cat_sim_fast {irt}

Computerized Adaptive Test (CAT) Simulation (Parallel Computing)

Description

cat_sim_fast function simulates computerized adaptive test (CAT) for one or many simulees. This function uses parallel computing, so, for large number of simulees, it might be significantly faster than cat_sim function.

Usage

cat_sim_fast(true_ability, cd, verbose = -1, n_cores = NULL)

Arguments

true_ability

True ability vector to generate item responses.

cd

A cat_design object that is created by function create_cat_design.

verbose

This is an integer that will print the stage of the test. For example, if the value verbose = 10, a message will be printed at each tenth iteration of the cat_simulation. Default value is -1, where no message will be printed. If the value is 0, only the start time and end time of the simulation will be printed.

n_cores

an integer specifying the number of cores to be used. The value should be 1 or larger. The default is NULL where the maximum number of cores of the processor will be used.

Value

If the length of true_ability vector is one a "cat_output" class output will be returned. This is a list containing following elements:

true_ability

True ability (theta) value to generate item responses.

est_history

A list where each element represent a step of the CAT test. It has following elements:

est_before

The estimated ability before the administration of the item.

se_before

The standard error of the estimated ability before the administration of the item.

testlet

TRUE if the item belongs to a testlet.

item

Item-class object that is administered at this step.

resp

The simulated response of the simulee for the item administered at this step using simulee's true_ability value.

est_after

The estimated ability after the administration of the item.

se_after

The standard error of the estimated ability after the administration of the item.

If the length of the true_ability is more than 1, a list of cat_output objects will be returned for each value of true_ability.

Author(s)

Emre Gonulates

See Also

create_cat_design

Examples

cd <- create_cat_design(ip = generate_ip(n = 30),
                        termination_rule = c('max_item'),
                        termination_par = list(max_item = 7))
cat_sim_fast(true_ability = rnorm(1), cd = cd, n_cores = 1)

cat_sim_fast(true_ability = rnorm(2), cd = cd, n_cores = 1)