What do people think about polling the membership?

June 29, 2004 7:42 PM

What do people think about polling the membership? A feature that would allow Matt (or others) to provide questions on popular issues/ideas/songs or whatever and get a sense of what the community thinks. The obvious example is topical stuff, e.g. who will the Metafilter users vote for in the election? This idea could either enhance the community, or divide it. The biggest negative is that statistics are misleading, they can polarize things, and they give people easy examples when they are generalizing, and that sucks. But, it would be fascinating to know how the community stands on a given issue. I still think this is the best community on the internet and the most interesting, and I think many people would like to know how we feel on this or that. If it were to happen in the style of this site, it would be a once every few months kind of thing. A side feature could be to have all discussion on that issue reside on that page (get them off the front).
posted by cell divide to feature requests at 7:42 PM (29 comments total)

So 1998.
posted by Stan Chin at 7:44 PM on June 29, 2004

it would be fascinating to know how the community stands on a given issue.

True, but you can get that with discussion, too. I fear that running the hard numbers would embolden the users who find themselves in the majority. Everyone seems to have decided that MetaFilter is mostly male, for example, and the "yay, it's a boyzone" foolishness gets annoying from time to time. That's a different example than the politics business, but other than mere curiosity, why do you really want to know that?
posted by scarabic at 7:47 PM on June 29, 2004

pollfilter.metafilter.com.....I think it could be done off the main site, as so many other things are? I know I wouldn't be doing them, or very interested in the results, in all honesty.

Of course, if you go through Miguel's I love you please don't be hurt posting history on the site, I think you'll find a number of those types of threads already exist. Maybe your question has already been answered!
posted by Salmonberry at 7:53 PM on June 29, 2004

MeFi Survey
posted by PrinceValium at 8:19 PM on June 29, 2004

I vote 'no' on polls.
posted by Witty at 8:54 PM on June 29, 2004

Bad idea. lets just do another survey because I wasn't in it.
posted by Keyser Soze at 8:59 PM on June 29, 2004

Sounds like a great idea! We can even have a CowboyNeal option on every poll. That'll be so funny!
posted by reklaw at 9:07 PM on June 29, 2004

"It says here that 75% of people think polls don't mean a damn thing!"

"Well, gee, if that many people think so, it must be true!"

[from Tom Tomorrow]
posted by scarabic at 9:07 PM on June 29, 2004

• No, this is not a good idea. A self-selected group within a non-representative sample that would mean...what?
• If Matt does decide to do this, the first question should be You inherit $5M the same day aliens land on the earth and say they're gonna blow it up in 2 days. What do you do?
• what reklaw said

posted by Stoatfarm at 9:19 PM on June 29, 2004

Count me as another vote against.

Aside from the fact that the internet is lousy with informal polling, it would seem to be against the stated purpose of metafilter ("best of the web" and all that) and would mostly encourage even more newsfilter and opinion posts.
posted by malphigian at 9:28 PM on June 29, 2004

Stop voting on this! It's what they want you to do!
posted by Krrrlson at 9:38 PM on June 29, 2004

There is no need for polls. Matt has been secretly collecting information on all of us all along. The MetaFilter Network knows everything about you. If you've been to a MeFi Meetup he even has your DNA. Right now he is using the information for good, but everyone has a price.
posted by birdherder at 9:51 PM on June 29, 2004

" The biggest negative is that statistics are misleading"

Considering that out of 17K+ Mefi users there are probably fewer than 100 who actually put "male" or "female" on the entry for gender in their profiles, I'd say that's a fair statement.
posted by mr_crash_davis at 10:17 PM on June 29, 2004

A self-selected group within a non-representative sample that would mean...what?

You are right that it would be self-selected (which is fairly common in polls), but it would be representative (if you had a certain number of users participate) to the MeFi population. So any findings would only be generalizable for MeFi users.
posted by Quartermass at 10:18 PM on June 29, 2004

Well, regardless of where we go with polls and whatnot, I'm the cutest.
posted by dobbs at 10:30 PM on June 29, 2004

pfft! *i'm* the most likely to succeed!
posted by quonsar at 10:39 PM on June 29, 2004

at what?
posted by dg at 10:48 PM on June 29, 2004

I'm a dark horse.
posted by taz at 10:55 PM on June 29, 2004

Kick a pic so we can vote on that, dobbs.
posted by scarabic at 11:08 PM on June 29, 2004

If my girlfriend's pimp would give me back my glass eye, I bet I could give dobbs a run.
posted by Opus Dark at 11:55 PM on June 29, 2004

Just ask Miguel
posted by matteo at 12:22 AM on June 30, 2004

Is there anyone who feels an actual lack of opportunities to vote in meaningless polls on the interweb? Because there seem to me to be more than ample opportunities of same.
posted by Sidhedevil at 6:04 AM on June 30, 2004

I'd take that wad and slide it right on over to my dad, because he's one of the top investment bankers in the state.
posted by vraxoin at 6:53 AM on June 30, 2004

I still think this is the best community on the internet and the most interesting, and I think many people would like to know how we feel on this or that.

And there the poster, inadvertently one must assume, captures in a nutshell the inherant communicative flaw of polling data.

What was once a community - characterized by intelligent discussion and [ahem] civil debate - is instantly transformed into a single, unified block. Polling that reports majorities give the false impression that the entire community supports that majority, or is at least complicit with it.

Once we say things like "85% of MeFites like sushi" it ends the discussion and labels all of MeFi as a fish-eating community. As I said, this gives the false impression that:

A) sushi consupmption is the defining characteristic of MeFi users (especially when alongside such stats as 50% of MeFi users wear sandals. The bigger the majority, the better that number stays with us as.)
B) nearly all MeFites are sushi eaters. While this is in fact true, through successive storings and recalls this quicly gets shortened cognitively to the much simpler heuristic: ALL MeFites like sushi.

and 75% of statistics are just made up on the spot, anyway.

There is a common statistics joke that goes: do more truck drivers or college professors read Shakespeare? The answer is, of course, more truck drivers do, simply because there are so many more of them. A small piece of a big pie is often bigger than a big piece of a small pie.

Even if only 10% of 17,000 believe something, especially if it is a vocal 10%, that's still alot of posting power.
posted by ChasFile at 6:59 AM on June 30, 2004

Well, cell devide, let me be apparently the first to support this idea. As a lark, it might be somewhat fun and interesting. Obviously, the sample size is too small, the chi square test won't work, the bell curve is skewed, statistics are malleable, statistics are the tools of The Man, blah blah blah. But we already essentially have this in the form of AskMe posts like "Is cheating wrong?" that appear to be popular, so I don't see why this would be any different.

/pissing into wind
posted by pardonyou? at 8:40 AM on June 30, 2004

Membership polls are:
[ ] Fun.
[ ] Divisive.
[ ] The prerogative of certain members.
[ ] In the bath.
[ ] About three times a week.
[ ] Broccoli.
[ ] All or none of the above.
posted by carter at 9:28 AM on June 30, 2004

[X] All or none of the above.
posted by caddis at 1:59 PM on June 30, 2004

[X] Pissing into wind.
posted by wendell at 5:06 PM on June 30, 2004

What election?
posted by dayvin at 2:39 AM on July 1, 2004