It is always best to use raw scores for computing the FrequencyTable.
They aren't always available - in that case, this function can be used
to simulate the distribution given its descriptive statistics.
This simulation should be always treated as an estimate.
The distribution is generated using the Fleishmann method from
SimMultiCorrData::nonnormvar1() function. The
SimMultiCorrData package needs to be installed.
SimFrequencyTable(min, max, M, SD, skew = 0, kurt = 3, n = 10000, seed = NULL)FrequencyTable object created with simulated data. Consists of:
table: data.frame with number of observations (n), frequency in sample
(freq), quantile (quan) and normalized Z-score (Z) for each point in
raw score
status: list containing the total number of simulated observations (n)
and information about raw scores range completion (range): complete or incomplete
minimum value of raw score
maximum value of raw score
mean of the raw scores distribution
standard deviation of the raw scores distribution
skewness of the raw scores distribution. Defaults to 0 for
normal distribution
kurtosis of the raw scores distribution. Defaults to 3 for
normal distribution
number of observations to simulate. Defaults to 10000, but greater
values could be used to generate better estimates. Final number of observations
in the generated Frequency Table may be less - all values lower than min and
higher than max are filtered out.
the seed value for random number generation