53

I posted an answer to this question: Writing ads to attract female PhD candidates, which boiled down to "How do I write job ads to attract female PhD candidates".

My answer was as follows:

Don't bother trying. Your ad is unlikely to make a difference either way; in order to do a post-graduate degree in Computer Science, one would presumably need an undergraduate degree in Computer Science. Since men who complete said undergraduate degrees significantly outnumber the women who complete said degrees (and the ratio gets worse the more developed your nation is, as the women are more free to pick the jobs that interest them, rather than the jobs that might be more economically viable), the number of potential female applicants is already too small to make a significant difference.

It got upvoted up to +11 votes, and was then deleted by Wrzlprmft and StrongBad. Since there doesn't seem to be any method of privately messaging them on here that I can see, I have to ask them publicly like this instead.

When I click the Help Centre link, I get the following list of reasons for answers to be deleted:

  • commentary on the question or other answers
  • asking another, different question
  • “thanks!” or “me too!” responses
  • exact duplicates of other answers
  • barely more than a link to an external site
  • not even a partial answer to the actual question

My answer fell into none of those categories. As I said in the comments of that question to another user, "Don't bother trying" is a valid answer to "How do I [do thing]"; it conveys useful information to the person who asked that question (namely, that attempting to [do thing] is a waste of time and effort, and that they would be better off not attempting to [do thing]).

As far as I can see, neither of the moderators who deleted my answer even made any comments explaining what problems they had with it; they just unilaterally deleted it without explanation.

  • 6
    FYI: Here is the pertaining review task. – Wrzlprmft Dec 22 '17 at 9:27
  • 2
    Also note that the answer was at +15/−4 and this. – Wrzlprmft Dec 22 '17 at 9:28
  • 7
    Okay, thanks, but that doesn't answer why it was deleted. What was wrong with my answer? It didn't seem to be breaking any of the rules. – nick012000 Dec 22 '17 at 9:38
  • Well if StrongBad was involved we know why. (I'm referencing en.wikipedia.org/wiki/Strong_Bad) in case someone gets offended. – LateralTerminal Jan 4 '18 at 15:14
18

"Don't bother trying" is a valid answer to "How do I [do thing]"

In general, I agree with you. An answer that questions the premise of the question is a valid answer. But that’s not what your answer does. The question does not aim at fundamentally changing the gender ratio in the field of computer science, but just at making ads more encouraging for women.

A valid (but likely wrong) answer that challenges the premises of that question would be:

Don’t bother trying. While there might be an impact of the wording on the gender-specific appeal of job ads, Doe et al. showed that it is strongly field-dependent and least prominent in the field of computer science where women are at most 5 % less encouraged by typical job ads than men.

Yes, your answer also starts with don’t bother trying, but the following elaboration makes clear that this refers to something else than the asker’s goal (with this question). Your answer addresses the question:

Should we bother to increase the gender ratio of computer scientists at the PhD level?

This is considerably different from the question in question. (It would also be too opinion-based for this platform.)

Hence, I considered your answer to fall in the category not even a partial answer to the actual question.

