Is a new Code of Conduct really going to solve the bad rep? Or is it a deeper issue? [closed]

Question

FINAL EDIT

Thanks for the great answers explaining the concept behind SE and its goals.

However, when it comes to the negative feedback (specifically the downvoting) I think - even in light of the explanations - it's still unhelpful.

The same curation goals can (absolutely IMHO) be accomplished with upvoting alone. But it is the encouraged negative feedback that creates the general mindset of negativity you're attempting to combat.

Put in another way; one can't say it's OK to throw stones, and then be frustrated when the general populace can't (or refuses to) differentiate between the stones you want thrown, and the stones you don't.

===Original post===

TL;DR Can a website which has been founded in the idea of supporting negative feedback as strongly as positive feedback - actually solve the now-very-negative attitude of the site with a CoC?

===

I've been a Stack user for many years as an IT and development professional, and have accrued a moderate reputation. I've also been a technical team leader and nerd herder for more than a decade.

While I applaud Stack's attempt to make the site more inviting, I wonder if a new CoC is not just bandaids and pretty wrapping over a deeper issue.

When Stack Overflow was founded in 2008, there hadn't been much research into how to manage the huge volume of information that passes through forum-style sites. The idea of a community-corrected site, using volunteer moderators - and where both both positive and negative feedback was employed - was highly successful.

However, the encouragement of the "negative feedback" mindset is - inarguably - what has led to the general feeling "unfriendliness" as seen by the internet as a whole. I was told directly by a moderator (during a Meta discussion) [paraphrased from here, see context in footnote*] that Stack's user base should be viewed as children who need to be punished for their stupidity (in order to improve their questions). This mindset is not unique in Stack.

I understand the need for a method to bring solid posts to the top, and drive poor posts to the bottom, however research suggests that positive feedback alone is more than sufficient to cause the same effect.

Quora and Expert's Exchange (before it went commercial) are prime examples of sites where quality posts cleanly rise to the top, and avoid glut and redundancy, without using a negative feedback model.

=== In conclusion... The bottom line is; many highly qualified and experienced members of my teams in the past loved Stack for when Googling a problem, but loathed asking questions - due to the inevitable hassling received for pointless and pedantic reasons. This is good for no one.

So, when the "unfriendliness problem" of the site seems to lie closer to its core mindset - will a sticker on the top saying "Be nice" really help?

• It was suggested this paraphrase was out of context. It is important to state that this is a paraphrase of the moderators opinion and not an exact quote. However, from the lengthy conversation that moderator and I had - I veleive firmly this is full this opinion of the user base.

EDIT === At the suggestion of the moderator, I'm going to add examples here. I hed initially left these out for the sake of brevity.

Down votes

While I understand the reason they were initially implemented - the web (and forum sites) are now well matured and researched. While "downvoting" may offer a small benefit for moving posts to the top and bottom of a feed - it has been shown by many, many sites that it is not really necessary to accomplish this task.

However it is inarguable that downvoting adds to the "punitive", "hen pecking" mindset of the site. There is - for lack of a better way to say it - a mindset amongst many people to criticize at all costs.

So, the question for the site owners and developer is: does the small benefit offered by "downvoting" towards one of their cited goals (curation of the site) outweigh the large destructiveness towards their other goal (site friendliness).

"Getting downvoted" is an especially large complaint, especially by noobs whose reputation my just be starting out.

Over-policing by moderators

The site MUST be policed to maintain quality. However, in recent years, many moderators seem to have devolved to the mindset that policing the site is their primary duty (versus answering questions or assisting users in becoming better users).

While following the guidelines is important, I find many, many questions that are downvoted or halted altogether for reasons of "policy violation that neither reduce the question's quality - nor would have made it better.

In fact, I can think of a dozen questions by team members in the past that following the question policies to the letter (or following the moderators suggestions) would have made the question harder to understand not better.

All regulations must strike a balance between following the "letter" of the law, and the "purpose" of the law. The site seems to have drifted strongly towards the letter, versus the goals.

Redundancy

This one is a difficult one. While I completely agree with the need to reduce a million questions asking the same thing, I have also experienced the site. And it is commonplace that one question in a subject is neither entirely correct, nor does it entirely cover the necessary problem.

