find_critical_pos: Find the critical point of stability

In fastpos: Finds the Critical Sequential Point of Stability for a Pearson Correlation

Find the critical point of stability

Description

Run simulations for one or several population correlations and return the critical points of stability (POS). The critical point of stability is the sample size at which a certain percentage of studies will fall into an a priori specified interval and stay in this interval if the sample size is increased further.

Usage

find_critical_pos(
  rho,
  precision_absolute = 0.1,
  confidence_levels = c(0.8, 0.9, 0.95),
  sample_size_min = 20,
  sample_size_max = 1000,
  n_studies = 10000,
  n_cores = 1,
  pop_size = 1e+06,
  replace = TRUE,
  precision_relative = NA,
  lower_limit = NA,
  upper_limit = NA,
  progress = show_progress(),
  precision = lifecycle::deprecated(),
  precision_rel = lifecycle::deprecated(),
  rhos = lifecycle::deprecated()
)

Arguments

rho	Vector of population correlations (can also be a single correlation).
precision_absolute	Precision around the correlation which is acceptable (defaults to 0.1). The precision will determine the corridor of stability which is just rho+-precision. Can be a single value or a vector (different values for different rhos).
confidence_levels	Confidence levels for point of stability. This corresponds to the quantile of the distribution of all found critical sample sizes (defaults to c(.8, .9, .95)). A single value can also be used. Note that this value is fixed for all rhos! You cannot specify different levels for different rhos.
sample_size_min	Minimum sample size for each study (defaults to 20). A vector can be used (different values for different rhos).
sample_size_max	Maximum sample size for each study (defaults to 1e3). A vector can be used (different values for different rhos). If you get a warning that the corridor of stability was not reached, you should increase this value. But note that this will increase the time for the simulation.
n_studies	Number of studies to run for each rho (defaults to 1e4). A vector can be used (different values for different rhos).
n_cores	Number of cores to use for simulation. Defaults to 1. Under Windows only 1 core is supported because forking is used.
pop_size	Population size (defaults to 1e6). This is the size of the population from which value pairs for correlations are drawn. This value should usually not be decreased as it can lead to less accurate results.
replace	Whether drawing samples is with replacement or not. Default is TRUE, which usually should not be changed. This parameter is mainly of interest for researchers studying the method in more detail. A vector can be used (different values for different rhos).
precision_relative	Relative precision around the correlation (rho+-rho*precision), if set, it will overwrite precision_absolute. A vector can be used (different values for different rhos).
lower_limit	Lower limit of corridor, overrides precision parameters. A vector can be used (different values for different rhos). If used, upper_limit must also be set.
upper_limit	Upper limit of corridor, overrides precision parameters. A vector can be used (different values for different rhos). If used, lower_limit must also be set.
progress	Should progress bar be displayed? Logical, default is to show progress when run in interactive mode.
precision	, use precision_absolute instead
precision_rel	, use precision_relative instead
rhos	, use rho instead

Value

A data frame containing all the above information, as well as the critical points of stability.

The critical points of stability follow directly after the first column (rho) and are named pos.confidence-level, e.g. pos.80, pos.90, pos.95 for the default confidence levels.

Examples

find_critical_pos(rho = 0.5, n_studies = 1e3)
find_critical_pos(rho = c(0.4, 0.5), n_studies = 1e3)

fastpos documentation built on Aug. 16, 2022, 1:06 a.m.