  • 15
    The question was how to increase the number of women that responded to their job advertisements; my answer was "Don't bother trying because the return on investment will be too small to make a significant difference". I'm not sure what the difference between my answer and your suggest answer is, in practice? It seems like hair-splitting to me. – nick012000 Dec 22 '17 at 9:58
  • 12
    The suggested answer says “don’t bother trying because the effect that you assume is negligible”; your answer says: “don’t bother trying because it’s not worth the effort”. The first answer may provide relevant information to the asker; the second is just challenging ethical decisions they obviously already made. It’s like suggesting a career in industry to somebody asking about the technicalities of graduate admissions. – Wrzlprmft Dec 22 '17 at 10:51
  • 30
    I think rather than deleting (which I think may come off as harsh/extreme), it would have been better to suggest edit in the comment, similar to the line of thoughts as what you wrote in this post. I agree with @nick012000 that the difference is small and the answer can definitely be salvagable. – Hosea Dec 22 '17 at 16:46
  • 7
    @Hosea: Edits are not supposed to change entire lines of argument. Moreover, I doubt that you can find any evidence or reasonable argument to back the suggested post up (not that the original post contained good arguments, IMHO) – it is really just an example to illustrate what a “Don’t bother trying” answer would look like. – Wrzlprmft Dec 22 '17 at 20:03
  • 1
    Well, there was this rather excellent documentary suggested in the comments, that I'd seen years ago and forgotten the name of. – nick012000 Dec 22 '17 at 23:39
  • 4
    @nick012000: I will not watch half an hour just to check your argument, but going by the title and first minutes, this goes along an entirely different argument. – Wrzlprmft Dec 23 '17 at 14:24
  • 1
    Some cited academic references exist. I'm not taking sides on the deletion issue, but kifinfo.no/en/2016/05/bigger-gender-gaps-rich-countries and scholar.google.com/… are places to start. – Fuhrmanator Dec 23 '17 at 20:39
  • @Fuhrmanator: Unless I misunderstand this, this at best provides evidence for a claim that is similar to one sentence in the original answer in question. I see no apparent connection to the answer I “suggested”. Anyway, the question is not whether the original answer or the “suggested” answer are correct or backed up but whether they address the question or should be translated into each other whether editing. – Wrzlprmft Dec 23 '17 at 21:42
  • 2
    The interpretation of the question as "how to balance the gender ratio in CS" is a reasonable guess based on the OP. It is considered acceptable on the SE network to look for the question behind the question -- see the "Giving Answers to XY problems" (no pun intended) part of meta.stackexchange.com/a/269222 . – darij grinberg Dec 28 '17 at 14:23
  • 2
    It is considered acceptable on the SE network to look for the question behind the question – If you have to and to some extent, yes. However, In this case, you do not have to as there is no XY problem: The asker clearly describes what they want to do and why, and it is by no means bizarre. Also the answer digs too deep: The question it presumes is too opinion-based for this site and several layers below the actual questions. As I already said before: It’s like suggesting a career in industry to somebody asking about the technicalities of graduate admissions. – Wrzlprmft Dec 28 '17 at 19:57
  • 9
    @Wrzlprmft: "The suggested answer says 'don’t bother trying because the effect that you assume is negligible'; your answer says: 'don’t bother trying because it’s not worth the effort'." - I am having a hard time seeing the effective difference between these two answers. Both can be interpreted to challenge whether attracting female applicants is a worthwhile goal as such, and both can be seen as explaining why the suggested targetting of ads is not a way the answerer would recommend to choose. – O. R. Mapper Jan 2 '18 at 9:30
  • @O.R.Mapper: Both can be interpreted to challenge whether attracting female applicants is a worthwhile goal as such – I don’t see how the “suggested” answer does this. It does not challenge the goal at all, just the method. — both can be seen as explaining why the suggested targetting of ads is not a way the answerer would recommend to choose. – The deleted answer does not talk about methods or job ads at all, but just challenges the goal of trying to change anything at the PhD level. – Wrzlprmft Jan 2 '18 at 9:54
  • 6
    @Wrzlprmft: The suggested answer says the effect aimed for happens only to a negligible extent, thereby possibly challenging the idea that even what you consider only a negligible effect could be worthwhile to the asker in some way - just like the deleted answer can be interpreted to ignore the asker's idea of a worthwhile cost-performance ratio. On the other hand, both the deleted and the suggested answer conclude that in order to change things at the PhD level, adjusting the application process at the PhD level is not a promising method. Thus, they are basically the same. – O. R. Mapper Jan 2 '18 at 13:51
  • @O.R.Mapper: The suggested answer says […] – There is a considerable difference: The “suggested” answer would provide new information that is actually about what the asker wants to know. Modulo rephrasing, it would be “there isn’t” to the asker’s “Is there a list […]?” — both […] conclude that […] adjusting the application process at the PhD level is not a promising method. – Again, there is a difference: The “suggested” answer is about the specific adjustment that the asker wishes to make; the deleted answer is only about the level. – Wrzlprmft Jan 2 '18 at 18:03
  • 6
    @Wrzlprmft: Both answers respond to "Is there a list (...)?". One says "None with significant enough effects.", the other one "None with a sufficiently good cost-performance ratio." One answer points out just how small the effect of the proposed method is, the other one shows an inherent limitation of the method. Both are equally valid ways of showing that the proposed method is not a promising way to achieve the goal. I still don't see why one answer would be better than the other, but it seems we won't reach an agreement on this one :) – O. R. Mapper Jan 3 '18 at 6:45
10

