Why internet polls are (possibly) important

It is oftentimes the case that a poll appears on the Internet.  Something like “Is the Earth flat?”, or “Should the government fund a state astrologer?” or whatever.

What happens next?  Someone skeptical notices, and puts a link to the poll on twitter or Facebook.  Other skeptics forward the link (it usually gets back to PZ at some stage) and we all go and vote to make sure the poll ends with the right answer*.  At the same time, other skeptics pipe up (not unreasonably) saying, well, it’s only an Internet poll.  The truth is not a democracy, it’s not really important, who cares if someone on the Internet is wrong, etc.

And the poll closes.  Sometimes with the right answer, sometimes not.

Today, giving evidence at the Leveson Inquiry, Rebekah Brooks said something interesting. Well, she said a few things interesting… but for the purposes of this blog post, the interesting part was (from memory):

We had a pretty good idea of what our readers’ views were. We had done a lot of polls.

Why is that interesting?

Well, because it seems to suggest that newspapers, at least on some issues, use the result of polls to gauge the views of their readers.  It seems to suggest that the outcome of these polls may influence the decision of the media to champion one side of the debate or the other.  It seems to suggest that if The Daily Ernest is persuaded by a poll that their readers are largely of the view that the Moon is made of cheese, they may not report so favourable on that unfashionable Rock Theory of the Moon in future.

Of course, I’m making some unfounded assumptions here. I’m assuming that the polls Ms Brooks mentioned included Internet polls. And I’m assuming that newspapers will sometimes adjust editorial policy to pander to the views of their established readership.

But it is something to think about.

* By “right” answer, I mean the one which best reflects reality based on the current data.

  1. #1 by Kevin Wakley on May 11, 2012 - 13:19

    Mike, are you maybe getting confused here between proper polling, performed by polling agencies, and internet polls performed by ……. ?

  2. #2 by Mike on May 11, 2012 - 13:23


    For one thing, I point out in the article that I’m making the unfounded assumption that her comment includes Internet polls. Maybe it doesn’t, and I’m full of shit.

    For another thing, in my professional experience, polls companies conduct on their own website are paid attention to. They aren’t polling readers for a laugh; the results may be statistically meaningless, but it isn’t uncommon for the company in question to take it as being indicative of their readers opinions.

  3. #3 by Kevin Wakley on May 11, 2012 - 13:25

    If that’s really the case then there’s ever more reason to Pharyngulate then!

  4. #4 by Clinton Huxley on May 22, 2012 - 16:28

    One would expect the opinions of ones readership to fall pretty much in line with ones editorial policy. Most people read newspapers that confirm their biases. It’s almost a chicken and egg question.

  5. #5 by AntRya.net on May 29, 2012 - 20:40

    Polls can be of great interest. As the article says though, they are pointless if pandering to the readership. I’m including one in my book, because I want as many views as possible :o)

(will not be published)