Sunday, April 06, 2008

Framing Atheism

 
Here's a video that presents statistics about atheism. Some of it makes me uncomfortable because the numbers can be very misleading. Is this an example of framing according to the ideas of Nisbet and Mooney? If so, to my mind it illustrates all of the bad things about framing that turn scientist off.




[Hat Tip: Friendly Atheist]

3 comments:

  1. I take your complaint to be that the film ignores "Atheism is true" in favour of "Atheists are smart, law-abiding, and have stable lives"? Fair enough, but in a climate of anti-atheist prejudice, there's also a place for saying, loud and clear, that the standard charges against atheism are lies. And there's nothing wrong with choosing, in a particular instance, to concentrate on one aspect and not the other (as long as, somewhere along the line, the other part gets discussed as well).

    Be that as it may, I think there are a few criticisms to be made about some of the stats cited:
    - "not affiliated" is not the same as "atheist". Lots of lapsed church-goers still retain some sort of personal theism.
    - Surely they can find a better source than Holysmoke?
    - the divorce stats could be questioned on the grounds that atheists may be more likely to cohabit outside of marriage. Hence when the relationship breaks up, it does not show in the divorce stats. (I'm treating the family breakdown as the "real" problem here, irrespective of legal status)

    ReplyDelete
  2. I didn't much like the video, either.

    I agree with Eamon's comment that demonstrating the false nature of charges against atheism is useful and important. But I don't think this particular video does a very good job of it.

    The video doesn't present any statistics, just a collection of numbers with no analysis. The viewer is dragged to the conclusion that atheists are somehow better behaved than are a disjointed collection of various religious groups. But, I saw not one example of an actual statistical analysis. No p-values, no permutation tests, no mention of variances or measurement error at all.

    For example, near the end of the video we see a description of divorce rates. Leaving aside the confounding factors as applied to atheists as mentioned by Eamon, the rest of this little data table is problematic: are Jews really more divorce-prone than most Christians? Without some estimate of the variance of those estimates (30% vs. 27% and so on), we have no way of knowing if those differences are significant.

    A slighly earlier segment in the video describes an apparent meta-analysis. But a meta-analysis is more than "X% of reviewed studies broadly agreed with conclusion Y", it must also include some examination of WHY very disparate studies would come to similar conclusions, and why the exceptions, those that found the opposite or a neutral conclusion, were exceptional.

    Overall, I was pretty disappointed. This collection of numbers would have worked much better as an essay, moving through themes like "national characterists" and "religion and family stability" or something. And, I didn't think the music or background images worked well.

    ReplyDelete
  3. The level of dishonesty or just plain ignorance of the person who created this video is abysmal.
    I strongly suggest he take a peak at http://atheismexposed.tripod.com.
    Mico

    ReplyDelete