The answer (1) expresses an unpopular point of view, (2) is poorly reasoned, (3) makes factual assertions without citing evidence, and (4) is not closely tied to what was being asked. The combination of factors 1-4 makes it seem trollish to me. If your intention was not to be a troll, and you actually want to convince liberals in academia to seriously consider a point of view that they are predisposed to reject, then you need to stop making mistakes like 2-4, which make it easy for them to dismiss you as a troll. SE is not an open-ended discussion forum, it's a forum in which people ask and answer specific questions.

  • 16
    "SE is not an open-ended discussion forum, it's a forum in which people ask and answer specific questions." - And any answer you disagree with on political grounds is obviously a discussion forum post. – Retired account Dec 24 '17 at 4:35
  • 8
    @A.W.Grossbard: And any answer you disagree with on political grounds is obviously a discussion forum post. – Any questions the answers to which are primarily based on political opinion should be closed as primarily opinion-based; this also applies to the question the answer in question tries to answer. However, this question is not the question to which this answer was posted. – Wrzlprmft Dec 24 '17 at 11:28
  • 6
    @Wrzlprmft "Any questions the answers to which are primarily based on political opinion should be closed as primarily opinion-based; this also applies to the question the answer in question tries to answer. However, this question is not the question to which this answer was posted" - While I can't help but envy your linguistic dexterity, I must disagree. It is not necessary for an answer to be primarily political for one to disagree with it primarily for political reasons. I think the answer in question is a prime example. – Retired account Dec 25 '17 at 2:42
  • 26
    "seem trollish" is a stupid reason to delete a post, sorry. – darij grinberg Dec 28 '17 at 14:17
  • 10
    I believe that these factors should be expressed in terms of downvotes, not deletion of answer (according to the list of reasons listed in Help page for deletion). Factor (4) may be worthy of deletion, but I think the answer still addresses the question partially. – Hosea Dec 28 '17 at 21:11
  • 7
    (1) shouldn't be a reason for deletion. – Simon Forsberg Jan 3 '18 at 18:57
8

The way the SE system works is we like to have our discussions about moderation in public so everyone can participate since we are really a community moderated site and so there is a level of transparency.

In general, regular users cannot delete an answer with a net positive score. They are forced to flag the answer. It then gets reviewed by other high rep users, and based on that review gets brought to a diamond moderator's attention. As a diamond moderator I have extra tools and privileges that allow me to delete up voted answers. The review of you answer is here: https://academia.stackexchange.com/review/low-quality-posts/57451

Basically, a high rep user raised a flag on your answer (it is not public who raised the flag, but they are a respected, trusted and valued member of the site). Then 7 additional high rep users reviewed your answer and 5 of them recommended it be deleted. Another user flagged your answer as rude. Basically 7 users voiced support of deleting the answer and 2 suggested it was ok.

Based on this review, I looked at your answer. I agreed with the first comment to the answer

Sans the first sentence, this is a poignant comment

And could see how the 7 users might be agreeing with the idea that your answer is not even a partial answer to the actual question.

Overall, the decision to delete the answer was a real struggle for me. My personal preference would be to keep the answer. In other words, if I had a non-mod vote the review would have been 7-3 instead of 7-2, even factoring in your presumably positive vote, only gets us to 7-4. I deleted your answer, because the evidence at the time was that is what the community wanted. If the community thinks the answer is worth saving (and I might even write an answer arguing it should be), I will happily undelete it.