In fact, nearly without exception, when I Google Stack only for answers - I find I have to read 2 or 3 very similar answers in order to find the complete solution that fixes my problem.

IN the "early days" when storage was expensive, databases overwhelmed, and Google algorithms imprecise - I would agree more with the redundancy issue.

However with the modern cheap storage, pinpoint searches, and powerful databases - killing a question outright for similarities to others seem unnecessary. Again, I can quote a half dozen questions asked by my team which were flagged as redundant (and closed down) when, in fat, they were completely dissimilar to the question referenced in the "already answered" link. Which relates to the final idea...

No method of review

I understand the moderators are volunteers, and individuals (with good and bad days), and whose time is valuable.

However, many moderators have developed a sense of inflexibility in their kingdoms. And, once a decision is made, there is ZERO recourse. I can understand this also helps avoid arguments. But in cases where the mod is, unfortunately, wrong - This leads to terrible frustrations amongst users, especially when the moderators follow them.

I myself once wrote a question, and had it closed down as redundant. However, the linked answer didn't address the problem. The automated message on the shutdown response itself suggested [paraphrased], "Try asking the question again, explaining how the answer to which it is redundant does not suffice".

I did, and it was immediately (within seconds) shut down by the same moderator saying [NOT paraphrased], "This is not how you make friends on this site."

While, again, I understand people, and bad days - this mindset of "fiefdom" has begun to be more prevalent in Stack. Which directly affects the operator's stated desire to make the site friendlier.

EDIT

@S.L. Barth "What might work better, is if you had references to actual scientific research that shows how less/no downvoting can still keep the poor content out."

Very valid point. While much of my information on this is a few years old (when I was running a web development team), Here is one link that helps support the idea. As all researchers know, "One paper does not a fact make" - but I will continue adding more as I find them again :)

Some more links, both pro and con

The Effects of Feedback and Habit on Content Posting in an Online Community

Rating Effects on Social News Posts and Comments

The rhetorical constitution of online community: Identification and constitutive rhetoric in the community of reddit

Answer 1

You've done a pretty good job of figuring this out, you just didn't realize it :)

due to the inevitable hassling received for pointless and pedantic reasons. This is good for no one.

A significant reason that your co-workers found the site to be so well represented in search engines was many of those things you described as pointless and pedantic. Stack Overflow started with the stated goal of becoming a very high-quality resource, and the entire product was built around accomplishing that.

Since everyone knew what we were doing, we didn't notice how the product itself didn't explain and reinforce that goal at every turn until people that hadn't heard of that mission statement started coming in as the dissonance grew louder. We just reinforced it. That's a dangerous disconnect.

We need the product to set people's expectations. Here's a great analogy, ready? Here's how the first 100,000 users of Stack Overflow perceived Stack Overflow when we started, and even more still today:

Here's how people that just really want a fast answer to their question often perceive Stack Overflow:

What's the big difference, other than one place being a little larger than the other, and, an actual castle? EXPECTATIONS. We're relying on decades old assumptions to continue to guide the user journey and that's a really big problem. Fixing that, the user experience from cradle to grave so people have some inclination that muddy footprints on the floors would be impolite is a top, top TOP priority.

But, it's not just about the downvotes.

People were getting really mean. And I'm not talking about comments like "That's just not optimal, and plain wrong", I'm talking about really mean stuff. And it wasn't just on the main site, it was in chat and meta too. Continuing to have no formal code of conduct in the face of so many reports of wildly hostile or inappropriate behavior is a tacit endorsement of that behavior, and that is simply not acceptable.

We need to get to a place where people don't walk up to something that people have spent tens of thousands of hours working on and see a pillow fort. Those folks need to know they're walking into the halls of knowledge, and contributions should be lasting ones.

But at the same time, we must admit that at least two graduating classes have passed since we started the site, needs are changing, and Stack Overflow isn't keeping up as well as it should. People need to see themselves in it in order to want to treat it with respect and in order for that to happen, we have to make some changes so they can at least visualize a scenario where they'd like to stick around.

The CoC doesn't solve our social issues, it admits that we have them. The real commitment is understanding that we failed to notice that we were still operating on a ton of assumptions and needed a reality check for relevancy.

