power4test_by_n: Power By Sample Sizes

In power4mome: Power Analysis for Moderation and Mediation

Power By Sample Sizes

Description

Estimate power for a set of sample sizes.

Usage

power4test_by_n(
  object,
  n = NULL,
  progress = TRUE,
  ...,
  by_seed = NULL,
  by_nrep = NULL,
  save_sim_all = TRUE
)

## S3 method for class 'power4test_by_n'
c(..., sort = TRUE, skip_checking_models = FALSE)

as.power4test_by_n(original_object)

## S3 method for class 'power4test_by_n'
print(x, print_all = FALSE, ...)

Arguments

object	A power4test object, or a power4test_by_n object. If it is a power4test_by_n object, the first element, which is a power4test object, will be used as the value of this argument.
n	A numeric vector of sample sizes for which the simulation will be conducted.
progress	Logical. Whether progress of the simulation will be displayed.
...	For power4test_by_n(), they are arguments to be passed to power4test(). For c.power4test_by_n(), they are power4test_by_n() outputs to be combined together. For the print method of the output of power4test_by_n(), they are arguments to be passed to the print method of the output of power4test() (print.power4test()).
by_seed	If set to a number, it will be used to generate the seeds for each call to power4test(). If NULL, the default, then seeds will still be randomly generated but the results cannot be easily reproduced.
by_nrep	If set to a number, it will be used to generate the number of replications (nrep) for each call to power4test(). If set to a numeric vector of the same length as n, then these are the nrep values for each of the calls, allowing for different numbers of replications for the sample sizes. If NULL, the default, then the original nrep will be used. This argument is used by x_from_power() for efficiency, and is rarely used when calling this function directly.
save_sim_all	If FALSE, the dataset in the power4test objects are not saved, to reduce the size of the output. Default is TRUE.
sort	When combining the objects, whether they will be sorted by sample sizes. Default is TRUE.
skip_checking_models	Whether the check of the data generation model will be checked. Default is TRUE. Should be set to FALSE only when users are certain the they are based on the same model, or when the model is not saved (e.g., save_sim_all set to FALSE when the objects were generated). This argument is used by x_from_power() for efficiency, and is rarely used when calling the c method directly.
original_object	The object to be converted to a power4test_by_n object.
x	The object to be printed.
print_all	If TRUE, all elements in x, that is, the results of all sample sizes examined, will be printed. If FALSE, then only those of the first sample size will be printed.

Details

The function power4test_by_n() regenerates datasets for a set of sample sizes and does the stored tests in each of them.

Optionally, it can also be run on a object with no stored tests. In this case, additional arguments must be set to instruct power4test() on the tests to be conducted.

It is usually used to examine the power over a set of sample sizes.

The c method of power4test_by_n objects is used to combine tests from different runs of power4test_by_n().

The function as.power4test_by_n() is used to convert a power4test object to a power4test_by_n object, if it is not already one. Useful when concatenating power4test objects with power4test_by_n objects.

Value

The function power4test_by_n() returns a power4test_by_n object, which is a list of power4test objects, one for each sample size.

The method c.power4test_by_n() returns a power4test_by_n object with all the elements (tests for different sample sizes) combined.

The function as.power4test_by_n() returns a power4test_by_n object converted from the input object.

The print-method of power4test_by_n objects returns the object invisibly. It is called for its side-effect.

See Also

power4test()

Examples

# Specify the model

model_simple_med <-
"
m ~ x
y ~ m + x
"

# Specify the population values

model_simple_med_es <-
"
m ~ x: m
y ~ m: l
y ~ x: n
"

sim_only <- power4test(nrep = 2,
                       model = model_simple_med,
                       pop_es = model_simple_med_es,
                       n = 100,
                       R = 40,
                       ci_type = "boot",
                       fit_model_args = list(fit_function = "lm"),
                       do_the_test = FALSE,
                       iseed = 1234)

test_out <- power4test(object = sim_only,
                       test_fun = test_indirect_effect,
                       test_args = list(x = "x",
                                        m = "m",
                                        y = "y",
                                        boot_ci = TRUE,
                                        mc_ci = FALSE))

out <- power4test_by_n(test_out,
                       n = c(100, 110, 120))
out_reject <- rejection_rates(out)
out_reject

power4mome documentation built on Sept. 9, 2025, 5:35 p.m.