statpsych (version 2.0.0)

random.yx.nonnormal: Generates random bivariate nonnormal scores

Description

Generates a random sample of y scores and x scores from a bivariate nonnormal distribution with a specified population correlation and a specified population mean, standard deviation, skewness, and excess kurtosis for each variable. This function uses the mvrnonnorm function in the semTools package.

For population excess kurtosis values greater than 0, the sample kurtosis values tend to be smaller than the specified population values because of their bias. The bias can be substanital for large excess kurtosis values even in large samples.

Usage

random.yx.nonnormal(n, my, mx, sdy, sdx, skewy, skewx, kury, kurx, cor, dec)

Value

Returns n pairs of y and x scores

Arguments

n

sample size

my

population mean of y scores

mx

population mean of x scores

sdy

population standard deviation of y scores

sdx

population standard deviation of x scores

skewy

population skewness of y scores

skewx

population skewness of x scores

kury

population excess kurtosis of y scores

kurx

population excess kurtosis of x scores

cor

population correlation between x and y

dec

number of decimal points

Examples

random.yx.nonnormal(10, 50, 20, 4, 2, .5, .75, 3, 4, .5, 1)

# Should return:
#    y    x
# 1  47.3 18.8
# 2  52.0 20.3
# 3  51.3 22.3
# 4  50.9 21.3
# 5  55.2 22.0
# 6  53.7 20.1
# 7  46.4 20.0
# 8  54.9 24.7
# 9  48.7 20.7
# 10 44.7 18.6