nonparametric method to calculate the probability how alike two samples are I have two samples with each couple of hunderd observations. I want to calculate a probabilty how much they look alike. I'm aware of tests like kolmogorov smirnov but I don't think I need this. I don't want to test if they come from different distributions. I'm actually want to show that they come from the same (continuous) distribution.
I looked around a bit on the internet, but I'm not sure how to approach this. Do I need bootstrapping? Also, I'm using R, so pointers to usefull packages are welcome.
thanks in advance
update:
As commented below, you can not really prove there are from a simmilar distribution. So the question would rather be, how can I assign a degree of belief that they are from the same distribution?
 A: Without a bit more detail it's difficult to know for sure what you need, but assuming you have two independent samples that are continuous data, a Mann-Whitney U test is a statistical significance test to test if two populations are the same.
There's an article on carrying out the test here:
http://www.r-tutor.com/elementary-statistics/non-parametric-methods/mann-whitney-wilcoxon-test
You can do the test with the stats package, so if you need to install it:
install.packages("stats"); library(stats)

Then, assuming your data looks something like this:
s1 <- runif(10); s2 <- runif(10)
View(d)

I.e. samples are separate vectors, you can run a version of the test with the following basic syntax:
wilcox.test(s1, s2)

If, instead, you have a dataframe that has all the observations in one column and another column to specify the sample the observation belongs to:
s <- runif(10)
f <- c(1, 1, 1, 2, 2)
d <- data.frame(s, f)
View(d)

then you would use the ~ syntax:
wilcox.test(d$s ~ d$f)

Hope this is helpful. Don't forget to mark your question answered if this solves your problem, or post further details if it doesn't!
