14

So this is mostly about this particular question. It has been very popular for a few days and both the question and several of the answers have gotten more upvotes than most other question on politics SE (disclaimer: I wrote one of the answers). So it seems the community likes the question and the answers provided. Now one moderator flagged most of the top answers and threatened to delete them. Per site rules his critique is completely justified. The answers are based on some general feel for public opinion and do not cite proper reputable sources.

Now I feel the rules are for the community and not vice versa so if a question that breaks the rules is super popular maybe it is time to question the rules instead of enforcing the rules against the community interest.

Providing sources where available is good and should always be supported. But there are questions where such sources are presumably not available. The given answers might look like personal opinions of their authors but they actually try to summarize general public opinion and they seem to do a good job of it. In this instance I would expect the best one could hope for is some social science researcher having thought about the same question and giving his educated opinion on the matter.

So my suggestion is that a question that asked 'what is your opinion on x?' is bad and should be deleted but a question that asked 'what is the general public opinion on x?' can be a good question and good answers will often just summarize what knowledgeable people perceive as the public opinion. For example the general public opinion in Germany towards nuclear power can probably be reasonably well summarized by anyone who has read a wide range of German press on the topic but that doesn't give any good source to cite.

2
  • 9
    It might be worth to take into account that this particular question was hit very hard by the hot network question effect. It got viewed by over 18,000 users. Most of them will be users from around the SE network who are not very familiar with the expectations on Politics SE and most of them have the right to upvote (due to association bonus) but not to downvote.
    – Philipp Mod
    Feb 28, 2020 at 10:22
  • 4
    Please note that as of this posting a number of highly upvoted answers (including my own) have subsequently been deleted, on white strike me as specious and pedantic grounds. I'll add my perspective as an answer below. Feb 29, 2020 at 13:31

2 Answers 2

9

Although this goes against my grain, I'm going to start by pointing out that I have extensive experience in philosophy, political theory, political science, and the philosophy of science. That much should be evident to anyone who's gone over my posts on the few sites I participate on here. I don't mention that — ever, for philosophical reasons — when I answer regular questions, and I'm only doing it grudgingly here because I feel I need to offset the social authority of a (to my mind mistaken) administrator. Because that is the problem here: that Philipp explicitly refused to engage in meaningful discussion, and plowed ahead with an entirely bureaucratic decision that did not take into account the sense or context of the question or the answers given.

The net result of Philipp's actions is that he actively created disinformation: suppressing a range of credible and reasonable answers and allowing only those answers that paint the situation in a particular light, because that particular light is the only way of viewing the question which can be accessed through the restricted methodology of opinion polling. I have no interest in speculating on his motives because I'm happy to believe he was simply following 'the rules' as he saw them. But I am pointing out the error.

I will go through this point by point, so it will be a long answer. My apologies in advance, but I want to be thorough.

1. The Original Question

The original question asked — the one which I answered — was titled "Why do Germans dislike Trump so much and consider him so dangerous?" There was a reference to a Pew poll in the first line, and the last line read "Are there some polls why Germans dislike Trump..." ('why' emphasized in text). The question title was modified to its current form two days after I answered.

My answer was based on the consideration that opinion polls are extremely poor tools for answering 'why' questions. Perhaps I should have put this explicitly in text, and likely I would have if the current title had been in place when I answered. But the fact is that opinion polls only gather Likert scale responses to pre-structured questions. 'Why' questions seek out motivations, and motivations are complex psychological phenomena that can only be accessed through some type of structured open-ended interview paradigm, or by inference through the application or reasoning to secondary, tangential information. My answer was perfectly appropriate for the original 'why' aspect of this question, since no opinion polls could ever suffice.

2. The Limitations of Opinion Polls, and the Danger of Disinformation

Opinion poll questions are inevitably practical and specific, such as:

  • "On a scale of 1 to 5, do you approve of Trump's tariff policies?"

