Maybe Internet Polls Aren't So Bad After All
from the small-sample-sizes dept
Many people seem to assume that any internet-based poll is, inherently, unreliable as compared to other polling methods. However, Nate Silver has taken a look at how a variety of polls fared in the 2012 Presidential election and found that many of the internet polls did quite well, outperforming other methods:
…some of the most accurate firms were those that conducted their polls online.
The final poll conducted by Google Consumer Surveys had Mr. Obama ahead in the national popular vote by 2.3 percentage points – very close to his actual margin, which was 2.6 percentage points based on ballots counted through Saturday morning.
Ipsos, which conducted online polls for Reuters, came close to the actual results in most places that it surveyed, as did the Canadian online polling firm Angus Reid. Another online polling firm, YouGov, got reasonably good results.
This isn’t to say (of course) that online polling is always accurate. It still very much depends on methodology (some online polls didn’t do very well at all). But, it should put to rest the idea that online polling is inherently flawed or inaccurate.
Filed Under: elections, internet, nate silver, phone calls, polling
Comments on “Maybe Internet Polls Aren't So Bad After All”
Hmm
Does that really mean that the polls are accurate, or does it just mean that some polls randomly happened to be correct this time? A stopped clock is right twice a day, but you can only know if any given stopped clock is correct in retrospect, after you know the actual result.
If it manages to target audiences in a proportional way, sure. But I would never trust a poll say from the Fox News website. Obviously that poll would be heavily biased because of the type of people that visit that site.
Re: Re:
That’s kind of a microcosm of my issue with polls entirely.
You’ve got a lot of factors with any site…audience, trolls, how likely it is to attract viral trolls from one camp or another….
but on a larger scale, i figure most people filling out polls online, (in general, irrespective of site specific slant)are likely netizens. Those of us who may not watch the same newsfeeds, but share the basic commonality of being plugged in, enough that this is the platform we use to share our thoughts and opinions.
I know a few things about my fellow netizens, going by the polls. We like weed. doesn’t matter who puts the poll up, Fox, huffpo, or costco….we will find it, and we will upvote the ganj.
We’re mostly ok with gays. even on /b/ “fag” has pretty much no emotional load to it. It’s hard to work up much hatred of m4m, when you just saw a horse skullfucking a monkey corpse while a midget has sex with a chicken on it’s back.
We like cats. We really, REALLY like cats.
And there’s a lot of crazy, and even more stupid, but those people tend to rather rant than fill out polls.
I guess my point was that it’s good to see that things like national Hate Chicken day wasn’t actually representative of a majority of the country anymore. Gays and weed gaining some decisive victories over FUD and bullshit was nice. if the internet polls mmatched how it came out, then maybe things aren’t as bad as they often look.
Re: Re: Re:
I apologize, this wandered all over the place. I had a point..and i had a couple good lines, and then i realized i had a couple more points…and by then i had to go refresh my idea of the first point. it’s just not my best work.
lol
If you take an online poll about whether or not you believe that online polling is accurate, have you reached the end of the universe?
I was wondering about this
During the election cycle, I declined being polled over the phone at least a half dozen times, but I did answer two online polls. For my personal case, the online polls certainly better reflected my opinion as the phone polls did not have access to my opinion at all.
I was wondering how many people do as I did? More than I suspected, perhaps.
Better Questions
I think it’s more important to ask better questions as opposed to differentiating based on the communication method.
Audience targeting
If online polls are gaining acceptance, in short time geotargetable mobile poll systems (see example) will become the norm. We increasingly all carry a smartphone these days…