I don't have it in matlab but mathematica. Use the excel spreadsheet in the paper to conduct the test in any language. Use email if necessary

You refered in the paper that you have used matlab ...Iam working on R and I have some difficulties for implementing the test

I have stopped working on Matlab but I still have the power tests on my old computer in my office. I will have to check this next week. Hopefully.

I found my programs on my  drive

Mathematica program + data example 50 gamma

Matlab programs normality test + weight computation + data 100 gamma

In the paper you have an example derived from Excel

It should suffice to allow you programming in any language

thanks alot for ur help but i want to ask two questions:

1. why s2y = var(y)*((n-1)/n) ... is it the sum of squares so it should be sum(x-mean(x))^2 .

2. how did u calculate the power of ols using monte carlo simulation ..

answer 1 :s^2 is the variance estimate

answer 2: explained in the paper. You sample a distribution 1 million times and compute the weights and check if they pass the test

plz, prof. haim how did u estimate the critical values for OLS test at different sample sizes when calculating type 1 error rate at 5% level of significance ?