Stack Exchange has expressed an interest in increasing the reputation earned from question upvotes from 5 to 10 on all sites.
Do you think this is a good idea?
Answers are more valuable than questions because they require knowledge, giving more reputation to answers rewards learning and acquiring that knowledge.
Users asking a question (trying to solve problem) will get answers besides the reputation, a solution to their problem which is a reward on its own.
Reputation unlocks moderation powers - reputation gained through knowledge means moderation will be done by more knowledgeable people and that reduces potential errors.
We don't need to add more incentive for people to ask questions, especially not on Stack Overflow, there are already way too many poorly researched and poorly asked questions.
Questions often get upvotes because of "me, too" problems. Someone has a problem, finds the answer on existing question and upvotes both even though question might be less than stellar or even rather poor.
As the end result, giving same reputation for questions as for answers will only reduce overall quality.
If there are specific sites in SE network where having +10 reputation on questions would be beneficial, then such change should be discussed and applied only on those particular sites, not SE wide.
Simply increasing question reputation as stated will make it easier for sockpuppets/voting rings to start. It's rather easier to copy a nice looking programming question from Quora / Reddit and post it on Stack Overflow (maybe changing a few words to avoid the simplest plagiarism checks) and after two upvotes (instead of three) you'll already be able to upvote yourself. Since those votes can come from the ring themselves, their growth rate will be more than doubled, it seems. This will make it harder for regular users, ♦ moderators and Community Managers to detect and take care of them.
So my two cents: it could work but only if the reputation needed for the upvote privilege increases as well.
There's one site in the network where a question upvote already gives +10 reputation: Stack Apps. Incidentally, it's also the site with the highest Good Question : Good Answer badge ratio, as this SEDE query shows. It doesn't have voting rings since it's small enough to moderate effectively.
(You recognize this? That's correct, it's a repost from another, now deleted question.)
Weighing up the pros and cons, I'd have to say overall "no".
It's very site dependent. Compare:
Stack Overflow and Math.SE attract users who repeatedly post poor questions, and slowly accumulate reputation. This change may encourage them, so I'd expect it's not beneficial at these sites.
At Chinese.SE, I can spend 20+ minutes writing a question, only to have it answered in seconds by a native Chinese speaker. Moreover, at small sites there's issues with having enough users with enough reputation to cast 5 close votes. I think it would be beneficial here.
At Islam.SE, which is also a small site, writing decent answers take a lot of research effort. I think it's safe to say that answers are more valuable than questions.
At Skeptics.SE and CodeGolf.SE, writing good questions is quite difficult. Perhaps it would be beneficial here.
I expect voting patterns would adjust accordingly. Knowing a question now gets twice as many reputation points would possibly lead to users voting on questions half as much.
I feel it's better to put more effort into encouraging users to selectively upvote worthwhile questions. Make them stand out on the front page. Draw attention to good, effortful questions.
Answers > Questions
Answers are more important than questions because people (including ones who posted a question) are coming for the answers. Nobody is interested in a wonderful question that has no answers. Someone can argue that an awesome question will inevitably get interesting answers but that's actually means that by definition valuable question is something that has valuable answer.
That said, I as an active user who answers a lot and actually asked questions quite a lot never was annoyed by 5 point for a question up-vote.
Any changes should be justified
We need to address first what exactly problem we have and will try to solve by increasing the value of a question up-vote. In my opinion we don't have a problem at all. It's not like someone was going to ask an interesting question and then like: "Hmmmm, 5 points per upvote just doesn't worth it".
Although it can be frustrating to ask a good, well researched question and get only 5 points, or a multiple thereof, and see a not particularly good answer get a multiple of 10 points, that is life, and with enough activity, it evens out. I'm not going to repeat the points in the other answers, except to say that writing a good answer is, on average, harder than writing any but an excellent question. I see this as "solving" a problem that does not exist.
It would be helpful if the OP told us what problem SE thinks the change would solve.
If a site's community wants to change a question's upvote value for that site it could be beneficial, but it's not clear all sites would benefit.
The issue isn't what to do with bad or even mediocre questions; SE has multiple ways of dealing with them, and bad questions tend not to get many upvotes. Good question asking varies widely in difficulty based on site nature and culture. For example, devising a puzzle is easy, but devising a good puzzle can be quite difficult. It wouldn't be unreasonable for Puzzling to value good questions more highly than a site where composing a good question takes less effort.
