I have many (20) normal distributions of the probability of a certain event occurring. Each of these 20 has n>1000.I have an observed proportion from a separate sample that I want to compare to calculate a p-value. However, I don't know which of these 20 it came from. How can I evaluate against all of these?
I do know that each my observed proportion had to have come from one of those 24 distributions. I also know the probability of it being each of those distributions. (ie. Dist1 - 5%, Dist2- 3%, etc).
I tried simply combining all 20 and creating an overall distribution, but the means of each of the 20 are quite distinct, so it ended up just looking like many normal dists in random series on the same axis.
Any ideas?
