Department of Pure and Applied Mathematics, Ladoke Akintola University of Technology, P. M. B. 4000, Ogbomoso, Oyo State, Nigeria. E-mail:
In Monte Carlo studies that involves multicollinearity, cointegration or any other form of correlation problems; it is often desired to generate variables that exhibit different degree of intercorrelations. In this paper, we provide equations that generate normally distributed random variables with desired intercorrelation matrix. To investigate the efficacy of the equations, we examined the average estimated correlation values and the number of type 1 error of Fisher's logarithm transformation statistic at some specified intercorrelation matrices in 1000 simulation trials. Results show that the average estimated correlation values approximate the true correlation value; and that the type 1 error rates are close to the predetermined alpha level (0.05).
Type 1 error, intercorrelation matrix, Monte Carlo simulation