This is a limitation of opinion polling, mandated by the desire to gain a comparatively large sample size with a comparatively high response rate. The more detailed and specific a question is, the more difficult it is to process on a large scale, and the more likely it becomes that respondents will get frustrated or bored and toss the form in the garbage. Moreover, trying to craft more deeply psychological questions on opinion polls creates other problems having to do with triggering. For instance (using my own answer as an example), consider an opinion poll question like the following:

  • "On a scale of 1 to 5, how much does Trump's rhetoric recall for you the political rhetoric of the Nazi regime?"

A question like that would both predispose respondents to particular answers and trigger emotional reactions; it is so methodologically questionable and so politically charged that any results from it would be effectively meaningless. I have occasionally seen poll-type questions that manage to thread this type of needle, but it's rare, and difficult, and it's risky for the researcher's reputation; and even then it is subject to a broad range of interpretations that limits its usefulness.

So, opinion polling is limited to simple, practical, intellectual questions, and thus limiting answers to those supported by opinion polls gives the false impression that all motivations are simple, practical, and intellectual in nature. Does anyone honestly believe that the the bulk of Europeans (or Americans, for that matter) evaluate Trump (or any political leader) on a calm, measured, reasoned analysis (CMRA) of their policies? No doubt that some people do, and that CMRA is a factor for many people, but implicitly asserting that CMRA is the only meaningful metric deeply misrepresents the nature of the situation, and inevitably creates a skewed, biased understanding.

3. Site Policy and Bureaucracy

Now, I am a relatively new member here — I've only been posting for about a month — so apologies if I am not entirely 'up' on standards and practices. But it seems to me that the purpose of a site like this is to inform readers, not misinform them. And yet, that is precisely what these deletions do. Now I could perhaps see the point of this (not agree with it, mind you, but understand it intellectually) if this kind of practice were a uniform feature of the site. But it isn't. I could look through any number of other questions on this site and find answers that are less up-voted, less intellectually sound, and less reasonable than my own, which have not been deleted or even questioned. For some reason this question caught Philipp's eye, and he decided to act in a manner that is inconsistent with general practice on the site; and yeah, I know, shit happens.

But as a general rule, when shit like this happens I expect a certain amount of discussion and reasoning to be applied. I do not expect to have an administrator flatly tell me "There is nothing to discuss", with the implication that if I do not comply with his wishes he will unilaterally delete my answer. I suggested we bring this to Meta or private chat to discuss the details (as I'm doing now); he refused. I gave him a précis of my objections to his actions; he declined to respond. The most he would say is: "Those are the rules and I am forced to follow them," a bureaucratic deflection of his own decision onto the site itself.

Is that the way this site is supposed to work? Because if so, that is — speaking frankly — stupid. And note please that I do not mean 'stupid' as a personal insult to any particular individual; I mean to imply that the entire site becomes stupid because it fails to fulfill any of the meaningful purposes such a site could have. It becomes simply an arbitrary collection of those answers that individual administrators have presumptively decided they will tolerate, for reasons of their own which they steadfastly refuse to express. If you can think of a better word for that than 'stupid', by all means do, but the point still holds.

I know that administrators have issues and problems I do not understand, and I am sympathetic to that. But my sympathy does not extend to endorsing haphazard, bureaucratic solutions. An administrator has an obligation to engage content, at least with members who are trying to be reasonable and earnest, and where administrators refuse to do so the effect is a consistent worsening of the quality of the site, and a chill on those trying to participate well.

I'm not entirely concerned about this particular answer (though Philipp's actions clearly annoy me). If this answer remains deleted I don't think I'll lose any reputation points, and if I do, I won't lose any sleep; I have the skills and knowledge to gain reputation points quickly and easily in this field. But I dislike the misrepresentations and malformations that inevitably arise from this kind of pedantic, unmethodical, bureaucratic thinking. If this is the way things are run here, then we seriously need a conversation about revising site rules and policies.

17
  • For some more background on related mod decisions politics.meta.stackexchange.com/questions/3848/… and politics.meta.stackexchange.com/a/4298/18373 I should probable make some mea culpa statements here regarding the original title. OTOH, I did not expect the question to get HNQ'd because I thought it was hard to answer (as actually asked in its body.) In retrospect I should have edited the title sooner... Mar 1, 2020 at 17:28
  • The cold war between top-down and bottom-up philosophies about what is on topic has been heating up lately. IMO, the top-down philosophy has always driven away users. (I entirely stopped using GameDev SE when its mods started rage closing interesting and valuable questions and interpreting what is on topic there as a strict subset of what is on topic on SO.) But this is doubly true now that the company wants to flood the site with new users and content irrespective of the value they bring. Mar 2, 2020 at 15:28
  • 2
    @Fizz: I hope you realize the way in which framing the question in this way creates disinformation. I mean, I look at the links you've provided, and they only confirm my suspicion that Philipp's perspective is jaded realpolitik; He is not dealing with the problem as a question of pedagogy or conceptual validity; he's simply trying to minimize the hassle and aggravation of dealing of dealing with political turmoil, and he's doing so at the expense of actual informativeness. Mar 2, 2020 at 16:06
  • 2
    @Fizz: had you edited the title earlier, I would have given the same answer — because my answer is intellectually and theoretically accurate — I simply would have prefaced it with an explanation of why what you were asking was a pragmatic impossibility. You think (perhaps understandably) that you are being more intellectually exacting by demanding 'evidence', but you don't realize that demanding the wrong evidence actually undercuts your question and forces responses to be biased misinformation. Mar 2, 2020 at 16:09
  • 2
    @user560822: I'm a stubborn, vocal idealist. No doubt I will get myself in trouble sooner or later (I've already got Philipp angry enough at me for calling him on his cr@p that he refuses to respond to me). I simply don't care about corporate or internal politics. Maybe I should, but I just can't find it in my heart to care. C'est la vie! Mar 2, 2020 at 16:13
  • Frame-challenge answers are in general quite OK here (and with me personally). The issues I see/saw with your answer is that it didn't go much beyond reiterating what could be called "Trump roasting", which has gotten better-sourced content mod-deleted on this site. (I see you've already found one of those related meta-topics and even posted your own answer to it.) Mar 2, 2020 at 17:06
  • @Fizz: When I talk about public figures, I rarely indulge in filtering. That means I make my comments based on overt statements, behaviors, and the natural inductions that can be made from such. Back in the Obama era I frequently got in trouble for pointing out that Obama was a centrist and statist with hawkish tendencies, not the progressive liberal he was generally believed to be. Now I get in trouble for pointing out that Trump is demonstrably a clinical narcissist, and that he promotes xenophobic nationalist policies. Mar 2, 2020 at 17:34
  • 3
    @Fizz: The fact that people dislike these comments does not make them incorrect, and does not make me guilty of bashing Trump or Obama. Things like this need to be discussed in order for people to be informed, and the complaints of those who do not wish to engage reality are not a concern Mar 2, 2020 at 17:37
  • By the way, your answer said that Trump's speech almost surely violated a certain German hate-speech law. Interestingly enough, it seems that no tweet of his was filtered out in Germany under the related NetzDG law: politics.stackexchange.com/questions/50536/… Mar 2, 2020 at 17:37
  • 3
    @Fizz: that's because Trump is not a German citizen, and did not (to my knowledge) make any of these tweets or comments while on German soil. German law does not extend to the actions and statements of those outside its jurisdiction. Mar 2, 2020 at 17:39
  • Your section on polling seems to completely disregard the possibility of open-ended questions. E.g. ‘Please rate Trump as a president from 1 to 5’ – ‘1’ – ‘Can you give us some reasons why you consider Trump a bad president?’ The art is then grouping freely voiced reasons into common statements that can then be published as a poll response.
    – Jan
    Mar 3, 2020 at 7:01
  • 2
    @Jan: have you ever administered a large-scale poll or survey instrument? Open-ended questions are difficult to design well, and require an inordinate amount of human time and effort to done, sort, and analyze. It's not like you can dump an open answer in to SPSS and hit a button to get a result. Further, an off-hand question like that can seriously skew a data set. People who dislike Trump may be energized, and more likely to complete the process; people who like Trump may be offended and inclined to stop. Response rate biases impact all of the data in the poll, not just the open answers. Mar 3, 2020 at 15:11
  • @Jan: Pollsters who use open-ended questions have to put thought into designing them, and even them mostly use them for 'flavor' rather than analysis: e.g., cherry-picking typical or evocative quotes as examples, not analyzing them for meaning or content. Mar 3, 2020 at 15:13
  • 1
    The most aggravating part is that a moderator's unilateral-permanent-deletion powers would be used in such nuanced a case where reasonable people can disagree on whether the answer is a good one.
    – lazarusL
    Mar 9, 2020 at 19:33
  • 1
    @lazarusL: That's aggravating primarily because Philipp refuses to discuss his actions in any meaningful sense. He either explains his decisions didactically after the fact, or refuses to discuss them at all. Based on what he's said, I suspect he's just tired and jaded. I haven't reviewed his decisions enough to know whether this is a persistent problem or just something that crops up occasionally, but either way he ought to find some way of revitalizing an interest in content. This heavy-handed nonsense is reprehensible. Mar 9, 2020 at 19:47
2

