pc_power_fast: Estimate power for proportion conflict effect using...

In CrumpLab/conflictPower: Simulation based power analysis for adaptive control designs

Estimate power for proportion conflict effect using monte-carlo simulation

Description

Estimate power for proportion conflict effect using monte-carlo simulation

Usage

pc_power_fast(
  subjects,
  mc_c_nmst,
  mc_nc_nmst,
  mnc_c_nmst,
  mnc_nc_nmst,
  num_sims = 1000,
  alpha = 0.05
)

Arguments

subjects	A number for the number of subjects in simulated experiment
mc_c_nmst	A vector for mostly conflict (mc), conflict trials (c) containing the parameters for an ex-gaussian distribution, c(n, mu, sigma, tau), where n is the number of trials.
mc_nc_nmst	A vector for mostly conflict (mc), conflict trials (c) containing the parameters for an ex-gaussian distribution, c(n, mu, sigma, tau), where n is the number of trials.
mnc_c_nmst	A vector for mostly no-conflict (mnc), conflict trials (c) containing the parameters for an ex-gaussian distribution, c(n, mu, sigma, tau), where n is the number of trials.
mnc_nc_nmst	A vector for mostly no-conflict (mnc), conflict trials (c) containing the parameters for an ex-gaussian distribution, c(n, mu, sigma, tau), where n is the number of trials.
num_sims	A number, simulations to run
alpha	A number, alpha criterion

Details

This function uses monte-carlo simulation to determine statistical power associated for detecting a proportion conflict effect, specifically a difference between two conflict effect, typically based on a manipulation of the proportion of conflict and no-conflict trials.

For every simulated experiment, a one sample t-test (two-tailed) is computed, and the p-value is saved. Power is the proportion of simulated experiments that return p-values less than the defined alpha criterion.

Value

A number, power: the proportion of simulated experiments that returned a p-value less than the alpha criterion for the conflict effect

Examples

pc_power_fast(subjects=10,
  mc_c_nmst = c(80,550,100,100),
  mc_nc_nmst = c(20,500,100,100),
  mnc_c_nmst = c(20,570,100,100),
  mnc_nc_nmst = c(80,500,100,100),
  num_sims = 1000,
  alpha = .05)

CrumpLab/conflictPower documentation built on May 17, 2022, 5:48 a.m.