Stack Overflow is always going to make people a little nervous, but we need to make it (again) be for really good reasons. We also have to deal with our industry as it is, which (right now) is experiencing some pretty painful growth pains when it comes to inclusivity and equal treatment, and we need to be there for everyone.

So yeah, it's a deeper issue, but in no way did we expect (or predict) that a simple social contract was going to change everything. That's just the start.

My prediction? Once our software sets expectations much more accurately, you'll see a lot less pent-up frustration leaking out at any convenient vent, and then we can get a better idea of the size of the problem that remains.

Answer 2

While I applaud Stack's attempt to make the site more inviting, I wonder if a new CoC is not just bandaids and pretty wrapping over a deeper issue.

It is neither. The CoC should be thought of as a part of a process, not the end. SE didn't go, "hey, we have a bit of a culture problem. Let's make our rules more clear; that will totally fix everything." I expect that there will be other things, but I also expect that the core aspects of the site are not going to change.

For the following reasons:

When Stack Overflow was founded in 2008, there hadn't been much research into how to manage the huge volume of information that passes through forum-style sites.

First correction: Stack Exchange is not a forum. We aren't here to promote discussion; we are here to create knowledge. Preferably accurate knowledge.

So, the question for the site owners and developer is: does the small benefit offered by "downvoting" towards one of their cited goals (curation of the site) outweigh the large destructiveness towards their other goal (site friendliness).

... yes. Especially since the benefit is not "small".

Being able to downvote bad content, rather than just not upvoting it, is vital towards the goal of creating useful knowledge. If someone answers a question with code that is flat-out wrong, or actively dangerous, there needs to be an immediate indicator that the code is in fact wrong and/or dangerous. Not some comment thread you see after you scroll down a bit; it has to be the first thing you see.

Such determinations are not "opinions" that can/should be shrugged off. Especially if lots of people hold the same view. And considering the consequences of disseminating bad/dangerous advice in programming, there must be some method of quality control. Not merely for good answers to rise to the top, but just as much for bad answers sinking to the bottom.

When I'm looking for questions to answer, I want to be able to know if it is a worthwhile question. The question score serves this purpose. It prevents me from wasting time with garbage, poorly-thought out, or otherwise useless questions.

It makes my life as a provider of content better, since I spend more of my time with well-conceived, interesting, and useful questions.

This is not a "small benefit"; it is a big part of why I am here. It services the primary goal of the site, and it does so effectively. And it makes me enjoy my time on the site.

From the comments:

The "marker" is, indeed, the num of votes that is has. And I do use that marker. However, the marker works provably just as well with "upvotes" only and - even if downvotes are arguably useful

Consider the following case:

You see a question with two answers. One has +10, the other has +2. It is obvious to you that more people think the +10 answer is better. OK, fine.

Now imagine the same question, only with just the +2 answer. The natural conclusion, since the answer has been upvoted twice, is that this is a decent answer to the question.

But if we're living in your "upvote only" world, that answer may in fact be terrible. People give upvotes to bad content; the concept of "pity-upvotes" is well-documented. So a bad post getting a couple of upvotes is hardly unreasonable. And there are cases where people just sometimes don't see the bug. It happens.

Any number of people might have recognized the bad answer for what it is. But without the option to downvote the post, there is little they can do.

Therefore, any answer that has no competition is equally "correct" in your world. And if every answer is correct... none of them are.

Oh yes, we could comment on the post, but commenting invites discussion. Many people can't handle being called out on their own mistakes on the Internet, and will argue fervently that their post is fine. That's a waste of everyone's time, when you could just stick a "this is terrible, don't do it" sign on the post and move on.

Furthermore, comments appear at the wrong place on a post for this purpose. Note that the vote total appears at the start of a post; it's the first thing you see about it. Comments appear at the end. So even if you do deign to take the chance on a comment thread argument, there are many people who will never see it. They see an upvoted post, they read the text, and the damage has already been done. A comment might correct it, but there's no guarantee of that.

Voting, both up and down, is a more efficient and effective method of communicating the quality of a post.

Your "upvote only" world is even less useful for questions. At least with answers there is the possibility of competition, with a default sort ordering based on votes. With questions, how can you tell if, behind a particular link, you will find a genuinely well-considered, well-stated question, or a piece of time-wasting filth? You don't.

+2 means nothing on questions.

