Improving signal/noise ratio

Question

Lubos Motl has fairly complained that the site has had an unfortunate number of questions which are of a pop-sci philsophical flavor (Why is there something rather than nothing? What came before the big bang?), and many homework style problems (What happens to this pully/weight system?).

In order to allow these questions to get answers, but at the same time to allow people to ignore them, I had the following perhaps silly ideas:

1. If a question can be answered in one line and involves nothing beyond Wikipedia facts: like "How to visualize the difference in scale between the nucleus and the atom", perhaps one can answer with one line, "The same difference as between a football field and a small kernel of corn", and close the question?
2. If a question cannot be given a logical positivist formulation, so that the answer is not clearly meaningful, perhaps it can be redirected to a different stackexchange, like philosophy.
3. If a question is homework-ish, if its solution requires no concepts beyond what is standard in the undergraduate curriculum, it can be given the "homework" tag, and omitted from the site list by default (so that only users who choose specifically to see such questions will see them).
4. If a question is pop-sci-ish (involving rubber sheets for General Relativity, or balloons for big bang), then tag it with "pop-sci" and omit it by default.

I know this website is more like "math.stackexchange" than "math.overflow", but it is difficult to navigate the repetitive soup of pop-sci questions and homework. Perhaps it is possible to have one site which functions for all people, but I think that this means that the default questions you see should be the kinds of questions whose answer is not immediately obvious to all physicists.

I am just throwing this out there, I am a new user, and this has probably been discussed before.

Answer 1

• I can assure you that there will be no content omitted by default as this goes against our basic design (making it easy for people to find content). There is a working ignored-tags function for users who do not wish to see certain content.
• You proposal for the use of homework is consistent with the practice on Stack Overflow, and is fine by me, noting that we have already decided against allowing "solve this problem for me" type problems in favor of "explain this physics 101 concept". The use of [homework] is already policy.
• pop-sci is also fine with me.
• We've been trying to chase the worst of the philosophical stuff away. Could you provide some examples of still open question to which you believe this policy should be applied? Indeed, examples would be helpful in all categories.

That said, there has always been some trouble with homework, in that some users object when it is applied to their question---either because they came by this particular pedagogical question in some other context or because they feel it is pejorative. I imagine we can expect the same from pop-sci though I'd suggest that people asking them just get over it.

Answer 2

A late answer, I know, but I feel that it is important. Most of the discussion has been around signal to noise in questions, but the same is true of answers. I've noticed a worrying trend that longer answers are sometimes assumed to be true, and it seems that a few up-votes can generate a runaway effect in this SE.

It may be worthwhile setting some guidelines for a good answer. One thing in particular is that the single most important piece of advice I gave in my years as a TA, especially to the younger ones, was to draw a massive diagram. It sounds silly, but a picture says a thousand words and it takes a few minutes to knock up a diagram in whatever package you favour. Even the most hardcore texts on field theory and the newest research papers use diagrams, so there is no excuse not to have them.

Perhaps there could also be general guidelines on how to formulate a long answer when one is absolutely necessary. This would be useful for the writers and editors to bring things into a format that can be understood and digested with less effort.

Answer 3

I totally agree with Lubos and this seems a general problem to me on scientific SE proposals. I asked this question on AREA51

https://area51.meta.stackexchange.com/q/2733/38770

Signal/Noise Ratio is the perfect term, thats why we have peer-review in top journals. This way important questions and information is highlighted to the scientific community. Why can proven experts like Lubos not have a higher vote count than 30 rep points lawman here? Doenst make any sense to me. I mainly search questions on any SE site by number of votes, but with current system on scientific sites with much less dynamics and user like huge stackoverflow the 1 user 1 vote system doesnt work to filter out these questions. My time is limited...

The question to me is not how to avoid pop-sci-fi questions, you get them anyway. But how to highlight top-notch questions Currently this is not possible by votes sorting nor tag sorting!

Answer 4

I agree with what dmckee said, in particular that you should try the ignored tags feature. But to add to that:

...that the site has had an unfortunate number of questions which are of a pop-sci philsophical flavor (Why is there something rather than nothing? What came before the big bang?)...

These questions get closed pretty quickly, and people seem to be getting the idea that they are not appropriate here. We did have more of a problem of this in the early days of the site, but not so much any more.

...and many homework style problems (What happens to this pully/weight system?).

That's pretty much addressed by the homework policy dmckee linked to. Basically we don't exclude homework questions entirely, because this site is effectively a PR machine for physics (well, that's the goal anyway), but we have stricter requirements than something like Physics Forums in that homework questions asked here should be of some use to other people who aren't doing that particular problem. Right now homework questions do constitute a minority of the questions we get.

If a question can be answered in one line and involves nothing beyond Wikipedia facts: like "How to visualize the difference in scale between the nucleus and the atom", perhaps one can answer with one line, "The same difference as between a football field and a small kernel of corn", and close the question?

If a question is to be closed, it shouldn't be getting answers. The entire point of closing a question is to keep it from being answered (and more generally, from getting attention), so answering and then closing would be pretty useless. With some of these questions there can be useful things to say beyond just what one could look up in an encyclopedia, so I generally favor leaving them open and just not voting them up too much; of course, there aren't that many truly trivial questions on the site. If we had a lot more of them, I'd probably think differently.

If a question cannot be given a logical positivist formulation, so that the answer is not clearly meaningful, perhaps it can be redirected to a different stackexchange, like philosophy.

Yeah, we generally do that already.

If a question is pop-sci-ish (involving rubber sheets for General Relativity, or balloons for big bang), then tag it with "pop-sci" and omit it by default.

I look at these as an opportunity to explain things more accurately, to rid non-experts of some misconceptions they may have picked up due to these analogies. It fits into the role I see for the site as a PR thing.

Answer 5

Lubos Motl has fairly complained that the site has had an unfortunate number of questions which are of a pop-sci philsophical flavor

I'm a little confused about this "fair" complaint when it seems that the entire point of stack exchange sites (altogether) is to improve the signal-to-noise ratio.

I think it's generally correct to note that better questions get higher ratings. If someone wants to spend less of their time on this site while being more effective, why can't they just filter out all the the questions with below 5 votes or something like that? If that sort of solution would be unacceptable, then I have to question if the problem is really a problem of site management or just a conflict of the knowledge levels of people using it.

I've noticed, btw, that many many physics and science oriented questions have been finding their way to the philosophy SE, although you might not notice b/c people have learned to ask over there in the first place.