Bad behaviour from a moderator has a negative impact on the community

Moderator behavior is a very crucial point for a community. This means that such people have to be evaluated in a very strict manner. All the time. Otherwise they tend to derange and misuse their power.

Summary:

• Moderator gives partially wrong answer and deletes the correction post.
• In the same step copies the answer in the next step and earns credit.

Why are admins allowed to delete correction-posts from users even if they gave a wrong answer? When it comes to math there is no negotiation what could be wrong or right. It should be possible to point out mistakes without getting punished.

It's about this post where "whuber" gave a partially wrong answer based on the question of OP. I replied with my post and explained why he is wrong. What followed is not a discussion. My comment simply got deleted without an answer. His post remains wrong. This is not only discouraging since I've seen many incorrect answer and now took an evening to gain some points which should allow me to comment.

The next thing is a plain copy of my answer in this post from the same person, just minutes after his deletion.

This is a very shameful and discouraging behavior. Such a moderator should be degraded or at least tainted.

Never give a person too much power for an unlimited time.

• Please take a few moments to review our help system so you understand how this site works. I deleted your post because it did not answer the question (and therefore was likely to collect downvotes very rapidly, BTW). That has been the fate of ten thousand such posts before yours and will continue to be the fate of such posts in the future. If you find that improving this site by removing such material is "shameful," then you might prefer participating in a different kind of forum. – whuber Sep 12 '15 at 20:24
• Dear whuber. It's better not to give any answer than a wrong answer. Your post treats RMS as standard deviation which is wrong. There is no negotiation in this. Correct it or delete it. If you don't understand how math works it is better for you to participate in a different forum since since wrong answers lead to bad science. The shameful part is the copy of my answer in the second post which initiated this post. – nali Sep 12 '15 at 20:28
• Re the edit to this post: your representation of the events is woefully incorrect and not borne out by the record. You are making unsupported and scurrilous public accusations of misbehavior. Please think hard before you continue this course of action. BTW, I did not copy any answer of yours--nor, as anyone who has any familiarity with this site will know, have I any need to. Whether or not either of us understands math is not a question for public debate, so any further posts on your part about such issues will be deleted without further notice. – whuber Sep 12 '15 at 20:29
• I'm glad we have timestamps so even at this point there is no negotiation. – nali Sep 12 '15 at 20:30
• And what do we learn from it? If you are a moderator you can give wrong answers, delete the correction, don't correct your own post for the sake of science and copy other posts without being punished. Such a lack of self-criticism is very dangerous. – nali Sep 12 '15 at 20:37
• I would agree--if anything like what you say actually happened. Why don't we try this: over the next few days, the more experienced members of this community will read your post and may investigate your claims. If they find merit in them, they will respond positively and upvote your post. If they do not, they will likely downvote this thread and perhaps even offer comments (or answers) explaining their reasons. These are the people who have the ultimate power on this site: no single moderator does. In the meantime, please be so kind as to visit stats.stackexchange.com/help/be-nice. – whuber Sep 12 '15 at 20:41
• Dear Dr. Huber, you don't want to argue with me that for a sample $RMS=\sigma$ don't you? I don't think you're that silly. Both of us know you're wrong. So let's take this point aside and correct your post. There is absolutely no "I would agree--if" inside for you. Take your moderator role more seriously. For the second point: I might be wrong but I posted the same answer first, which is my point. After you deleted my answer without a correction, you posted the same answer in short form. There is also no negotiation on this point. Your lack of self criticism is bad for a scientific community. – nali Sep 12 '15 at 20:53
• Downvotes don't count when it comes to judgement between right or wrong. It's a logical fallacy to believe if the bigger mass believes in something it must be right. I'm writing it down here for the sake of documentation and to rise awareness of possible abuse of moderator rights. You are a reputable person (here) but this does not mean that you have the right to derange at some point. – nali Sep 12 '15 at 20:59
• First Point: Due to the lack of some points I wasn't able to comment and point out that your answer is wrong. Should I keep quiet if someone posts a wrong answer and, as you state it "the more experienced members of this community" up-vote it? This is ridiculous. It is okay to delete my comment as long as you correct your initial post. Otherwise it's better to delete your answer to avoid bad work being derived from it. But obviously you have no problem with wrong answers since it's still not corrected. Second point: Your answer was 175,-10 with $n$=100 which I posted around 3 hours before you. – nali Sep 12 '15 at 21:24
• Re "Should I keep quiet?" -- until you have spent the small amount of time and effort needed to gain the reputation to comment, actually that's exactly what the site is designed to try to make you do. [You'd also need to do a much better job of explaining why you think you're right than you managed there.] – Glen_b Sep 12 '15 at 23:21
• It's silly to keep quiet if someone posts a wrong answer which leads to wrong results and the "experts" upvote it ignoring the fact that it is a wrong statement. My answer was: "In general RMS≠σ. To compute the standard deviation σ from a sample, as asked in the OP, you need to divide by the degrees of freedom which is usually n−u : number of observations minus number of parameters." This is explanation enough. – nali Sep 13 '15 at 0:52

Nali, I do owe you an apology. The proper way to proceed when I encountered a flag on your first post was to (1) recognize that you are a relative newcomer and welcome you to our site and (2) add a comment explaining why your post had to be deleted. I could plead extenuating circumstances, but that would just be an excuse. I am sorry about the misunderstandings that arose.

