Because it was founded by computer programmers, SE seems to have a unique approach to what constitutes good and bad questions and answers. Specifically, my understanding is that a good question is one that leads to a (short) series of answers that "converge" on one best answer after one or two iterations.

One strategy for constructing such a question is to have it lead to "binary" outcomes. The classic examples of this might be a true-false question. An example of such a question (from an earlier debate on Politics) went something like: "True or False, the clash of egos sometimes trumps more rational considerations in policy formation." Such a question can be answered "true" by one or two examples.

I tried to edit this question. https://politics.stackexchange.com/questions/1522/what-is-the-current-situation-in-turkey#comment7563_1522.

It was initially broad, but I tried to reduce it to a "binary" question, the gist of which was, "Is it true that the "official" reason for the Turkish riots (closing down a popular park) was the REAL reason?" In doing so, I left the OP's first sentence in place, and added a new, "second" sentence (deleting six or seven words in the process).

Is creating binary outcomes a good strategy for creating a sufficiently narrow question in general on the site? Did I succeed in this particular instance in improving the question so that five people would be willing to vote to re-open it?

  • A lot of topics simply aren't binary in nature. A lot of the SE sites seem to struggle with that. – user1530 Jun 10 '13 at 17:30

The key criteria for a question is not its binary nature, but rather its answerability. Ultimately, whenever you write a question, you should be able to articulate what an answer would contain, even if you don't know the right answer.

For example, I recently wrote this question: How high must a craft fly in order to not be in foreign airspace? I didn't know what the answer was, but I did know the terms in which it would be phrased - there is a set limit over which spacecraft are no longer subject to international airspace. In my mind, "an" answer would have been "International Airspace no longer applies when you are X miles above the ground."

That isn't binary - but it is answerable. In general, if you know the terms in which an answer would be given, you have an answerable question. If you are asking for an open-ended narrative, you don't.

Notice, for example the difference between:

  1. "What is the current situation in Turkey?"
  2. "Is it true that the "official" reason for the Turkish riots (closing down a popular park) was the REAL reason?"
  3. "What is the cause of the riots in Turkey?"
  4. "What is the proximate cause of the riots in Turkey?"

Question 1 is way too open ended. I could very legitimately answer question #1 with "It is currently 70 degrees F, with a 10% chance of rain" and be completely accurate as of 9am EST 6/6/13.

Question 2 is in fact binary, but the answer is necessarily opinionated. The real reason for the riots is X. The media are reporting Y. Therefore, the answer is {Yes | No}. Ulitmately, I could answer "Yes" or "No" but it wouldn't be a good answer, and frankly would begin to get into conspiracy theories and that sort of stuff. As such, it's better than #1, but still not great.

Question 3 is in fact, what I think you really want to know. "The riots are being caused by X. {MSNBC | Fox News | The Onion} is saying it, therefore, it has to be true. :) Okay, I'm being a little bit facetiuous (and admittedly, I'd prefer 'The Economist' says or 'BBC is reporting' but ultimately, since you are sourcing the claim, the answer is "good", but nonetheless, I digress. The point is, you can find a source that will claim it. If "I" am the source, the answer sucks, plain and simple. If "I" am quoting general knowledge, it's better, but still not as good as "This source says." And herein lies the rub - Questions 1 & 2 are not likely to be sourceable. Question 3 is, and is therefore better.

Question 4 is probably the question in my mind, because you move from a secondary source to a primary source to get an answer. "According to the protestors, the riots are because X." "Why" people are doing anything can get fuzzy. There are all sorts of historical reasons - good reasons to be sure, but a lot of them. What is the proximate cause / What was the spark that started this riot is much more defined. Again, because the question tells you what a good answer has, its better.

So, in the end, what makes for a good question? The answer is simple - a Question has set the criteria for a good answer.

  1. The answer itself can be sourced.
  2. The answer will be limited to a specific event or reason
  3. The answer will not be an "I think" but rather "This is expert analysis" (I don't care what "You think"!)
  4. A good answer is sourceable, verifiable fact, not opinion. Note - I never even said facts have to be right. They just need to be stuff that can be tested. A fact can be wrong, an opinion can only be heard.

You must log in to answer this question.

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