@ssdecontrol Leaving aside the "repeatedly" part, I find answering on CV is typically a great deal more work than on SO. Often an SO answer is a couple of lines of code, or a two sentence explanation. That's rarely going to answer the question on CV, where typically a page of text is needed (and often a diagram or something similar is needed to really convey the concept). I feel like my average post here is about two paragraphs for every sentence in my average SO post.
It's fairly common that I'll be reading a research paper or conducting a simulation to answer a question on CV, and often several of both. I don't think I've ever done anything so involved for SO; occasionally I check the help on a function to make sure it works the way I think it does.
I also find questions on math.SE tend to have shorter answers than here; they're usually more specific and focused.
Votes per keystroke tend to be relatively low too; you have to earn your reputation. There are some other SE-sites where I might have to work as hard to post a good answer, but I'll reap many times the votes for my effort.
It's not just lengthy answers, reading papers, conducting simulations and making diagrams. Many of my posts have carefully constructed counterexamples; these don't just grow on trees, sometimes they take a good deal of thought and sometimes algebra or experimentation to find a case that does exactly what is needed to illustrate the issue.
Fortunately, a good counterexample can often be adapted to answering several kinds of questions, so there's at least some re-use value for some of them.
There are some answers here where I have but a great deal of thought into and gotten very little response to them. But there are also some where that I just wrote what came to mind and I got a lot of credit for them.
It relates to a story about a committee reviewing plans for a nuclear power plant spending most of its time discussing the bicycle shed.
It comes from a book, Parkinson, C. Northcote (1958). Parkinson's Law, or the Pursuit of Progress.
Which book I read ... I think in my late teens , many decades past
I've seen it first hand in reports of local council meetings, where there will be gigantic discussions of trivia while hugely important issues pass without comment.
The book itself has some nifty insights (the eponymous law of the title relates to the adage that work expands to fill the time available); I read it at about the same time I read "The Peter Principle" (the eponymous principle there is that in management, people are promoted to their level of incompetence) and a couple of other works in similar vein.
When I read them I found them entertaining but possibly deliberately cynical for humor value. Now I've had lots of experience of management, these authors appear to be romantic pollyannas who underestimate the psychopathology of a large number of managers. I've come to the conclusion that anyone who seeks such a position should on no account be allowed to hold it.
The best managers I've encountered have always been people who really didn't want to be in charge but felt there was no decent alternative/took it on because the alternatives were all worse.
Not that I think not wanting to do it is necessarily a sign of a good manager, so much as wanting to do it seems to be a marker of one of a few very particular kinds of people.
hey @Glen_b, about your comment on my question (here: stats.stackexchange.com/questions/176636/…) I wanna ask a question. But I'm sorry if I sound too "green" in all this (as you may have suspected from my own question there).
I wanna ask the following:
What kind of test should I take if I want to compare medians of two populations, and they do not share a similar distribution?
@Sosi: Ask technical questions in the forum please; that is what it's for. You would need to flesh out that question and even more importantly check out what has already been posted on such issues. Most methods for comparing medians are rank-based and do not make assumptions about precise distributions, so there are already probably dozens of threads to explore.
@Glen_b Nolo episcopari is an old principle. Supposedly, those worthy of being bishops should not be those wishing to take on the role. The practice was and is often different.
@Sosi yes, probably best to ask a new question (assuming a search doesn't turn up an answer already - it probably will), but you don't necessarily need the same shape in the data to use the WMW as a test of equality of medians as long as your alternative doesn't assume the same shape; for example, it would be perfectly valid to use a WMW to test equality of location of samples from beta distributions which were equal under the null but had a larger α (alpha) parameter under the alternative;
in that case you wouldn't necessarily see the same shape unless the null was actually true.
That can be a test of equality of medians against an alternative of inequality of medians, but not of median-shift; the whole shape changes under the alternative.
Indeed, with the WMW two distributions don't have to have the same shape under the null, but it's harder to guarantee it's really a test of equality of medians then.
Thanks @NickCox and @Glen_b. I am now searching that scenario on the website, even though I'm not even sure that may happen in my data I'm just curious
(and sorry to have brought the question here. In the Mathematica forum people usually use the chat for quick/small questions and I was used to that)
@ssdecontrol Sometimes I post something and am confident that there will be a fair number of votes, but otherwise there is no real relation between your own evaluation of an answer and how many votes you get. I reckon that in total almost everyone gets about the right reputation, but sometimes for the wrong specific reasons. Otherwise I would probably not stick around.
@Sosi asking on the main site means you're not just relying on one or two people at one particular time for an answer (answers might trickle in over days or longer), and it means many other people get to benefit from the question and answer(s) possibly many months later.
Indeed the fact that it's a permanent searchable repository is one of the reasons I expend effort here rather than on other forums; I don't have to keep answering the same question over and over -- I can write one good answer and point to it.
So all the benefit seems to point to asking on the main site.
@ssdecontrol (or anyone who cares...) My earlier comment about post length vs SO is borne out by the data -- comparing Q1,Q2,Q3 for both, answer length is nearly twice as long on CV.
This is a minor suggestion relating to Meta: need some help from users with privilege to suggest (or make) the 'formatting' tag to be a synonym of the 'markdown' tag. I have checked out questions tagged with 'formatting' and the newest ones refer to 'markdown'. The oldest one is more about 'editing', but it was also tag 'formatting'. Tks.
@ssdecontrol: Hmm, I am actually not super happy with that particular thread :) Too many answers, some are really not that good or not that to the point. Top answer is... good in its way, but IMHO does not really answer the question (rather, it answers a different question), etc.
@ssdecontrol, my understanding is that two of them are relating to markdown issues of the StackExchange system, the oldest one is about editing (which include the use of markdown). Maybe I lack explanation in my previous comment, the 'format' tag is ambiguous; in the main Meta SE site it is a synonym of 'markdown'. So, I checked if it would make sense following the same pattern from there in our meta site, and I concluded it would. But it is true the words are not mutually exclusive.
