As a follow up to “Lee’s post on Napier and Jost”:https://themonkeycage.org/2008/07/post_97.html#more, this graph, from Larry Bartels’ _Unequal Democracy,_ is pretty striking.
It shows how different levels of political information affect conservatives’ and liberals’ beliefs over whether income inequality between rich people and poor people has increased. The facts are unambiguous – economic inequality between rich and poor has increased very substantially. Yet, as Bartels describes it, increased political awareness (as measured by correct answers to the usual kinds of questions about which party had more members in the House, which party was more conservative etc) has different effects on conservatives and liberals. In Bartels’ words (p.155):
bq. At low levels of political information, the figure shows that conservatives and liberals were about equally likely to recognize that income differences had increased over the past 20 years. However, the perceptions of better-informed conservatives and liberals diverged significantly. Among liberals, recognition of increasing income inequality rose markedy with general political awareness, to 86% for people of average political awareness … and a near-unanimous 96% at the highest information level. However, the proportion of extreme conservatives who were willing to admit that economic inequality had increased actually decreased with political information, from 80% among those who were generally least informed about politics to 70% for people of average political awareness to a little less than 60% among those at the top of the distribution of political information.
As Bartels goes on to note (p.157)
bq. Rather than contributing to accurate apprehension of [a seemingly straightforward objective fact] by conservative and liberal observers, political awareness seems mostly to have taught people how the political elites who share their ideological commitments would _like_ them to see the world. In particular, what is most significant in figure 5.2 is the effect of political information among conservatives. Rather than being more likely to recognize the reality of growing inequality, those conservatives who were most politically aware were most likely to _deny_ that income differences had increased. In this instance, political awareness did more to facilitate ideological consistency than it did to promote an accurate perception of real social conditions.
Update: Thanks to Andrew in comments for linking to “this post”:http://www.stat.columbia.edu/~cook/movabletype/archives/2008/07/education_and_t.html describing a similar phenomenon among Democrats.