So, if you want to play around with this, you know just fiddle around with it like I normally ask you to do. Try different multiples of the standard deviation, try adding in more outliers, try adding in outliers that aren't quite as outlier-ish as Donald Trump. You know, just fabricate some extra fake data there and play around with it, see if you can identify those people successfully.

So there you have it! Outliers; pretty simple concept. So, that's an example of identifying outliers by looking at standard deviations, and just looking at the number of standard deviations from the mean or median that you care about. Median is probably a better choice actually, given that the outliers might be skewing the mean in and of themselves, right? So, by using the standard deviation, that's a good way of identifying outliers in a more principled manner than just picking some arbitrary cutoff. Again, you need to decide what the right thing to do is with those outliers. What are you actually trying to measure? Is it appropriate to actually discard them or not? So, keep that in your head!