monte1: Simulate Multivariate Non-normal Data by Vale & Maurelli (1983) Method

Description

Function for simulating multivariate nonnormal data by the methods described by Fleishman (1978) and Vale & Maurelli (1983).

Usage

monte1(seed, nvar, nsub, cormat, skewvec, kurtvec)

Value

data: The simulated data.
call: The call.
nsub: Number of subjects.
nvar: Number of variables.
cormat: The desired correlation matrix.
skewvec: The desired indicator skewness values.
kurtvec: The desired indicator kurtosis values.
seed: The random number seed.

Arguments

seed: An integer to be used as the random number seed.
nvar: Number of variables to simulate.
nsub: Number of simulated subjects (response vectors).
cormat: The desired correlation matrix.
skewvec: A vector of indicator skewness values.
kurtvec: A vector of indicator kurtosis values.

Author

Niels Waller

References

Fleishman, A. I (1978). A method for simulating non-normal distributions. Psychometrika, 43, 521-532.

Olvera Astivia, O. L. & Zumbo, B. D. (2018). On the solution multiplicity of the Fleishman method and its impact in simulation studies. British Journal of Mathematical and Statistical Psychology, 71 (3), 437-458.

Vale, D. C., & Maurelli, V. A. (1983). Simulating multivariate nonnormal distributions. Psychometrika, 48, 465-471.

Examples

Run this code



## Generate dimensional data for 4 variables. 
## All correlations = .60; all variable
## skewness = 1.75; 
## all variable kurtosis = 3.75
 
cormat <- matrix(.60,4,4)
diag(cormat) <- 1

nontaxon.dat <- monte1(seed = 123, nsub = 100000, nvar = 4, skewvec = rep(1.75, 4),
               kurtvec = rep(3.75, 4), cormat = cormat)
 
print(cor(nontaxon.dat$data), digits = 3)
print(apply(nontaxon.dat$data, 2, skew), digits = 3)
print(apply(nontaxon.dat$data, 2, kurt), digits = 3)

Run the code above in your browser using DataLab