7

Following up a recent meta post wondering about the direction this site is taking, I wanted to discuss the possibility of changing one of our current habit, which is to close "technical" questions as off-topic.

While some are clear cuts, say someone asking about the content of an algorithm, some are actually effectively about the academic process. One example are questions pertaining to good practice when designing a study. We might then benefit from a "study-design" tag.

A few examples:

Creating a research questionnaire with repeating sections

What tradeoffs are there between internal and external validity?

Is it possible to "validate" a simulator without comparing it with the actual tool?

Yes these questions are field-specific, but a lot of the existing one are as well. Apparently we're capable of dealing with that.

with the pool of clever and successful academics participating to this site it would be great to leverage that knowledge. Also, it would be a refreshing change to all the questions along the line of "someone stole my work/my GPA is not so great how do I still get into grad school/something mildly entertaining happened in class" for which we now have a comprehensive set of answers.

In a typical meta fashion, vote this question up if you're in favor, down if you're against.

  • We already have an experiment-design tag. Are you referring to questions like those? Or more subject-specific? Can you give an example of a closed question that you think should be acceptable? – ff524 Jan 16 '17 at 19:05
14

My feeling is No. There are a couple reasons I think this way:

  1. Academia is, in many ways, a site about the practice of academia, the process of being an academic, etc. rather than highly technical questions about narrow aspects of research. I like that, and there's not actually firm evidence in the question you link in your question that Academia isn't "sustainable" with that scope. Stepping away from that has the potential to dilute what people are interested in coming here for with technical questions, effectively decreasing the signal:noise ratio. This has happened on other sites I have been on - for example, when CrossValidated took in the machine learning questions from a failed site, my interest dropped pretty dramatically as I had to do far more sifting to find questions I was interested in. This risk is, IMO, higher on Academia, as you're not just talking about "one adjacent" fields, but all fields.
  2. Academia does not have good field coverage. There are lots of people in CS, Applied Math, etc. There are some in the biological sciences. There are vanishingly few in say, Medieval History. This is important not only for "can someone answer this" but for field specific norms. An economist and I might be able to answer the same question, but we'd answer it differently, and potentially miss field-specific nuances (for example, epidemiology as a field is pushing pretty hard against p-values). You can already see this happening occasionally (for example, the LaTeX vs. anti-LaTeX soapboxing).
  3. It serves the technical sites poorly. At least one of those questions is more suited to CrossValidated. The StackExchange network is best served by directing people to the site with experts, rather than trying to take on those questions on more generalist sites, which at best ends up fragmenting "Where can I find answers to questions about X" and at worst results in lower-quality answers. Because many of us are probably decent Python programmers does not mean we should be answering programming questions when SO exists.
4

In principle I agree with you, and I upvoted your proposal.

In practice, I'm worried about how we could ensure uniformity of treatment between different disciplines, when people might not understand the potential generality of a certain relatively technical question related to a certain field, and where to draw the line of technical but not-too-technical.

Will we be spending time reminding people, in comments or here in meta, that certain questions are on topic?

Just to give an example from the recent past, several people would vote to close questions about writing style and formatting. These are somehow technical questions related to what is clearly a major activity of ours. How can we ensure uniformity of views with more borderline questions?

4

I have a concern about "procedural" issues, not about the idea you're proposing:

By this point, some votes have accrued on this question. But they are not so useful for moderation purposes, since you have not clarified what kind of question you're referring to:

  1. I doubt all of the people voting on this question have the same thing in mind, so it's hard to point to this post as evidence of "consensus" for something.
  2. I (and other users) use meta posts as a reference (in comments) when I see close votes on a question that I think should stay open. I couldn't use this post as a reference to explain why a particular question should stay open - it's too vague.

Moving forward with this suggestion, can you (or someone else) open a new meta post with a much more specific "type" of question you think we should be more tolerant of, with an example? You can look at closed questions for ideas - if we are actually getting and closing questions of the type you're referring to, there should be some examples in there.

(Ideally, answers to this new meta question should explain why the category of question is good for the site, if in favor, or bad, if opposed. That way the post is useful as a reference for purpose #2. Then votes on the answers can show consensus - purpose #1.)

  • I added a few examples. – Cape Code Jan 19 '17 at 8:21
  • @CapeCode thanks - that definitely clarifies what you meant! But it's still hard to say if the 8 people that already voted (7 up, 1 down) had the same thing in mind when they voted on the previous version. I guess we'll have to mostly look at the new votes that come in now. – ff524 Jan 19 '17 at 8:25
  • Agreed. Lesson learned for my next meta post. – Cape Code Jan 19 '17 at 8:33

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .