Closed macartan closed 5 years ago
In the block designer there is something like this:
u_0 = rnorm(N) * sd_i_0, u_1 = rnorm(n = N, mean = rho * u_0, sd = sqrt(1 - rho^2)) * sd_i_1
I believe this is incorrect: the rho approch we use here works properly if the base variable has unit s.d. but here it does not
I think we need this:
u_0 = rnorm(N) * sd_i_0, u_1 = rnorm(n = N, mean = rho * u_0/sd_i_0, sd = sqrt(1 - rho^2)) * sd_i_1
@medinali possible to check this and also see if we do this elsewhere?
you're right!
I think having u_1 = rnorm(n = N, mean = rho * scale(u_0), sd = sqrt(1 - rho^2)) * sd_i_1
u_1 = rnorm(n = N, mean = rho * scale(u_0), sd = sqrt(1 - rho^2)) * sd_i_1
would be neater.
I'll correct and check if this also in other designs.
In the block designer there is something like this:
I believe this is incorrect: the rho approch we use here works properly if the base variable has unit s.d. but here it does not
I think we need this:
@medinali possible to check this and also see if we do this elsewhere?