optimal_p {ionr}

Find an optimal p-value for SONE

Description

a wrapper that runs the maximum and minimum scenarios using scenario_sim and provides the optimal p -value

Usage

optimal_p(sizes, n_sim = 100, plotting = "", n_indicators = 8,
  to_min = (round((n_indicators/2), 0)) - 1, ...)

Arguments

sizes

An array of sample sizes to be simulated. Can be single value.

n_sim

number of simulations. 1000 is a start, 10000 was used in paper, but takes a long time

plotting

Plots the result with optimal_p_out. Defaults to ”. Possible options: ” - no plot; 'yes' - a regular plot; 'file' – writes the plot to a tiff file in working directory. If sizes is a single value, plotting is disabled.

n_indicators

How many many indicators are there in a scale. The package is tested with 8 indicators (default), but should work with other number.

to_min

How many indicators relate to the outcome in the lack of ION condition. In optimal_p defaults to (round((n_indicators/2),0)) - 1), i.e close to half the number of indicators.

...

further tweaking of the scale simulator, see scale_sim for details.

Value

Returns the P criterion, as well as the p values for max and min scenario for each sample size. If min pvalue > max pvalue, then p criterion is NA.

Examples

set.seed(466)
n_sim=100
ptm <- proc.time()
a=optimal_p(sizes=750, n_sim=n_sim, n_indicators=8, cor_to_outcome=0.25)
stp=proc.time() - ptm
print(paste("Currently elapsed:",round(stp[3],1)))
print(paste("Time estimate for n_sim=5000:",round(stp[3]*5000/n_sim,1)))