> The groups found by this sort of "mycelial" finding are of course not > statistically valid samples. > Hi Birrell, While the sampling method you employ does not permit estimates of sampling error when making inferences to the total population, it can be very useful. One of the things to look for is the additional variance provided by each additional case. When additional cases provided by this sampling method provide no new variance, I think you could make the case that your sample variance may be a pretty good reflection of the population variance. Just a thought. Dave Nasatir %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%