As far as your second concern about "copying" goes, please recognize that any answer not only has to be correct to be of any value: it also has to communicate. When two people post the same numerical answer to a question but one explanation is difficult (or impossible) to understand and another gets right to the point with a clear explanation, then it is much more valuable. Thus we all strive to be not only right, but also to be understood. This is especially the case with statistical questions, because statistics is not mathematics. (And even in mathematics, clear communication is valued.) Many questions permit multiple interpretations and may have differing correct answers. For these reasons we welcome multiple answers to questions and sometimes those answers can look superficially similar.

• Sorry I've just read this post. There might be some hot arguments from my site which you can find prior to this answer. You don't need to apologize. I also overreacted. But honestly I never thought that a basic thing like penalty for the loss of degree of freedom is not taught at the university level. Wish you a nice start into the week. – nali Sep 13 '15 at 23:46
• "any answer not only has to be correct to be of any value: it also has to communicate" That is true of any SE/Q&A site in my opinion. – Erik Sep 16 '15 at 23:40

Added in Edit: Now that two of us are accused of abusing our position as moderators, I'm not sure I can contribute anything further.

You have said of each of us that you think we're abusing our positions; you have the right to email the people that oversee us (yes, even us completely unpaid volunteers). See this meta post which explains how to do that.

1. In respect of the first post you link to, the stackexchange network is deliberately designed (not by any of us talking here) to work a particular way. This includes that -

b. novice users don't have the ability to make comments; this is deliberate -- to make sure new users at least understand how the system works before they start commenting. The system worked correctly in this case because it seems you're not yet in a position to understand the appropriate site behavior (including that you don't comment in answers), and so it looks to me like you shouldn't in fact be free to comment until you've learned more about how it works. The site works this way for a reason.

c. moderators are not only given the ability to delete comments posted as answers, that's one of the things they're tasked with doing. That's what moderators are elected for. whuber not only acted as he should, he was chosen by the rest of us to do exactly that, and the site would be far poorer without his efforts.

d. If he had not deleted your non-answer, I would have as soon as I became aware of it. I expect any of the other moderators would do the same. It's our job. In a typical day, I will delete several posts like that.

So in the first case, you've accused him of wrongdoing when he did what he's supposed to do.

[The correct action for a user without the reputation to comment if you feel his answer is wrong (it isn't actually wrong but that's not the present issue), is to post a better answer. Many of the active users - the people who bother to vote - are very experienced and knowledgeable statisticians with the ability to tell a good contribution from a poorer one. If the people who vote on the site agree that your answer is actually better, you'll get some upvotes, and it only takes a small handful of upvotes to earn the reputation required to be able to comment. If your answer is less spectacular or does not improve on existing answers, you might require several more answers to get the needed reputation.]

Now you've drawn my attention to the question, I shall certainly be upvoting whuber's answer, since it's the best answer I see on the page; it's better than yours in several important respects, and I think Repmat chose wisely in awarding the tick. In his position, I'd have done the same thing.

I see nothing whatever that supports your second accusation against whuber. Unless you have something substantially better to base it on than the similarities we'd expect to see in answers to the same question, I don't see there's anything to respond to.