Other sites have a culture where most questions die quickly. Merely making the minimum standard to survive can be quite challenging. It isn't unreasonable to give more reputation for writing a question that manages to make it on such a site.
Questions need answers and answers need questions, but Stack Exchange culture values having fewer questions over having many interesting but unanswered questions. Having a higher numerical reputation value for quality answers encourages some users who would otherwise write questions to write answers instead. It could be interesting to experiment with a "congestion pricing" style of valuing questions relative to answers, but I would imagine that too few users would pay attention to the current payoffs to make it worthwhile.
The meaning of upvotes is heavily overloaded, in practice if not in theory. Question upvotes are even less strongly tied to quality than answer upvotes. On the technical sites, even a barely adequate question will garner upvotes if it's found by users with similar issues, while a well-posed question about an obscure problem is unlikely to get many upvotes at all. While there is certainly value to the site having the solution to common problems (and that requires questions about common problems!), stumbling into a lot of common problems doesn't seem to be what reputation was intended to mean.
No, because increasing the value of an up-vote doesn't matter if nobody votes on a question (or answer).
One of the purposes of a site like Stack Overflow was for users to treat it a bit of a "developers notebook" in which problems were documented along with their solution. While this can be useful for generating useful information over the long term, obscure problems are rarely rewarded.
Case One: Recent question, self-answered, with zero up-votes. However, it resulted in a patch to a popular library.
Case Two: Old question, 45K views, likely getting a lot of "drive by traffic" from Google. However, limited votes (10) on the question and even the top answer doesn't have that many votes (21).
Case Three: Old question, 4k views, high up-votes for both the question and top answer. Compared to Case Two this tends to imply that the value of a question isn't tied to the popularity (as measured by views) of a question.
In short, there is a lot of evidence on Stack Overflow that votes are inconsistently applied and good work for obscure problems isn't necessary rewarded. As such, increasing the value of an up-vote on a question isn't going to resolve an underlying cultural issue.
I generally agree with @Glorfindel's answer, but have some additional context. The main issue here is the large difference between the cost of a downvote and the benefit of an upvote (to the poster receiving the vote). I think this concern is highlighted by looking at some posts the community has strongly disapproved of in the past:
This illustrates the effect of changing the ratio of the impact of an upvote to the impact of a downvote from 5:2 to 5:1. As "rough measure of how much the community trusts you", this does not seem to match with reality at all. Yes, they are internet points, but they are designed to reinforce behavior and control access to site privileges, and should produce a meaningful result.
Note there are reputation change caps that effect the net rep seen by the user, especially in these extreme cases (e.g., the user who posted An apology met both the upvote and the downvote caps at various times on the day it was posted). These caps aren't considered in the numbers above.
I lean towards "no, it should not be increased", that's probably a bad idea, unless...
First why not: because I don't think the common assumption that increasing rep earned from questions means "more incentive to ask better quality questions" is necessarily true. The incentive is probably just to ask more questions, quality be jiggered. Increasing the rep earned may have the opposite effect and may actually incentivize bad questions.
One of the important things the +5 rep weighting for questions does is keep it closer to the -2 rep weight for downvotes. This keeps serial question abusers (which are much more common than answer abusers) from making too many low quality posts and taking advantage of the community. If the delta value was to change from Δ3 to Δ8, that would make it firmly in the best interest of anybody to keep posting as many low quality questions as they felt like without looking back. The vote weight disparity is less of an issue for answers because there are fewer other incentives for posting bad answers. Sure some people do it for the virtual internet points, but the "do my homework for me" and "I'm too lazy to debug this problem myself" questioners have motivation coming form outside the point system. The reputation should roughly reveal or equate to what they are putting into the system, and a bunch of low quality questions is not something you want to add up to a large rep score.
...now for that "unless".
Given that I think the issue with raising the question rep score to +10 is more about the delta between up and down votes, you could easily fix this concern by also raising the downvote weight to -5. That might actually equalize the incentive for asking good questions vs. writing good answer, but still keep the stakes high so people don't get lazy with their questions. They already expect people to volunteer their effort to answer, there should be high stakes for that.
