As per the moderator my question isn't appropriate: Variance formula amplify outlier instead ignoring it!

I am learning stats and I tried many resources over internet and books but I could not get cleared my doubt behind the variance formula. In all of my wisdom, I believe I have asked a useful question for the community but I am not able to get off from the hold put by the moderator on my question

I need help from someone who could verify this first question of mine in this community and let me know if my question really doesn't have any merit or it is just a person's opinion.

Please help and kindly justify the moderator action or if I am wrong. I am bringing this up as if the moderator behavior is inappropriate then this is adverse to the essence of great community that all stack exchange community are known for.

  • 5
    $\begingroup$ It's not an issue of merit--it's purely an issue of expressing yourself so you can be reliably understood by most site visitors. Please clarify your question if you would like it reopened. As you likely know, because you seem familiar with SE, we offer guidance on what is on topic and how to ask a good question in our help center. $\endgroup$ – whuber Feb 28 '19 at 0:04

The moderator gave you the opportunity to clarify your question, so that it would be more likely to be understood by the community. This serves two purposes. First, it helps you personally, because it makes it more likely that you'll receive a useful answer. Second, it serves the community at large, because questions and answers should be useful to a wider audience who might run into the same or similar issues. This requires that the question is stated in an understandable way.

Note that your question is not definitively shut down. It can be reopened if you update your question to be more clear. If it is unclear to you where there is room for improvement, there are many people here who would be happy to help you with this.

While a question being put on hold may be frustrating, it is not intended as a personal attack on you. Please assume good faith, and bear in mind that our moderators volunteer their time and effort in order to keep our community in good shape. Lashing out at them is not a productive response, even if you believe you have been wronged.

  • 1
    $\begingroup$ If i really know, where is the room for improvement in the given question I would have done it. Quite sarcastic to me- people are just downvoting and also happy to help! I am done with this, and I have no motivation to take help or improve this question for this forum. Thanks but you comments sounds of NO USE to me again, $\endgroup$ – VISHAL DAGA Feb 28 '19 at 17:19
  • 7
    $\begingroup$ @VISHALDAGA At least 4 people, including Ruben above, have offered you feedback on your questions. They are being generous with their time in an attempt to help you. I suggest you calm down and take what they are saying to heart if you want to have a productive time on this site. $\endgroup$ – mkt - Reinstate Monica Feb 28 '19 at 22:06
  • $\begingroup$ I think in its current form the Q is perfectly clear. Do you agree? If so, please vote to reopen (as I did). I also upvoted it because I think it's unfair that it has -2 score now. $\endgroup$ – amoeba Mar 1 '19 at 14:07
  • $\begingroup$ +1 For "it is not a personal attack on you" $\endgroup$ – Alexis Apr 16 '19 at 16:44

The topic of whether variance (which is square by definition) is a good measure of dispersion or deviation is not without merit and is being discussed a lot.

Your question, however, is very unclear. That is why it is closed. See the text that explains the reason for closing

"Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking."

  • $\begingroup$ I think in its current form the Q is perfectly clear. Do you agree? If so, please vote to reopen (as I did). I also upvoted it because I think it's unfair that it has -2 score now. $\endgroup$ – amoeba Mar 1 '19 at 14:07
  • 1
    $\begingroup$ @amoeba I still find it difficult to find the essence of the question. Is it about the use of squared difference to benefit outliers (which is a false statement, in my opinion, or the term 'outlier' should not be taken explicitly, as I already sort of "answered" in a comment under the question)? Is it about the particular example given in the question (which lacks any information about background, circumstances and detail)? Is it about the benefit of using squared error in other circumstances? $\endgroup$ – Sextus Empiricus Mar 1 '19 at 14:15
  • $\begingroup$ @amoeba I'm on the fence about this. I think the question in its current form needs considerable experience and imagination to make a good guess about what it's trying to ask. My guess is that it might be inquiring how one can use outlier-detection tests that are (somehow) based on a sample variance. This is a question that AFAIK has not appeared on our site, and so I would favor opening it, but in its current state it is so vague that the possibility of misunderstanding and confusion seems large. I'm not voting to reopen it. $\endgroup$ – whuber Mar 1 '19 at 14:16
  • 3
    $\begingroup$ In addition I find the behaviour of the OP very uninviting to open the question. The toxicity should be removed from the question as well before I touch my hands on it. $\endgroup$ – Sextus Empiricus Mar 1 '19 at 14:16
  • $\begingroup$ I think I already sort of "answered" the question (at least some interpretation and part of it) in two comments under the original question. And here I have hinted that a different use of terminology may help (variance is by definition the square) and that a question about dispersion or deviation would not be without merit. Yet, the OP has not shown any interrest to follow up on this. We could start answering your version of the question or my version of the question, but I rather see that the OP clarifies his question better. $\endgroup$ – Sextus Empiricus Mar 1 '19 at 14:20
  • $\begingroup$ @whuber I have a feeling you are overthinking it! The OP seems to be asking a very basic question: why is "variance" (mean squared deviation) a useful estimator when it's so obviously non-robust, and why wouldn't one rather use mean absolute deviation instead, given that's it's more robust (in OP's simple toy example). $\endgroup$ – amoeba Mar 1 '19 at 23:45
  • 1
    $\begingroup$ @amoeba I appreciate your position. Mine is that we ought to take a poster's language seriously, even when it makes little or no sense, and that in this case the references to "outliers" strongly suggest the underlying concern might not be about variance per se. This difference of interpretation, whose roots are plainly evident in the original text, is precisely the problem: this question can look entirely different to different people. That sort of post often leads to good statistical questions once it has been clarified, but as it stands it is not appropriate for an SE site. $\endgroup$ – whuber Mar 2 '19 at 13:35

You must log in to answer this question.

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