There are two unrelated issues that seem to be conflated within this meta-question as asked.

On the topic of answering questions while disregarding what the question says

It has been pointed out that the question from day 1 requested poll data. This means that Fizz wanted some other party to have performed the research, put it together and publish it. Fizz wanted to be linked to these resources. Fizz did not request the respondents on SE to come up with a reasoning, however accurate that may be.

There are many reasons why somebody would ask a question with that explicit disclaimer. I myself have done so though mostly on other SE sites. Generally, this should not be questioned and answers should be given as OP desired.

It is possible and accepted to provide frame-challenger answers. These answers would ideally then point out how or why the original question’s phrasing would make it impossible or very difficult to answer. None of the deleted answers to Fizz’ question did that, they proceeded immediately evaluating public opinion or what they perceived public opinion to be. Thus, none of these answers actually answered or even attempted to answer the question and all were rightfully deleted.

On the topic of permitted wordings versus opinion-based wordings

This meta question ends with:

So my suggestion is that a question that asked 'what is your opinion on x?' is bad and should be deleted but a question that asked 'what is the general public opinion on x?' can be a good question and good answers will often just summarize what knowledgeable people perceive as the public opinion. For example the general public opinion in Germany towards nuclear power can probably be reasonably well summarized by anyone who has read a wide range of German press on the topic but that doesn't give any good source to cite.

I agree that a wording asking specifically for your opinion on a subject is opinion-based. Asking for the general public opinion is a mixed bag. The more diverse opinions on a subject are, the more difficult it becomes.

Taking Fizz’ question, as only 13 % of Germans consider Trump doing the right thing in world affairs, asking for general public opinion is probably fine as there should indeed be a general opinion that can be extracted. In the UK, there is already a 1:2 ratio between those who think he’s doing the right thing and those who think he’s not; this might already be more difficult.

However, given my (probably not fully accurate) perception of the current state of the United States, even a 50:50 split might warrant itself to a general opinion question if the two sides are relatively uniform within their opinion.

Of course, there are cases where if you ask three people you will get four different responses; such topics would not lend themselves well to general public opinion questions. I don’t think a line can be drawn easily between one and the other.

Where a general public opinion question is permitted, answers will be unlikely to include sources as general public opinion is not polled as extensively – which loops us back to Fizz’ question where apparently only agreement/disagreement with Trump has been polled extensively but not so much the motivations of the respondents. Had Fizz not included the request for references, I likely would have upvoted far more answers and I would not have felt the need to point out where answers did not state sources.

You must log in to answer this question.

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