Adjusting the prior for correlations - matlab code
I have been using the matlab code below to calculate BFs for correlations - I am looking to adjust the prior scale by sqrt(2), but to be honest the maths are confusing me - could anybody provide any insight into what I would need to change and how?
The code I am using is:
%%Original function fro, Sam Pen Du. % Function to be integrated F = @(g,r,n) exp(((n-2)./2).*log(1+g)+(-(n-1)./2).*log(1+(1-r.^2).*g)+(-3./2).*log(g)+-n./(2.*g)); % Bayes factor calculation bf10 = sqrt((n/2)) / gamma(1/2) * integral(@(g) F(g,r,n),0,Inf)