Public Information Writeup: Wrong Answers Are Worse Than No Answers

mcmonkey

Administrator
Staff member
Helper
Wrong Answers Are Worse Than No Answers!



Short version: Don't answer questions from users if you don't absolutely know the answer for sure. It's much worse than simply waiting for somebody else to answer.







Long version:

If somebody linked this to you, it's probably because you're answering questions / trying to help, when you don't actually know the answer yourself.

This happens quite often on our Discord, and almost always comes from a place of purely good intentions. Somebody needs help, and you're trying to help! Whether you're just a helpful person, or you feel an obligation to give back to a community that has helped you, it's of course a very good and understandable thing to try to help. We (the team here) of course love when users are able to help other users - in fact, our helper team is entirely made up of people that joined as new users and simply started giving so much good help consistently that we offered them a rank and a name color to symbolize how helpful they are!

Unfortunately, there is a difference between intentions and reality. You want to help... but if you're telling people things that are wrong, in reality you're not helping. If anything, you're doing the opposite.
When somebody receives a wrong answer, they spend time and effort attempting to achieve what you told them is the way to go, only to get frustrated when it goes nowhere. Users might even grow generally frustrated with the project or the Discord community when they ask for help and get sent down rabbit holes that take them nowhere.

Here's where that "wrong answers are worse than no answers" fits in: you might think that if your uncertain answers are right 70% of the time and wrong 30% of the time, that's 70% very good and 30% neutral (after all, if you give a wrong answer and it's later corrected, what harm is done?)... but in reality, that's 70% slightly positive and 30% very negative.
That 70% of right answers, you're helping a user figure it out 10 minutes (or however long) faster than they would have if they had to wait for somebody more experienced to answer. That's mildly positive, but not incredibly so.
The 30% of wrong answers, however, is times when users are getting lost, frustrated, and confused - wasting time and effort pursuing things that won't work. Time and effort they could have spent better on other parts of their work while they simply waited for a good answer. Even worse, however, is the times when there isn't a correction to the wrong answer later on. Far too often, users receive an answer that "works" (with heavy emphasis on the quotation marks) - answers that seem at the time to achieve what they need, but have unintended consequences or unknown imperfections - refer to The "It Works" Fallacy for more detail on this.

So, what can you do to avoid this problem?
  • Know what you know! It's important to have the awareness of your own mind to know whether you know something, think it, or it's purely a guess.
    If this part might be difficult for you, refer to Footnote 1.
  • Don't answer when you don't know the answer! If you think the answer, or you are guessing it... just don't. Don't speak. Wait for somebody who does know to come along and give the proper answer. If you really want or need to give an answer to that user, get to researching. Find the true answer and confirm it's right - in the documentation, by asking somebody to help you help the user, or as a backup plan, test your idea.














As an additional note, because I know somebody will make a snarky comment about helpers sometimes giving "I think" answers:
A key point of this is predicated on the fact that there is somebody 'higher up' who can give a better answer. When you, a user, can't answer... a helper can come along and answer instead. Helpers, however, don't always have a fallback. Generally with Denizen questions, I (mcmonkey), am the 'last line of defense' - if everyone's fallen back to me, and I don't know the answer, quite possibly nobody has the answer. As such, helpers, including myself, are not held to as strict a standard here. I will sometimes answer user questions with things I think are true for two reasons:
  • If nobody has answered before me, and I don't know the answer... well, not many people are going to know the answer (or just nobody does). I've from time to time avoided answering questions that I don't have any good answer to but think there might be users out there that do know an answer to, only to wake up the next day to 3 pings from that user or others asking me what the answer could possibly be. I wish I could avoid answering when I don't know for sure myself... but when I don't know, quite often nobody else does either.
  • A helper's "think" is often closer to other's "know". This isn't meant to be egotistical - it's a natural result of years of experience. Even when we don't know the specific answer to the specific question, we understand how all the moving parts inside work, so we can logically conclude the most likely answer with a pretty high degree of accuracy - whereas newer users, even with a decent high-level understanding, often don't know as much about how it all pieces together on the inside.









Footnote 1: If you don't know what you know... don't answer people, at all. Just don't try to give help. It's hard to hear, but... if you can't definitively differentiate between proven facts and casual assumptions, you're not going to do a good job of giving help to others.
If it's a goal/hope of yours to be able to help people, please take time to work on understanding:
  • Where did your idea come from (eg: is it something you thought of, or something that somebody once said, or is it information that you found in the documentation, or...?)
  • How well you know that information (are you certain of it? Have you used it yourself? How trustworthy is the source that first told you it? How certain was that source when they said it - an offhand remark, or a documented fact?)
  • Is that the truly best answer (or just one that technically works while not necessarily being ideal?)
  • How recent is this knowledge (is it actually the current way to achieve the goal, or is it possibly outdated and superseded by a newer, better way to do things?)
  • etc. along the same line of thinking.
While you don't need to be able to know all the answers to these types of questions for every bit of information you give, practicing thinking about this with information you might have given as an answer to somebody is a good way to work towards being able to know what you know well enough to be able to eventually give reliable answers to people.