Do you think this is a good idea?
Not at all. All the true quality folks are welcome to skip this answer, you probably won't like it, as this is pretty emotional, and not so much focusing on the aspects the quality folks uphold.
Thing is: gamification is still a large factor that drives this network. Many people that contribute a lot of content do that to hunt reputation, and badges.
Sure, I would love to make more from asking questions. But changing that part of the system would also change "the physics" of the game.
I do dislike various aspects of the those "game physics", but well, that is what we have here. Same rules for everybody, 10 years ago, 5 years ago, today. Changing that part would feel like tilting the game. Leading to "subtle" questions, like:
Seriously, I try to contribute high quality input on Stack Overflow, but the game part is an essential part of that. Screwing with that will definitely impact the motivation of this top 0.03% contributor, and I think that I am not alone with that.
I'm on the fence about this proposal
Since the reduction of the question upvote value in 2010 was not an efficient measure to suppress low-quality questions, I don't expect an effect in favour of good questions by undoing this change. I expect no significant change in user behaviour here.
So the net effect is giving out a lot of reputation at once and by proxy giving out a lot of privileges at once. It will also make earning privileges easier, since even a controversial question voted +3/-3 already earns 24 reputation points. I don't know whether this is a good thing or not, maybe it will lead to other necessary adjustments like increasing the question downvote value or adjusting the privilege ladder steps.
Yes if and only if we can do something about the "pity upvote." The pity upvote is too common of an occurrence, and the change would make it more likely for an otherwise terrible question to net positive rep.
I don't know what it would be, but maybe if a question has more than twice the number of downvotes, the upvotes should be worth 5. Something like that.
There is a really basic best practice approach that wasn't even broached in the 13 November Stack Overflow employee-written blog post announcing the change (although there were lots of comments about it from Stack Overflow user/contributors): Do a statistical survey of a sample of Stack Exchange users' preference regarding the change from 5 to 10 points on questions. Start with SO. Do not sub-sample on Meta SO users only, or high rep users only. Do it properly, and disclose the inclusion methodology. Well, document it for internal purposes if you don't want to reveal the details to us.
I am mystified why this easy, inexpensive first step wasn't done. Even Wikipedia, a non-profit organization, does surveys of editors before making changes to the editor interface. SO has a data scientist and three coder employees working on this initiative. I'm sure they know how to do a website survey.
In direct response to the question, I do not believe such a major retroactive change should be made to SO let alone all the SE sites, without a data-driven justification. There is none at the moment, not that I can discern despite digging around. Perhaps there is, but the SO employee team did not choose to disclose it to us. A data-driven rationale would enhance credibility and trust from us, me in particular.
I don't want SO to move fast and break things, good things that worked really well for a long time! That is my concern in making this global change on all SE sites. This is especially a concern for SE beta sites. The change in question up vote points will cause confusion for new beta users (and probably have confounding effects on the calculation about whether the site moves out of beta eventually, or is shuttered).
Firstly I will declare that this change will increase my rep by about 50% on Stack Overflow. However, I care much more about most questions being unfit to take up space on the home page, hence having long ago driven away many of the experts.
I think the rot started when we were no longer allowed to close question as "you don't show enough skill to make us believe you can understand an answer". Hence if you still care (I gave up years ago) the only option is to use the voting and closing system in a way that will trigger the maximum number of question bans, remembering that targeting more then 2 or 3 votes at a user on the same day will get your votes automatically rolled back.
But there is hope. At present I get the feeling that many questions get upvoted to 0, that would not have got an upvoted if they were not on a negative value. Maybe the experiment with hiding negative values will remove this problem.
Giving more credit for answers than questions gives out the message that the site cares most about the experts. If this is no longer the case, and the site just cares about maximising short-term engagement then making them equal would be logical.
UPDATE November 18, 2019
In answer to Ellie Kesselman's observation that a survey should have been first conducted, a few days ago, I asked Jon Ericson the following question:
What happened to the SE philosophy of: "This is what we're thinking of doing, and these are the reasons why we are doing it. But before we go through with it [edited], we wanted to know what you thought. Here are some ideas we have come up with”? @Mari-LouA (Nov'15)
[emphasis in bold mine]
@Mari-LouA: I think we still do that, but only when it makes sense. For instance, if you are following Yaakov's work on post notices (or status banners as I call them), you'll see he's incorporating community feedback. This change however is different. It's not really fair to ask for feedback knowing full well we aren't going to change the decision. One of the reasons it hasn't been done before is we know it would be unpopular for many. I don't think y'all needed the drama of us fake-asking on meta.
