Site Meter

Saturday, May 18, 2013

Biased Assimilation


Brad Plumer demonstrates the power of biased assimilation.  A definition "people with opposing attitudes often assimilate evidence in a way that is biased towards their existing attitudinal position"

Plumer supports this hypothesis with the case of --  Brad Plumer. He wrote "If anything, it just makes people even more polarized." and (a paragraph later) linked to an abstract which includes " we observed no evidence of attitude polarisation".

OK yes there is the super weasel phrase "If anything," but I think he demonstrated the power of biased assimilation by finding the (second) linked paper and concluding that no evidence of polarization can be summarized by "if anything ... polarized".

What is happening ? I think one thing is that psychologists are people too.  The abstract clearly presents the hypothesis before noting the absence of evidence "which may lead to attitude polarisation." What we have here is a prior which motivated the study and which is reported in the abstract.  I'd guess that there is evidence somewhere that biased assimilation causes polarisation and evidence somewhere else that information reduces polarisation in spite of baised assimiliation and biased assimilation of the evidence.  This is based on my assumption that the psychology of psychologists is similar to that of economists.

I think another thing is that Ballance has infected wonkblog.  The abstract provides evidence that climate change non skeptics are open minded and skeptics are closed minded

we measured participants’ scepticism about climate change before and after reading two newspaper editorials that made opposing claims about the reality and seriousness of climate change [skip] In both groups, attitudes towards climate change became significantly more sceptical after reading the editorials, but we observed no evidence of attitude polarisation

The result is as asymmetrical as a result can be.  But it is not OK for wonks to say that the social science evidence suggests that conservatives respond to arguments by become more extreme while non conservatives respond rationally.  The non-skeptics started from one opinion that anthropogenic climate is occuring and ending with 2 that it is and one that it isn't moved towards skepticism.  This is open minded, humble and even rational.   But the post (which is a discussion of how best to convince the public of the objective fact that we are causing climate change) slips into Ballance asserting that evidence that debate makes conservatives more extreme demonstrates "polarisation."

1 comment:

Brad Plumer said...

oh argh. this study was the one i meant to link to there: http://www.nature.com/nclimate/journal/v2/n10/full/nclimate1547.html

which does find more polarization among people with greater scientific awareness, which i think speaks to kahan's point.

but your point about the other study is a good one.