Dear All,
I have known for sometime that, if X is standard normal, then
E(X|X>t) = f(t)/[1-F(t)]
where f() is the density function and F(t) is the distribution function of
the standard norml.
My query is: if X and Y are bivariate normal with means 0, variances 1, and
correlation r, is there a similar expression for E(XY|X>t,Y>u)?
By analogy with the above simple expression I suspect that this could be
expressed in terms of the bivariate density and distribution functions.
Perhaps this is just being hopeful, but I wonder if anyone knows this answer.
With best wishes
Pak Sham
Pak C Sham
Division of Psychological Medicine and SGDP Research Centre
Box P080
Institute of Psychiatry
De Crespigny Park
London, SE5 8AF
Tel: 020 7848 0892
Fax: 020 7848 0866
|