis_scale = NULL,
  n_dodge = 1,
  ...
)

Arguments

resp

Response matrix.

item

The column number, column name or the 'ID' of the the item that should be plotted.

bins

An integer larger than 2 representing of ability groups examinees should be grouped into. The default is 10. The maximum value of bins + 1 is the number of possible total scores.

binwidth

If 'theta' scale is used, the binwidth determines the width of each bin of the theta scale. Within each bin, there might be different number of examinees.

ip

An Itempool-class object needs to be provided if expected ICC desired.

theta

A vector of examinee abilities.

title

Title of the plot. The default value is "".

suppress_plot

If FALSE the function will print the plot. If TRUE, function will return the plot object. Default value is FALSE.

x_axis_scale

Set the scale of the x-axis. The default value is NULL. For total score it will be defaulted to "percent".

"percent"

Percent interval.

"number"

Numbers between 1 and bins

"theta"

Theta values equally divided into bins. the middle value of the bin is shown in the x-axis. For example, if bins = 10, the first tick of the x-axis will be the mean of minimum theta value and tenth percentile theta value.

n_dodge

The number of lines the x-axis tick labels should be written to. This is especially useful if the x-axis tick labels overlap with each other. The default value is 1, which means all of the labels are written on the same line.

...

Extra parameters that will pass to geom_line.

Value

Depending on the value of suppress_plot function either prints the empirical item or test characteristic curve or returns the plot object.

Author(s)

Emre Gonulates

Examples

ip <- generate_ip(model = c("3PL", "GRM"), n = 20)
true_theta <- rnorm(2000)
resp <- sim_resp(ip = ip, theta = true_theta)

# Provide item ID
plot_empirical_icc2(resp = resp, item = "Item_5")
# Provide item number
plot_empirical_icc2(resp, item = 3)
# Change x-axis scale
plot_empirical_icc2(resp, item = 3, x_axis_scale = "number")
# Change number of bins and x-axis scale
plot_empirical_icc2(resp, item = 3, bins = 11, x_axis_scale = "theta")
# Use bin width
plot_empirical_icc2(resp, item = 3, binwidth = 2)
# Use theta scores instead of raw scores
plot_empirical_icc2(resp, item = 3, binwidth = .2, ip = ip,
                    theta = true_theta)

# A GRM item
plot_empirical_icc2(resp, item = 4)
plot_empirical_icc2(resp, item = 4, x_axis_scale = "percent")
plot_empirical_icc2(resp, item = 4, x_axis_scale = "number")
plot_empirical_icc2(resp, item = 4, binwidth = 4)
# Use raw score and custom binwidth
plot_empirical_icc2(resp, item = 4, x_axis_scale = "percent", binwidth = 4)
# Use theta score
plot_empirical_icc2(resp, item = 4, binwidth = .2, ip = ip,
                    theta = true_theta)
# Add arguments for 'geom_line'
plot_empirical_icc2(resp, item = 4, binwidth = .2, ip = ip,
                    theta = true_theta, size = 1, alpha = .25)