I wrote up a general approach to calculating quantiles here. Median is of course an example of a quantile. Given how your data is structured I had to tweak that slightly. BTW, I make it that the median is 5, not 3 (I haven't actually checked the data, but I'm fairly confident in that calculation).
For calculating the mode, you can use this table calculation:
BTW - I see you're still on 6.0 (so I did the attached in 6.0), but it's well worth upgrading to 7.0.
Edit: I should have pointed out that although I called the calculated field "Median", it is of course a general quantile, so you can just change the parameter p from 0.5 to 0.95 to get the 95th percentile (or whatever).
distributiontest_rl.twbx 159.1 KB
Wow - amazing. Thank you very much Richard! We are in process of upgrading our server to 7 which is why I haven't upgraded my desktop yet. We're definitely on board though!