Determining what is normal for laboratory values is not as easy as it might seem. First, there is the problem of a reference population. Should one use healthy young adults or a group representing the population from which diseased persons come? Then, once the statistics are available, how should they be handled? These are some of the questions considered here.