We seem to be conflating two variables: whether one can smell the difference, and whose urine it is. The number of people, even spouses, willing to sniff each others' urine to test these separately must be too small to form a meaningful data sample.