Jon Ericson♦ (Nov'15)
The change wasn't implemented so that better questions would be asked. It was implemented so that users who post more questions than answers, would feel welcomed and as valued as those who primarily post answers. That is not my opinion, it's an unbiased explanation. As can be seen in Jon Ericson's response, the survey would have been of absolute no use.
The OP (Robert Harvey) asked
Do you think this is a good idea?
Premise: How that change was communicated, abruptly, and without forewarning (ignoring the infamous leak) led to several users posting similar questions such as: “Awarded double yearling badges”, “My reputation suddenly raised by 10k!”, and “Not complaining, but I just gained 500 points without reason”. But for some time, SE has stopped communicating one-to-one with its user base…
You'd think there were no such thing as “sand answers”. Is anyone complaining about lazy answers, which consist of a few lines, being upvoted and their authors earning privileges that they don't deserve? Have users on Meta Stack Exchange (herein MSE) and Stack Overflow forgotten that questions can also be downvoted?
Questions often get upvotes because of "me, too" problems. Someone has a problem, finds the answer on existing question and upvotes both even though question might be less than stellar or even rather poor.
Upvoting a question where you found the answer to your problem, is exactly what the Stack Exchange model is all about. The person upvoting did not need to post a question because they found the answer. I agree, often a question may not be stellar, but it was definitely very useful if it helped somebody.
When a low-quality question is upvoted and earns 10 reputation points, there will be a hoard of disgruntled hi-rep users marching to downvote the post, effectively telling its author and the community that the question is “defective” as it fails to meet the minimum requirements laid out in the Help Centre.
But recently there has been some exciting developments, the newly introduced Question Wizard is meant to stem the flow of no-effort, useless questions hitting the stackoverflow home page in the first place.
From stackoverflow meta, the following announcement earned a very respectable 684 upvotes, with just 10 users opting to downvote the initiative.
- Guided mode means better questions on the site overall.
Based on our experiments, we found that question quality improved when using guided mode compared to traditional mode. In our latest experiment, we found a 5.12% decrease in bad-quality questions, and a positive change in neutral-quality questions (2.26% increase) and good-quality questions (1.12% increase). We also saw an overall 3.42% decrease in overall question volume, which correlates pretty well with the decrease we saw in bad quality questions.
The author feels encouraged, gratified and the person will likely repeat their positive behaviour in order to get (hopefully) great answers as well as further gratification in the way of reputation points. This is precisely the type of newcomer that every site on SE needs and should welcome with enthusiasm. Every SE site needs new blood, if the community aspires to survive long term.
If the aforementioned proposal goes forward, getting two upvotes (20 rep) on a first question will generally send a positive message, if that question also gets a downvote (-2 rep), which frequently happens, the new contributor will feel the sting less. Under the new system, the acceptable question with its two upvotes and one downvote will earn the asker 18 reputation points while the visible score will remain at +1. The number of votes will be unaffected.
What happens if the new
blood contributor becomes lazy and thinks it they can get away with posting homework-like type of questions. Those off-topic questions can still be placed on hold, closed and eventually deleted. New contributors soon discover at their own expense the punishment for submitting repeatedly, low-quality questions.
What is a question ban, and why is it implemented?
As stated in the about links on every page, Stack Exchange is a network of question and answer sites, not help forums. This implies that all posts are expected to have some value for later visitors too. To enforce that, and to prevent help vampires making the answerers turn away from the communities, low-quality questions and answers are blocked. This includes posts from:
- users who can't be bothered to form sentences
- users who don't do the most basic kinds of research themselves
- users who barely even explain what it is they are trying to do An automatic filter is in place to ban questions and/or answers from IP addresses or accounts with a history of extremely poor posts.
To avoid bypassing the filter its internal rules are a secret, but it is partly based on downvotes cast by other members of the communities. If the other members of the site consistently give your posts a low ranking, you should try to identify the reason(s) for this.
