Since the above combination gives me the NaN value I tried to solve that as follows.
mu=3.82;
sigma=1.023;
x=normrnd(3.82,1.023,39,1);
skew=skewness(x);
kurt=kurtosis(x);
r=pearsrnd(mu,sigma,skew,kurt,39,1);
Could someone please tell me that method is correct or not?