• There is a lot of text. It does not matter since. A wrong answer is a wrong answer. If it's not corrected, it harms more than it helps. You're the second who thinks that the mass is right. A huge logical fallacy. I guess that's the difference between a valuable journal and online sites where moderators and the crowd can decide what's wrong and what's right. Upvoting and experience don't prevent people from doing mistakes. Judge yourself is he right or wrong? Where is your experience? See what I mean? There should be always an option - even for newbs - to point out wrong answers in an easy way. – nali Sep 12 '15 at 22:39
• Where did I say "the mass is right"? I don't think that at all. Only a small proportion vote on the site, and it's easy to identify who many of the main ones are. Those journals that you think are valuable, take a look at the people who edit them and publish in them, and then look at who votes, edits and comments here, and you'll see a lot of overlap. (In some cases it may take you a while to figure out who is who but others make their names plain enough that you could figure out that in many cases it's a lot of the same people. Outside that we have a good number of highly experienced people) – Glen_b Sep 12 '15 at 22:49
• In that first post, I did already look at whuber's answer and decided it was a good contribution back when I first voted on it, a couple of years ago. If I could vote again, I'd upvote it now as well. The reason why new users need 50 reputation isn't because new users can't be correct, and that's not what I said. – Glen_b Sep 12 '15 at 22:51
• +1, but to be completely honest, one should add that moderators often choose to convert answers that should rather be comments into comments. In this case, @whuber decided to delete OP's answer instead of converting it into a comment under his own answer. As you said in (1.d) that you would have done the same, I wonder if you want to explain why. – amoeba Sep 12 '15 at 23:43
• @amoeba If you mean "why would I not choose to convert it to a comment?" -- It's not required, but I do it sometimes as a matter of courtesy, under consideration of two aspects that help: (i) the poster is a brand new user who has no idea how things work (which doesn't apply in this case); (ii) the comment is a good contribution which otherwise follows the rules and site conventions. I was pondering whether to do so in this case, but the exchange under the question here made me think I'd have to delete the entire conversation anyway. Given the nature of the discussion here, it would ...ctd. – Glen_b Sep 12 '15 at 23:51
• ctd... at best go into a chat room. Ultimately the reason I haven't converted it is that I think we'd simply get the same argument in comments there, and there's little benefit to that. The OP here is 2 rep points short of being able to comment, and I'm prepared to wait that out in the hopes that the comment, when it finally comes, is more measured, or at the least better argued. If conversion to a comment seems likely to lead to more heat than light, I'm not going to do it without very careful thought. Nothing I've seen so far makes me think the site would be better for it. – Glen_b Sep 12 '15 at 23:51
• @amoeba It's just occurred to me that there's an aspect of my answer that may be misleading you -- deletion of a comment-as-answer occurs either way; I'd have done that as a matter of course, since that's our job. The only aspect that may change is that some of those can end up as a comment. That's not something I would rush into in the present circumstances. However, if you feel it should happen now (rather than wait out the two reputation points), you don't need a mod -- simply find one of the OP's answers (5 to choose from) that you think is worth an upvote. – Glen_b Sep 13 '15 at 0:30
• @Glen_b: The answer was wrong. Point. A reputable journal would accept any correction-submission in any form. Point. It's still not corrected. This fact disqualifies both of you as moderators. – nali Sep 13 '15 at 0:44
• @Glen_b: "why would I not choose to convert it to a comment?" -- It's not required" Are you serious? It's a wrong answer and my correction was short and straight to the point. This is a clear miss-use of moderator privileges. The right behavior from whuber should be a short change, and honestly I wouldn't mind if my comment is deleted afterwards. Now since you are in the game you start to talk around instead of getting to the point and accepting that the answer is wrong and that whuber should have corrected it. I'm glad this is documented, it shows the quality of such portals and their mods. – nali Sep 13 '15 at 0:58
• Glen_b, my point was not so much that @nali's comment-answer should be converted into a comment now; my point was that if it had been converted into a comment right away (instead of simply deleting it) then this would likely not have alienated OP so much. I am also a bit surprised that this did not happen, given that such comment-answers are often being converted into comments. I guess it had to do with this comment-answer not being seen as helpful or correct; so perhaps this discussion should rather center on whether it is in fact correct or incorrect. – amoeba Sep 13 '15 at 9:21
• @nali, you have already repeated a couple of dozens of times that you believe that whuber's answer is wrong; as you say yourself, math is not about negotiation or arguing, so perhaps you can attempt to prove that it is wrong. You probably noticed that EVERYBODY here seems to disagree with you; don't you find it strange? If $x_i$ is a sample of size $n$ and $\bar x$ is the sample mean, then $\frac{1}{n}\sum (x_i-\bar x)^2$ is the maximum likelihood estimator of variance. You seem to think that this is false. If so, you are confused. – amoeba Sep 13 '15 at 13:24
• How are those aspects for you different from any other person in the thread? I have a PhD in stats, have published in a number of high-ranked journals and refereed for others, supervised a number of postgrad theses (including PhD), have taught statistics & stats-heavy subjects at several universities over more than 25 years (mostly to postgrads; right now I'm teaching a research subject to masters students to help prepare them to write their thesis), and I have worked in stats research in industry over an even longer period. That's nothing compared to whuber; take a look at his CV some time – Glen_b Sep 14 '15 at 0:41
• Almost all the usual active participants in meta are highly qualified, many much more than me; several are editors for high ranked journals. Which is to say, you can bet that almost everyone voting in this thread has very impressive qualifications (it will be the same people voting as usually participate). If the sort of qualifications you mention really count for you, you should be reconsidering your position. – Glen_b Sep 14 '15 at 0:47
• Personally, I don't put that much weight in qualifications; I think performance here is a better indicator of value; people who consistently generate very highly ranked answers over a long period of time demonstrate not only understanding, but ability to communicate that understanding. If you want to identify the people who know what they're doing, I'd say try something like count the "good answer" badges per hundred answers. If you want to have some idea of breadth of knowledge, look at tag badges per hundred answers. These are not the only indicators, but might be a starting point. – Glen_b Sep 14 '15 at 1:00
• @nali, You appear to be under a delusion that nobody here understands this $n$ vs. $n-1$ issue. This is, frankly, ridiculous. Everybody here knows this well. Maximum likelihood estimator for variance is biased, and the Bessel's correction is needed for an unbiased estimator. Yes. So what? You did not even start explaining why you insist on the unbiased estimator; what if one prefers the maximum likelihood one? Also, the formula given in whuber's answer was for standard deviation. Even with Bessel's correction it remains biased. Come on, you need to invoke better arguments. – amoeba Sep 14 '15 at 9:23