In reply to Timmd:
> It's like the term 'standard deviation', it implies there's a none standard kind, when apparently it's just a term.
I’m not a statistician, but aiui (this is probably not quite right, but I think in the right ballpark) it’s called that essentially because it is a way of expressing deviation that gives us a standardised value by which we can work out the probability of different values occurring in the dataset.
Because it’s a standard/standardised calculation, if someone tells you that a value is 3 standard deviations from the mean you can understand how much of an outlier that value is without needing to know much else about the dataset. If I just told you that a value is 7 hoojamiflips from the mean, its still expressing deviation but it’s meaningless information without a lot more context.
Edit: cue any number of corrections of the specifics above from the more mathematically inclined on here, but for example: 3 SDs from the mean means the same thing (with the assumption of normally distributed data at least) regardless of whether we are measuring someone’s IQ, weight, height or toenail length. Thus “standard”.
Interesting question though, not something I’ve thought about before.
Post edited at 10:48