Saturday, July 09, 2005

Prigs who like 6th grade math   posted by Razib @ 7/09/2005 07:17:00 PM

Over the past few days I've been commenting a lot over at ParaPundit and Sepia Mutiny. I have been a bit of a pest, but I've really started to agitate for more "6th grade math"1 in public policy discussions. Over at Sepia some individuals seem to live in terror of hate crimes because of Islamist bombings. As someone who has spent their whole life in regions where the lowbound frequency of whites is 90%, and more often 99%, I really don't see a Rosewood happening anytime in the near future. In contrast, over at Randall's blog I've made myself a thorn in the side of some by demanding precise numbers for various assertions (I usually know the numbers beforehand and am aware that it would take 15-30 seconds to cough them up using google or wikipedia). In general I tend to agree with Randall's readers on many points...but I get irritateted by the tendency to couch observations impressionistically, and also engage in hyperbole for effect. As far as I'm concerned, this only undermines your case and adds distortion into models that might have utility for public policy. For a specific example, one reader asserted "and reliable projections point to an actual non-White majority by mid-century." Using the UK Census site I back-of-the-enveloped that the absolute number of whites in Great Britain would have to drop by 20 million individuals in 2 generations, assuming the relatively high rate of immigration and non-native total fertility rate remaining constant, for this contention to be defensible. Why is this relevant? After Randy posted his piece debunking the more extreme claims about the Islamicization of France many liberals were jubilant, as far as they were concerned there was no problem to worry about because Randy's analysis had thrown cold water on the more dire demographic "projections."2 I think such a response is wrong-headed and premature, but in the current climate of hyperbole and imprecise formulation of hypotheses, it is entirely all too easy.3

PS: Being a negative pest is a lot easier than putting forward a positive model. I'll put up in the near future....

1 - Steve has generated a large body of work that basically applies arithmetical principles to polling data. It's ludicrous that mainstream journalists can't remember the math they were supposed to learn in 6th grade for the life of them when their "sources" feed them garbage, but that's the way of the world.

2 - Projections are in quotes because they were usually verbal assertions. Sometimes a number would be thrown in for effect, "4,000 Frenchmen are converting to Islam per year," but totally without context and often with little relevance.

3 - Randall points out that a proportion of Muslims on the order of 10-20% of a electorate has consequences. I think the 10-20% is far more plausible than a Muslim majority. This number is something you can use when you look at other nations as "case studies" (India and Singapore are two nations in this range). Unfortunately, I do not see this sort of precision most of the time.