I suggest you either provide an answer to this meta question arguing why the answer should be undeleted because in fact it is more than just a comment and provides at least a partial answer to the question. If the community agrees (by up voting, since votes on meta are interpreted differently), we will undelete the answer.

One difficult aspect of judging the community wishes is the net vote count does not always reflect the wishes of the community. Questions that make the HNQ list attract a lot of attention from SE users whoa may not be active members of academia.SE. These users lack the reputation to down vote questions and answers, but can upvote. This bias is a known problem: Prevent questions on Hot List from being upvoted by casual visitors (only rep is from association bonus). There is also the issue of down votes costing rep, so there tends to be an upwards bias. In general, 4 down votes is a lot for the main site. It is also a lot on meta, so this answer (and my actions) has clearly struck a nerve.

  • 31
    I'm not really comfortable with the idea that any content is deleted just because the community wants it to be. Such an approach leaves too much room for drowning out less popular points of view. This seems antithetical to the goals of SE. If your deletion was based not on community opinion but on less arbitrary criteria, I recommend deemphasizing the notes about community opinion being a deciding factor and putting greater emphasis on the actual criteria. (You do mention that you felt this answer was more a of comment, but you don't unpack that thought at all.) – jpmc26 Dec 28 '17 at 1:46
  • 1
    @jpmc26 I am not sure your views are consistent with a community moderated site, but you should probably ask this as a new question. – StrongBad Dec 28 '17 at 2:07
  • 10
    I think they clearly are. Even though the community moderates the site, we're still given instructions and expected to follow them. (This is why we have review bans.) Our votes are supposed to be based on applicability of the rules rather than our desires. Yes, we trust users to exercise their moderation abilities according to the rules, but a diamond moderator has an even greater responsibility to be as objective as possible. Note that I'm not saying your actions were wrong (or right), merely that your answer here does not emphasize the rules to which all users are supposed to adhere. – jpmc26 Dec 28 '17 at 2:53
  • 4
    As a procedural point, it's worth pointing out that the linked review would have straightforwardly deleted the answer had it been negatively scored. Similarly, that answer's score needs to be considered carefully: the question was on the Hot Network Questions sidebar for a substantial period, and those are well-established to attract upvotes (but not downvotes!) from a much wider cross-section of the SE network userbase (including many users with a much smaller investment in this site) than regular posts. – E.P. Dec 29 '17 at 23:34
  • 6
    "I deleted your answer, because the evidence at the time was that is what the community wanted". The community? 7 vs. 2: p-value = 0.84 ;) – Orion Dec 30 '17 at 15:48
  • @Orion along with a strong prior based on the content and who was voting. Based on the answers and votes on this question, makes it seem like I understood what the community wanted. Also remember that we close and delete questions with only 5 votes and ignore up votes completely. – StrongBad Dec 30 '17 at 16:07
  • 7
    One thing I would like to point out: flags are anonymous to the general public for a reason. Even stating that the flagging person was a respected, trusted and valued member (implying that they had way more than the 15 (!) rep needed to flag) is too much information in my humble opinion. Please ask yourself: If a 15 rep user had flagged the post and the review result was the same would your choice have been a different one? I hope the answer is ‘no’. – Jan Jan 1 '18 at 15:21
  • @Jan the reputation, and past actions, of users can play a role in decisions. As for the review of this answer and the actions, the review was clear enough that it didn't really matter to me that it was high rep users voicing their opinions. As for leaking too much information, I disagree, but you should ask that as a separate question. – StrongBad Jan 1 '18 at 17:24
  • I was specifically asking about the person who raised the flag. The reviewers are, per definition, relatively high in rep as they can only review from 2000 rep upwards. I also don’t really consider the flagger’s reputation and trustedness ‘leaking’, only irrelevant. – Jan Jan 1 '18 at 17:30
  • 9
    I find it strange you'd delete any answer with a net positive score of >10 based on the "wishes of the community." The wishes of the community are most represented in the answer score, not in the review queue with its extremely small sample size. – user168715 Jan 2 '18 at 18:16
  • @user168715 see edit, voting tends to be more biased and less reflective of the community than reviews. This is especially true of questions that make the HNQ list. – StrongBad Jan 2 '18 at 18:58
  • 6
    The 7-vs.-2 figure doesn't make sense; flags can't be counted since users can't flag things as okay. The review process itself is biased (I mean, when we see a post flagged for review, we're going to start from a place of skepticism about it), so even 5-vs.-2 would be overstating the signal. Not to take a position on this Meta thread, just the signal analysis doesn't follow. – Nat Jan 2 '18 at 23:39
  • Or, I guess users can kinda flag things as okay by up-voting them, and this post did get a lot more up-votes than down-votes. But that's also a biased signal. So, again, not to take a position on this Meta thread, just trying to correctly qualify the signal. It's actually kinda neat to think about the modeling phenomena that go into this situation. There're interesting phenomena like social inertia in both the voting and flagging processes, etc.. – Nat Jan 2 '18 at 23:48
6

