[rossdavidh]

"statistics is something of a blind spot for programmers, who tend to think of themselves as numerically proficient but often dismiss statistics as unimportant “stamp collecting” for people who can’t do “real maths”" ...I've literally never heard anything like this from any programmer (or anyone else, actually). Has anyone else heard a programmer express such an attitude towards statistics? Mostly I've heard people say that statistics is hard, or counter-intuitive.

[oliwarner]

That's definitely a viewpoint amongst other branches of mathematics. Statistics isn't too bad, but applied statistics is viewed as a crap shoot of confounding human factors, widely practised by charlatans trying to prove their brand of science by abusing numbers.

[screye]

Isn't all of machine learning, economics, industrial engineering and statistically social science a form of applied statistics ?

That's a lot of fields to dismiss by simply handwaving.

[oliwarner]

Yes. These are exactly the sorts of soft science a pure mathematician would snub.

It's not, I think, that they're without utility, it's that the science involved in discovery is often accidental or observational.

[haloux]

Your sample population sounds skewed ;). There are _many_ intersections of statistics in computer science. Off the top of my head: data analytics processing, anything in the realm of big data, computational theory, security timing attacks, industrial computation in the natural sciences and so on.

[rossdavidh]

Absolutely. I've just never heard the attitude that statistics was for people who can't do "real math", or anything similar. But, my sample population is undoubtedly skewed in some way or another.

[padthai]

I had. Algebra and analysis are still considered by some people harder and "purer" than let's say combinatorics, applied or statistical mathematics.