User:Rspeer/Pollzilla, Destroyer of Consensus

So you're having a dispute on Wikipedia. The discussion is going nowhere fast. Both sides argue that the other one doesn't have consensus, and it all comes downto the simple fact that some people believe one thing while some others believe its opposite, and they don't want to change their minds.

Then someone gets a bright idea: "Why don't we have a poll? Then one side wins, the other side loses, and the issue is over!" At that point, you'd better run for cover, because whatever small consensus you had ever found in that discussion is about to smashed into tiny pieces by Pollzilla, Destroyer of Consensus.

You don't want to summon Pollzilla. It's easy to avoid doing so -- just keep in mind that polls do not create consensus, they destroy it.

How do polls destroy consensus?

• Polls make it undesirable to find a compromise. People who hold polarized opinions are generally going to support those polarized opinions, so that the opposite polarized opinion doesn't end up with more people.
• There is no perfect procedure for turning poll results into a consensus decision. Except in places with a well-established tradition of frequent polls, like WP:RFA, your poll will turn an argument over the issues into an argument over what the poll results mean. If 66% of the people agree on something, does that mean the other 34% should be left out of the discussion from then on? What if it's 80% versus 20%? What if the 20% are giving well-thought-out justifications while the 80% are just signing their names? (The correct answer to all of those questions is "Does it matter? When did we go from writing an encyclopedia to arguing over percentages?")
• The result will probably be even less clear-cut than that. When people do look for a position besides the two polarized options, they'll create more options to vote on, and more options makes it difficult to interpret the results. There are clear reasons why this is the case. If people are just supporting their favorite options, how can you find a compromise that will satisfy a majority of participants? If people are listing their first and second choices, how do you aggregate these? If you agreed on a voting system beforehand (very rare), how do you ensure that people aren't manipulating it? The consensus you get out of such a procedure will be shaky at best.
• Even though it frequently turns out not to be the case, polls give each side the impression that they can score a complete victory over the other side. Then, the participants aren't even hoping to find a compromise anymore, they're hoping that enough "right people" show up to make the "wrong people" go away.
• Who are all these people, anyway? If a lot of people show up to the poll, why weren't they participating in the discussion before? Do they actually care about the issue that much? Do they understand the issue, or are they just arguing for overly simplified positions? On the other hand, if only a few people show up, is the result really legitimate?

A typical poll will just distract from the process of finding a consensus. At worst, it will polarize the issue and destroy whatever consensus was there previously.

What's the alternative?

You're on a Wiki, not a Nomic, so you'll get more done by editing than by voting. The bold, revert, discuss cycle is perhaps the best tool we know of for building a consensus on a Wiki. Yes, you have to be willing to talk to people you disagree with, and to get reverted by them sometimes, but the result is often a far stronger consensus than what would come out of a poll.