Once you have posted too many poorly-received questions or answers, you will be banned from posting more, and you will see the error message.
However, suspension is a much more blunt instrument for dealing with folks who, for whatever reason, simply cannot seem to ever post a useful question. Unlike suspensions, you can "work your way out" of a q-ban: just post some helpful or insightful answers. […]
Philosophically, I like this sort of merit-based restriction a lot better than suspension. Even if in practice both mean than someone can't post questions for a few months, the former puts this decision squarely in the hands of the participant themselves and the community they wish to be a part of.
Shog9♦ (Sep 17 '13)
Overall, questions can be downvoted, placed on hold, closed, locked, and/or deleted. A person asking their first ever question on SO is going to be intimidated by all the roadblocks put in place, including the Question Wizard. Questions are subjected to far more rigour than answers, no wonder many potential good contributors decide to leave the site. Their questions are considered trivial, as seen by the number of points awarded: 5 points compared to 10 for an answer.
On EL&U, I sincerely believe in helping people find good answers to their English language questions, some of the ways of doing that is by editing. Since 2013 I have edited users' posts 3,634 times, and cast a total of 23,799 votes. On my profile page it says I have cast 19,798 upvotes and 1,063 downvotes and offered 96 bounties for a total of 12,950 reputation points. Since I have virtually stopped contributing on ELU, October 9 was my last post, the number of questions and upvotes have seen a sliding trend. It probably signifies that other high rep users such as myself have been disillusioned by the company's treatment of Monica Cellio, and have stopped curating the site we once loved.
I have access to EL&U's site analytics, so for those who thrive on data, here are a few telling line graphs.
The first graph plots the number of questions posted weekly between November 11, 2018 and November 11, 2019. The week beginning December 3, 2018 saw the first peak (469) followed by January 21 and the second week of March: 467 and 465 respectively. However, between September 23 and October 14, 2019 the number of questions plummeted from 332 to 247. The number of questions began to rise slightly in the week of October 21 but thereafter steadily decreased, showing a mere 257 questions in the week of November 4.
The second graph displays the total number of posts (questions and answers) submitted in the last five months: May 10 - November 10. From a peak of 447 answers posted in the second week of May, that number dropped to 294 in the week beginning November 4th.
The third and last graph, shows the number of upvotes and downvotes cast during the same period.
Without questions, a site will over time wither and eventually die. Users will become less invested, visitors will find answers to their own questions but will not be motivated to post a new question because, frankly, the home page with its zero votes and one or two posted answers looks bleak and lifeless. In order to incentivize visitors, it is absolutely necessary that the home page looks productive, we can achieve this by upvoting more frequently. When a question receives an upvote it creates a sort of domino effect, it increases the likelihood of getting a second upvote and fosters an air of positivity. Above all, it shows respect to the new querent. It goes without saying that low quality questions should never be rewarded, but sometimes there are hidden gems, and it's well worth the effort to edit those posts so they can become on-topic and interesting. Add an upvote and users become curious and are gratified when they see the question improved.
A site's home page shows visitors questions with its number of votes. It doesn't show visitors the number of upvotes a brilliant answer received that only appears if they click on the question. If users on SO had edited and upvoted a few more questions and showed some good will, perhaps management would not be thinking of augmenting the points for questions across the network.
Who needs interesting and thought-provoking questions the most? Your hi-rep users, the ones who started their career by posting answers to easy questions, encouraged by the positive response i.e. upvotes, they moved on to answer the trickier, more challenging questions. But today they're not finding challenging questions because users who are able to produce quality questions are not staying.
Who stays? The “help vampires”, the gimme the codez users who demand answers, as if it were their right. “Help Vampires” don't care about keeping a site clean or building a repository of useful Q&As.
“Answers are more valuable than questions because they require knowledge”
TEST DRIVE the initiative on Stack Overflow for two months.
P.S A few things mentioned above were echoed, indubitably more eloquently, in Cody Gray♦'s answer to What was the context of the decision to lower the value of upvotes to a question?
The fact is that the changes made back in 2010 simply weren't having the intended effect of improving question quality. Reducing the amount of reputation gained from upvotes on questions did little or nothing to address the problem of users earning reputation from low-quality questions. Worse, it merely served to make it that much harder for users who were asking useful, high-quality questions to earn privileges.