You'd have to come up with ad-hoc methods for estimating quality, like comparing the time it was posted to its score. If it was posted 2 hours ago and has a low score (<=2), then it's clearly crap.

... unless of course the question is posted on low traffic tags. In those cases, the question may simply not have been seen by very many people. So that assumption breaks down. I see 0, +1, +2. How do I know which ones are good and which ones happened to be visited by particularly gracious users who just "like" everything they see?

With downvotes, even in a low-traffic tag, bad questions will still get one or two downvotes, which is enough to let you know what is not worthwhile.

Your system only works if we assume that everyone will upvote everything that is at least minimally passable. That's not a good world even if everyone does it, because it divides everything into only two groups: bad and not-bad.

"Good" has no place in such a world, because you cannot tell a truly exceptional post from a "not-bad" one.

No, at the end of the day, an upvote-only world is a terrible place to be if you actually care about the quality of content.

Back to the post:

(versus answering questions or assisting users in becoming better users).

Second correction: our goal is not to "assist users in becoming better users".

The Stack Exchange Q&A model is designed to minimize interaction. Think of it like if you scrawled your question on a bathroom stall. Some time later, someone else comes in to do their business, sees your question, and decides to answer it. And later, when you come back to do your business, you find an answer.

That is the ideal of how SE is meant to work. We are not here for hand-holding or to provide personal assistance. The more of that we do, the further our model breaks down.

Indeed, this breakdown a big part of the negativity problem on SO. If a user asks a question poorly, then two things happen. Answerers get annoyed because they have to deal with yet-another bad question. And the asker gets annoyed because they're not getting what they need.

If the asker had just asked a good question to start with, if they had done their research and made a good-faith effort on their own to solve the problem first, if they asked a complete, answerable, properly-scoped question, none of that would have happened.

While following the guidelines is important, I find many, many questions that are downvoted or halted altogether for reasons of "policy violation that neither reduce the question's quality - nor would have made it better.

In fact, I can think of a dozen questions by team members in the past that following the question policies to the letter (or following the moderators suggestions) would have made the question harder to understand not better.

All regulations must strike a balance between following the "letter" of the law, and the "purpose" of the law. The site seems to have drifted strongly towards the letter, versus the goals.

While I agree that there is a bit of lawyering that goes on when it comes to site policies (usually attempts to stretch the interpretation of close reasons so that it covers stuff that the person doesn't want on the site), I don't think this problem is as big as you suggest.

Yes, there are times when someone reflexively asks for an MCVE for a question that doesn't warrant one. But these aren't that common, and someone else usually counters such requests.

In fact, nearly without exception, when I Google Stack only for answers - I find I have to read 2 or 3 very similar answers in order to find the complete solution that fixes my problem.

That is how the site is supposed to work. If your problem is not exactly identical to the OP's question, you may need to see several different approaches in order to massage it into the form you need.

Encouraging users to think for themselves is not a bad thing.

IN the "early days" when storage was expensive, databases overwhelmed, and Google algorithms imprecise - I would agree more with the redundancy issue.

Dupe-voting has nothing to do with storage. It has to do with not destroying the will of users to contribute.

See this question? It gets asked at least once every 2 weeks in my tags of choice. It gets asked in different forms, but it's always asking for the same thing. It's just that the OP probably doesn't realize that.

Now, why should I (or anyone else) waste my time providing the same answer to all of them?

By closing them as duplicates, the asker gets the information they need, and the answers don't have to waste their time repeatedly answering the same question over and over again.

Having to answer the same things is one of the ways information providers on forums burn out. They get so disgusted with saying the same stuff that they just leave. Stack Overflow prevents that.

This site has become what it is by making the lives of those who provide answers easier. Undo that at your peril.

I myself once wrote a question, and had it closed down as redundant. However, the linked answer didn't address the problem. The automated message on the shutdown response itself suggested [paraphrased], "Try asking the question again, explaining how the answer to which it is redundant does not suffice".

Then the automatic message should be corrected.

citation

I could not care less. You can site whatever studies you want, but SO is not a forum. The negative aspects of the site are important.

Now, they don't necessarily have to be done in the way they currently are, but there needs to be some equivalent for them. However, the Powers that Be are unlikely to radically revolutionize everything about SO, so you're going to have to live with it.