For many questions on UX.SE I see the answer given "The only way you will know the answer to this is to test..".
Sure, I do believe this is the case for some questions, the question may be too unique or the research simply hasn't been done and the only way the poster will know is to test with his or her users. However, for the sake of this site, I don't believe coming in to a question and reading the accepted answer of "you will have to test" is going to help anybody.
It is also a bit of a blanket answer that could be applied to any question and technically be right, testing which is better will tell you which is better. So when some people see those as accepted answers they may feel more inclined to repeat it for questions with better answers available.
Presuming that the answer itself provides enough reasoning and context to otherwise not be closed/downvoted as not an answer or poor quality, would providing no concrete answer other than to test it with your users count as a valid answer?