And for an explanation as to why the value of questions was increased twofold
Recently, increased effort has been targeted specifically on addressing the issue of declining question quality, including […]: a wizard to guide users through the process of asking questions, improved post notices that do a better job of communicating why a question was closed, and better moderator tooling behind the scenes to deal with less-than-stellar contributions.
With these changes, and more in the pipeline, the Stack Exchange team felt it made sense to reassess the reputation system, and roll back an old policy that never managed to achieve its own lofty goals while unintentionally hurting users who participate effectively by asking useful questions.
@Cody Gray♦ (2019, November 13, 18:59)
I am in favor of changing the upvote value to be equal. Here's why.
We hold questions to... pretty high standards. They can't be too opinion based. They can't be too broad. They can't be soliciting recommendations. It has to be self-contained. They have to be on topic. They should show prior research, be clear, interesting, and useful.
That's a lot to ask of a post on the internet.
Writing good questions is hard. Not everybody can immediately walk in and ask a good question. It takes a bit to develop the skill, and even then, you don't always succeed. Writing questions is hard.
And yet, questions are the life blood of a site. Without questions, there can be no answers. The site needs those questions for people to be able to write answers. A question and answer site isn't very useful without the questions.
When a new site is starting out, what's the most important thing for them? Questions. Because without those all important questions, the site does not exist.
Now, I may be biased. I've asked hundreds of questions here on Stack Exchange. A large percentage of them come from me learning about a new subject, having difficulties, and asking questions. I've learned the hard way what happens when you ask under researched questions, or boring simple questions. I have experience with sites such as Puzzling.SE, where the question is a challenge to the community, and designing a question can take months at a time.
I know just how hard it is to actually write a decent question - and it is at least as hard, if not harder, than writing an answer. And questions are of at least, if not more, the same value as an answer - because one cannot exist without the other.
Questions are an integral part of the site that are currently not being recognized the same way answers are. I think it's high time that changed.
Good questions are vital for the success of any site in the SE network.
Without good questions, a group of answerers are merely twiddling their thumbs and making smart remarks on the site's meta.
Some users are reticent to ask questions because smart users apparently shouldn't need to ask any (this is a typical problem with people in tech -- you're not allowed to not know anything!)
You can see some of the high profile answerers will have less than 10 questions among thousands of answers..
Awarding good questions better will help increase their quality (it worthwhile since the multiplier is more rep) and their number (which gives the site better "gdp" or "exchange")
A lot of us were really mad at the great recalc of March 2010. But that actually caused me I think to put more effort into generating answers.
Now the increase in rep for questions will hopefully increase the number and value of questions again, which you cannot deny will help the network.
There are already many good answers, I'll only throw in one thing, that I find changed to a better:
You might think, this is not a big issue. But for low rep users like me, it kinda is.
I find that it gets harder asking question within the last couple of years. If you don't get some upvotes or a qualified answer in the first hour, the question will vanish. Also if you got one or two wrong answers - which happens more often then you might think - even less people will even open the question. I guess they think it's not worth checking then. Also on sites like Code Review with way less traffic than SO it's hard to attract attention.
So, I often tend to award bounties to attract attention to the question and it helped me a lot and it also resulted in some great answers and dicssuions.
This is now easier as you only need 5 upvotes to compensate the smallest bounty instead of the previous 10.
This is really helpful.
That being said, with all the other arguments presented here the answer should still be no, but I just wanted to add at least one positive change.
No, it that is not a good idea.
First of all, Answers are more valuable than Questions - a question helps (mainly) yourself, an answer helps (almost exclusively) others. Thus while we should reward good questions (i.e. those that other people with similar issues can discover and use the answers already given) we should reward good answers more.
Secondly, let's look at bad questions: A downvote counts as
-2 reputation, so currently, if 3 out of 4 people think your question is bad with will negatively affect your reputation (and eventually stop you from asking bad question - i.e. question ban). If a question upvote is now worth
+10 (instead of
+5) 6 out of 7 people need to agree that your question is bad, i.e. bad questions with five downvotes that attract one random upvote will not impact your ability to ask additional (bad) questions.
Finally, it usually is much easier to ask a question (especially a bad one) than to answer one. So answers should be more valuable than questions.