I am really hoping that someone can help me out on this since I am not a statistician. I am trying to determine the statistical significance (or some other measure of significance) of certain values. If the distribution of values was normal then I could use a Z-Score and it tells me what I need to know (as I understand it, how many stdev away from the mean a given data point is). However, if I understand it correctly, I cannot use the Z-Score when dealing with skewed distributions. For example, if have 160 data points, 156 of which are 0, 2 are 20 and 2 are 60 then how can I assign some statistical weight to those non-zero data points? Any help would be appreciated. Thanks!