If a site bases best answers on the number of votes those answers get then are they really the best, most accurate answers or just the answers people like the most? Historically, has the most progress come out of what the masses like best or out of accurate research executed and disseminated by people willing to risk the derision of the masses in pursuit of truth?

  • The second part isn't even really on-topic at Meta. Apr 3 '16 at 18:26
  • 5
    @DJClayworth for some reason I'm get the feeling that someone's got an axe to grind...
    – Shadur
    Apr 3 '16 at 21:31

Skeptics is a bit stricter than most sites on the Stack Exchange network because even if an answer is popular, if it doesn't meet strict, specific standards unique to this site in terms of being based on 3rd party verifiable evidence, it gets deleted.

I know this, because a while ago I had a very popular and pretty uncontroversial answer (+24 votes in less than one day) deleted in this way (see discussion). The question was about an aerial photograph of uncertain origin allegedly of the passenger jet that was shot down in Ukraine, which was being used in a highly politicised way. My answer showed this photo was a fake by pointing out details such as, that the wings were measurably at different angles, along with various other signs of having been digitally tampered with. While no-one disputed this, and everyone agreed with the conclusion and argument, it wasn't an answer based on 3rd party evidence, so it got deleted.

I've seen plenty of questions on controversial topics where answers on both sides of the debate have been deleted for not meeting the standards. I've also seen answers at -10 votes or below not deleted because they are based (albeit illogically) on 3rd party evidence (albeit flawed or misinterpreted) - here's one recent example.

  • 2
    Thank you for your gracious answer, it's very well thought out and explained. Having had to delete my own answer at +25, I know it must not have been easy :-)
    – Sklivvz
    Apr 4 '16 at 1:22

are they really the best, most accurate answers or just the answers people like the most?

This is a false dichotomy.

Of course, at the base level, it is the answers people like the most.

But it is not just the answers people like the most. We have an reading audience that is strongly (not completely) biased towards skepticism, that are looking for answers involving empirical evidence and good science to upvote. Where there is no empirical evidence, the answer may be deleted. Where there is bad science, there will be downvotes and comments leading to more downvotes. So, in practice, the best, most accurate answers tend to be upvoted.

Is this system perfect? NO, but no-one is claiming that. All we are claiming is that the results speak for themselves.

This is analogous to Wikipedia. Just thinking about the problems for a second would lead one to believe that the whole idea of crowd-sourcing an encyclopaedia has no chance of succeeding. I'll admit, it lead me to believe that. However, the result, as improbable as it seems, it very high quality. If you ask me, it remains incomprehensibly high, beyond all sense! However, I have to believe my eyes over my own theory. (Flaws remain; and as a tertiary source, it is a poor reference for controversial facts, just like any other encyclopaedia, but I happily refer to Wikipedia many times per day.)

Historically, has the most progress come out of what the masses like best or out of accurate research executed and disseminated by people willing to risk the derision of the masses in pursuit of truth?

I don't know how to answer that definitively, but I suspect that virtually all progress is science is done inch-worm style: One small team of scientists researching one small area, and putting another small brick in the wall of knowledge, without any controversy. When I speak to practicing scientists, they tend to be working on unnoticed areas, building on the work of other people in unnoticed areas, and it isn't until you step back that you realise how much incremental progress that there has been in the last 100 years.

So, it isn't about fighting derision or winning over the masses. It is about offering evidence to other people also deep in the field that they can understand, replicate and build upon.

Of course, by the nature of this site, we tend to address the controversial, politically-influenced and/or popular science issues, so that's not the impression you might get from reading this site.

But the take-away message should be: Go ahead, risk derision, and post controversial answers here - just so long as they are based on empirical evidence and good science. You may get downvoted, but go ahead and try to convince a couple more mavericks to join you, and start the ball rolling for changing consensus.


Votes are not the only characteristic this websites uses to distinguish answer quality. There's also the accepted answer button and the delete feature.

Neither of those mechanisms is perfect. If you want to know everything about a topic it makes sense to read all the answers to a question. On the other hand if you just want to read one answer that's most likely the most informative answer you can read the highest ranked answer. Votes are a mechanism for sorting.


The bigger problem is that there is a bias toward early answers.

If someone gives a different answer years later, especially after an answer has already been accepted, the late answer has little chance of appearing as the "right" answer.

For example for this question: Does marriage between first cousins double the risk of birth defects?

There is one answer "no" from 2013 and one answer "yes" from 2016.

  • The evidence presented by both seems compatible though.
    – Sklivvz
    Apr 5 '16 at 23:08
  • @Sklivvz the underlying references are compatible, but many casual viewers will just look at "no" versus "yes" and which is accepted and upvoted.
    – DavePhD
    Apr 6 '16 at 1:34
  • @Sklivvz Ok, I'll do another experiment, on this question: skeptics.stackexchange.com/questions/22208/… The old answer is "no". I will answer "yes" and see how easy/difficult it is to bring out the truth.
    – DavePhD
    Apr 7 '16 at 12:19
  • 1
    @dave another example: skeptics.stackexchange.com/questions/32622/…
    – user30557
    Apr 29 '16 at 15:52

You must log in to answer this question.