How to improve the problem with identification questions?

Question

As many of you know, one way or the other this site has a problem with identification questions.

The large and rising influx of identification questions, which are often of rather low quality, steadily keeps on deteriorating the image of this site and the brand of what it stands for, to which identification questions have always been considered rather a "necessary evil" than the bread and butter of what this site is about.

The rather low votes those low-quality questions gather in turn lead to many of those questions (often asked by new users) ending up negatively voted and/or closed. This, while ultimately a consequence of the questions' quality, leads to an image of the site being "unfriendly" to new users, especially since new users might not always grasp the workings of SE and the meaning of downvotes immediately and might interpret them as more personal than they are. This is not a pleasant experience for those new users and neither does it draw the picture of an inviting and open community.

Add to this, that those reservations against supposedly scaring away new users have recently lead to a counter trend of giving mere sympathy upvotes to rather low-quality, and therefore downvoted, questions in order to encourage the new users. While the motives for such votes are pure, they in turn also have a bad effect on the site as they make it seem as if there was a rising sentiment in favour of bad questions and actively encouraging posting such questions, which is again detrimental to the site's image and ultimately its quality.

In addition, those identification questions, be they good or bad, but especially when lacking in detail, often also encourage rather unexplanative and bad answers that rarely have a motivation to be improved with further details (because, well, they already answered the question afterall) and thus again lead to downvotes, which again shed a bad light on the site, its community, its quality...

All those aspects come down to the problems of the many low-quality identification questions we face each day and that we need to tackle in some way or another. We do already have some measures for tackling them in place, which have their individual shortcomings, though:

• The help center as well as the tag wiki excerpt try to encourage giving as many details as possible and provide some guidelines as to what we're looking for in a proper identification question, but of course this requires reading those not immediately accessible guidelines.
• There is an automated popup in place that appears whenever someone tries to ask an identification question. However, it has some problems on its own, like only being shown for one tag (identify-this-movie), not featuring all the information we might want to know, or only appearing when the tag-edit loses focus (which is usually right before clicking the Ask button). And in general it is just not configurable at all without employing help by the SE übermods.
• There is a specific close-reason for identification questions that significantly lack effort. However, it's specific interpretation is not always entirely clear-cut and requires some personal judgement sometimes.
• There is a manual maintenance task for removing old inactive unanswered identification questions. This however is only a mere cleanup task long after the fact.

Some of those measures we have in place only act after a bad identification question was asked and thus fail to tackle the problem of low-quality and thus low-voted questions and the resulting impact on site quality and user discouragement. Tackling the problem of bad identification question before they are asked might even be better (albeit much harder, of course).

I'd thus hereby like to encourage the community to share any possible ideas on how we can further improve tackling the general problems we're facing from those identification questions.

Answer 1

As someone who's seen a lot of ID questions, I have no doubt there are plenty of Good ID questions (though ID detractors might imagine Good ID in quotes). But it's almost equally evident that Bad ID (brief, 1-2 line questions lacking basic details) has been increasing, sometimes coming in waves as if something triggered it. Yes, we can improve them sometimes, the same way we do with problematic non-ID Qs, but the recent amount of these is frustrating.

So I support Napoleon's call for preemptive measures. (I won't discuss ID's place on M&TV in general as I feel we've done this already. Let's say I believe in fixing issues, not in band-aids nor amputations. And if we can treat the problem at the source, all the better.) So here are a few ideas I came up with for such measures; they're currently tentative suggestions as I admittedly have little knowledge about implementing them, so any feedback will be most welcome (and I'll improve these accordingly).

• A more aggressive pop-up message. As much as the initiative was appreciated, I agree with Napoleon and feel that it wasn't fully implemented. The message is tucked down and to the side by the tag line, is literally in camouflage colors, and might not contain enough information. Changes like a stricter tone (like a warning that undetailed Q's would be closed), a stronger font, design and placement and a more detailed list of requirements could make this more effective.

• General warnings in the question form. Similarly to the previous suggestion, I think a more aggressive tone might deter sloppier ID Q efforts. Brief reminders in the righthand box, or even above the title, that ID Q's should be detailed as possible and ones that aren't would be closed.

Possibly Pipe Dreams (I don't even know if these are doable):

• Including ID instructions in the Tour. AFAIK, detailed instructions about ID Q's are only located in the bottom section of a Help Center question and in the tag wiki. However, I'd hazard a guess that the Tour is a more popular stop for new users [a shiny badge for completing it doesn't hurt ;)]. It seems odd that we keep expecting ID OPs to know exactly what ID Q's should look like when the information is tucked away like that. So I wish we could add a short, illustrated guide to ID to it, to explain what's expected of them... But I don't currently know if we can even touch it.

[Update: Napoleon informed me that the "Ask about\Don't ask about" section can be altered by mods. That seems like a good possibility; even just adding a line to the "Don'ts" (like "Identification questions that are too brief and lack basic details like year, language, genre etc.") might help a lot. And the Tour won't only show ID askers but also ID readers (who sometimes answer, comment and vote on them) what good ID questions should look like.]

• Overhauling the question form (though in the least intrusive manner possible). This is probably the best idea I have, and probably the least feasible. A brief scan reveals that SE question forms seem to be pretty uniform. But I believe that integrating ID into it might solve a lot of the problems we've been having, if it's done in a way that won't disturb others. Like an 'ID questions' button that forwards the asker to a bottom section or another form, which will then include mandatory fields, mandatory ID tag buttons, and a strict minimum character count in the message body. This will solve many of the quality issues in ID Q's, but again, has to clear some mighty hurdles to work.

[Update: I did some reading about this on SE Meta, but unfortunately couldn't find a relevant answer. But the more I read, the more infeasible this seems, I'm afraid.]