If I understand correctly, the question had +11 votes, as such it seems that the "community" valued this answer. In addition, even if that was not the case, as @jpmc26 said, "I'm not really comfortable with the idea that any content is deleted just because the community wants it to be. Such an approach leaves too much room for drowning out less popular points of view".

Regardless of the last statement, 15 people liked the answer, 4 did not, it means (to me) that those 15 people thought the answer did provide some value. So, while I appreciate the point of view of the moderators that the answer did not answer the question, it seems that at least 15 people did find some value in the answer.

I am just a lowly user, but, for what it is worth, these are my 2 cents, and I would vote for the answer to be re-instated (if I had the opportunity).

To summarise: why do I believe the answer should be undeleted?

  1. It was up-voted by 15 people.
  2. The answer does answer, at least partially, the question asked. In other words, people reading that answer (who may have the same question) may still get some feed-back (they might agree with @nick012000 that it is useless to bother trying).
  3. The reason I go to SE is to read different answers, and I would not want that particular answer deleted.
  4. The answer is not completely off-topic, nor is abusive or insulting
  • 7
    For what it's worth, votes on posts that end up on the "Hot Network Questions" list and get a lot of views through that, tend to be very skewed. These get a disproportionately large number of views from users who are not regular contributors and have only the association bonus (100 points). Those who want to upvote can, but those who want to downvote can't because they don't have enough rep. (See this meta post for more details.) This post was on HNQ list and got lots of views, so I wouldn't read too much into the vote ratio. – ff524 Jan 2 '18 at 3:52
  • 2
    @ff524 That is something I had not considered. – user Jan 2 '18 at 11:13
  • This is your opportunity to vote to get the answer undeleted. You need to provide a clear reason that convinces enough users that the answer should be undeleted, in which case I (or another mod) will undelete the answer. That said, the other answers, and the support for them, suggests that the community is okay with the answer being deleted. – StrongBad Jan 2 '18 at 14:42
  • 1
    I am not directly involved in this, just an interested user in how SE works. I offered my reason on why I believe the answer should be undeleted: it was up-voted by 15 people. The answer does answer, at least partially, the question asked. In other words, people reading that answer (who may have the same question) may still get some feed-back (they might agree with @nick012000 that it is useless to bother trying). This is why I go to SE, to learn about different answers, and I would not want that particular answer deleted. – user Jan 3 '18 at 12:12
-9

Don't you know that moderators are executive, legislative, and judiciary all in one? Don't you know that the SE moderation is also winner-take-all? Don't you know that moderators determine what "rule of law" is?

Since you can't be pro-law today anymore without people assuming that you're being pro-opposite-party when the "law" doesn't match with popular opinion, I do lean liberal.

Math (my community) has the same problem. I would be thrilled if math gained more venerable women mathematicians. Math needs as many brilliant minds as possible. However, attempting recruitment at a PhD level seems to be misdirected. There's research that suggests women tend to lose interest in the sciences between middle and high school. This needs to be researched more.

The way I interpret nicks answer is that you can't fish a dry pond; rather the people upstream controlling the dam need to think more downriver. And this is a perfectly acceptable answer to the question.

You must log in to answer this question.

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