how to calculate expected value of sample information