In conclusion: I feel that ID is still integral and is capable of drawing many readers and potential users to this site. But due to its unique nature, I also feel it requires some quality (and quantity) control to ensure a moderate stream of serious questions from commited users. Going back to this discussion, I think it raised a lot of good points that were never followed up. And a lack of action on the matter led to the out-of-control state of ID now, where it always threatens to become our most popular tag, instead of being what it truly is, IMO: just another important service we provide.

[Again, I'd love to hear suggestions, corrections and any other feedback. I'm planning on asking about some of these on SE Meta, but I'll have to do some reading beforehand.]

[Update: Unless I'm mistaken, it seems the Identify-This-X Questions in our Help Center was recently expanded with more information. This is most welcome and definitely a start, so many thanks!]

Answer 2

This may be largely a rhetorical point, but will you ever resolve to start banning these questions completely? You have been fighting with this problem for years, but as identify-this questions are surpassing 50% of the questions being asked (yes, I counted), the discussion always seems to be the same: "We need more signs with bigger letters so that people will listen louder."

The way I see it, this is a failed experiment. That's okay; it happens, and it is totally okay to own up to that. The vast, vast majority of "what was that movie I saw..." questions are terrible. Trying to save them out of some sense of preserving those rare (albeit, still mediocre) gems does not make sense to me. I don't get it. You don't need these questions; you don't need that traffic.

Respectfully, I am not going to respond to another round of how beneficial these questions actually are. We've had that conversation. If a city has a growing graffiti problem, I just don't see the benefit of advertising better art schools. I half-jokingly considered closing this question symbolically as a duplicate of how many times we've had this discussion before, but I don't think this issue has been resolved. So I'm taking this opportunity to point out that there's nothing wrong with creating a thoughtful close reason to say we once tried this, but it simply did not work out:

"...interesting question, but unfortunately questions asking to identify a particular movie, TV-show, or actor just do not work well on this site."

And with that, I'm out. Good luck!

Answer 3

My perspective

I mostly lurk here, and definitely enjoy reading many of the better questions and answers. However, as you noted in your question, "identify this" questions are typically not good quality. Sometimes I have been introduced to new material to watch or read from those questions on this site and at SciFi.SE. But there are reasons why I largely just ignore them:

Be specific

If you ask a vague question, you’ll get a vague answer. But if you give us details and context, we can provide a useful answer.

As has been hashed and rehashed, many of these questions are vague. That makes them difficult to answer, and difficult to relate to: maybe I am curious about the same movie, and your "identify this" question might jar loose some memory from 30 years ago and pique my interest. But if the question is vague, it will only jar loose my desire to downvote the question and close the browser tab.

Make it relevant to others

We like to help as many people at a time as we can. Make it clear how your question is relevant to more people than just you, and more of us will be interested in your question and willing to look into it.

The vagaries of the previous point make these questions less relevant to other people (e.g. me). If I have no idea what the movie or show is, or even any clear idea what picture you have in your head, why do I care?

For example, a good question might clearly describe a famous scene from a classic horror movie. Just to pick on one of my favorite decades, the 1980s had a ton of horror movies with many memorable scenes. A clear description of something I might have seen and would be interested in would definitely make it relevant to me (and others).

---

What do I think should be done?

Why not borrow an idea from Stack Overflow and implement triage review here?

If...

• A question is tagged with any "identify this" tag;
• Is asked by a user with less than 200 reputation;
• Or the user has less than 3,000 reputation (enough to have close votes) and has not asked an identification question before...

...then put it in the triage.

The triage allows users to help filter these questions without needing to cast votes. The system can approve or disapprove these questions, and users can edit them to help improve them before they hit the front page.

This will help prevent the following users from tossing low-quality identification question onto the front page:

• New users, including those with association reputation;
• Users who have been around and made a few positive contributions but maybe have not asked this type of questions before;
• But not users who are trusted with close votes.

Yes, this would require programming assistance from Stack Exchange employees, but I believe it would help quite a bit.

Answer 4

Realistically I don't think you can fix the problem with more popups/educational noise on the questions. People who ignore the current warnings will just continue to ignore the more obnoxious warnings IMO.

Based on Robert Cartaino's answer, and reading between the lines in your question/comment I think we need to do one of two things:

1. Aggressively close identify-x questions that don't meet the warning's requirements.
2. Ban the questions outright, and close them with the standard text that Robert Cartaino suggested.

Personally I think Option 1 is the approach to take at this point. If this approach fails to change things then Option 2 is a logical next step. That being said, I'm new here and I get the feeling this war has been waged for quite a while. As such jumping to Option 2 seems reasonable too.

Answer 5

I hope you'll forgive me if I end up repeating some of what has already been said but there is a lot to read here and I think I've gathered the main gist. I wanted to add my two cents about ID questions and start by saying that I think it is the responsibility of more involved site members to coach newer members and help them and give them the opportunity to improve their questions. I think that Movies and TV has done this very well from what I've seen in my time here.

My personal philosophy and approach to a bad ID question is to ask the user for more details and try and point them to site guidelines. If no effort is shown by them to improve or learn then I think a downvote and question deletion is justifiable.

In regards to the everlong debate of ID questions being on or off topic, I have always felt that this should be determined case by case. There are good and bad ID questions in any SE site. Unfortunately, I participated for a while in one SE community that determined to simply remove all ID questions as off topic. I asked a question that was very specific and detailed and that I'd been researching for years without luck in hopes that someone would be able to help me find some more clues in identifying it. This turned me off completely to the SE site and I've stopped visiting it.

It is not our jobs to be a bunch of geek nazis. We should always balance improving site quality with providing a good experience to casual users and hopefully helping them learn and become more valuable members of our community.

End Soapbox Rant.