Unintended consequence of cheap downvotes and possible solutions

Question

In our pursuit of great answers, we seem to have forgotten that even good questions can start out a little rough, or even very rough.

Downvotes are very common, and based on the comments following in their wake, they seem to discourage people asking questions more than encouraging them to improve their questions. This issue has been addressed many times here before, one of the best suggested solutions I've seen here is: How to defend new users from unconstructive downvoting

In that post Clockwork suggests requiring comments justifying a downvote. I think that because the normal posters here on SU tend to be less tech-savvy on average than those found SE or SF, for instance, we need to be at least as welcoming as they are. As I understand it, SE requires a justifying comment on a downvote, we should too.

But another issue is the cheapness of the downvote for the person actually doing the downvote. While Clockwork tried addressing the weight of a downvote to the person receiving it, I think the downvote ought to cost more to the person giving it.

Finally, while there are scripts in place that check serial downvoting, as I understand it, these are triggered based on the number of downvotes received by a specific user, not by the number given out by a specific user.

We should be encouraging engagement and improvement rather than drive-by cheap downvoting, and I believe the solution is a combination of all three of these:

Downvotes should cost more to the person giving them. This will encourage conviction that the question is actually bad enough to be worth sacrificing X amount of your points to.

Downvotes should require justification with the intent of improvement. This will force engagement and improvement rather than the casual snobbery they currently embody.

Downvotes should be limited per account per day, just like close votes. For a young questioner, a downvote can be as discouraging as a close vote, and usually come with less accompanying information. Limiting the number of these that can be given out will also encourage people to engage with what they consider poor questions in a more proactive way.

Answer 1

The problem with requiring "justification" is that the best you can do is either provide a drop list of pre-canned reasons, or a text box. There is no way to force a downvoting user to enter a personalized, specific, and clear justification that exactly fits the question and will maximally help the user.

If you go the pre-canned reasons route, the reasons available may not fit every scenario, so people will look at the reasons, click one that seems to be vaguely related to the problem (whether or not it actually is helpful to the user is another matter entirely), and move on.

If you go the text box route, users can, and WILL, enter something like "jfafafgasrgaffoq" or anything necessary to bypass the filter. No algorithm can be written that will ensure that the text the user types into the field will be helpful to the OP. For example, I could write, "The quick, brown fox jumps over the lazy dog." -- even if you spellcheck it and deny it if too many words don't match, that one would slip right through the filter, no problem. Then you'd just put that on the clipboard, and off you go. You can play whack-a-mole all day with the automated filters on this, but people who are determined to downvote without providing a constructive comment, will find a way to do so.

Also, if these justifications are going to be provided as comments, then this removes the anonymity from voting. In my experience, I've found that certain people tend to get particularly angry when their questions are downvoted. If you post a comment that is generic and unhelpful (as in selecting from the dropdown or radio list), the user tends not to appreciate it. The only comment that is truly helpful is one that is personally hand-crafted to the question, pointing out exactly what the user is missing in their post, which we can't force people to do. The people who want to do that are already doing it, and those who don't have the patience or frame of mind to do it, will never do it.

You can't design a technical system that interacts with human language or behavior in more than the most simplistic ways, because once you start dealing with problems that have to parse (understand) natural language, the whole problem space explodes into an intractable problem. You can make heuristic corrections to the system daily, and hundreds or thousands of comments will still slip through the cracks on a daily basis. Consider, for example, the futility of email spam systems. Tons of spam still makes it through, once your email address is out there.

The problem you identify -- users not offering to help querants with their question to improve it so that they include all the relevant technical details needed in order for the question to be answerable -- is a valid problem, but it cannot be solved with a technical solution. It is a social problem that needs to be solved with a social solution.