Academics are hugely Left-leaning.  Is it because they have higher IQs?

One explanation that has been put forward to explain the overrepresentation of individuals with left-wing and liberal views in academia is that they tend to have higher intelligence. The theory is that academic advancement requires very high intelligence, and since few individuals with right-wing and conservative views possess very high intelligence, such individuals are comparatively scarce within the academy (Solon 2014; Solon 2015; Charlton 2009; Gross 2013).

Several recent studies from the US, where the academy also has a sizable left-liberal skew, have concluded that intelligence does not contribute much to explaining the tilt (Gross & Fosse 2012; Gross 2013; Fosse et al. 2014). On the other hand, using a slightly different method, Carl (2015b) found that intelligence may account for more than fferent method, Carl (2015b) found that intelligence may account for more than half of the overrepresentation of socially liberal views, but may not account for any of the overrepresentation of economically left-wing views. His finding is consistent with evidence that cognitive ability is positively related to both socially liberal beliefs and at least some measures of economically right-wing beliefs (Carl 2015a).

Unfortunately, there do not appear to have been any surveys of British academics asking about specific policy issues, either economic (e.g., nationalisation of industry) or social (e.g., immigration). Only the distribution of party support among academics is available, which as noted above points to an overrepresentation of both left-wing views and liberal views.

To see whether intelligence contributes to explaining the left-liberal skew of party support among academics, I calculated the distribution of party support for individuals within the top 5% of IQ, using data from the Understanding Society survey. This is shown in Table 3, along with the distribution of party affiliation within the general population and among academics, also calculated from the Understanding Society data.



Note that the distribution within the general population differs from the outcome of the general election; this is probably due to the phrasing of the question posed in Understanding Society, to the sample being slightly unrepresentative, to the timing of the data collection, and to differential turnout by party.

However, what is of primary interest is the comparison between the figures for the general population and those for the top 5% of IQ, which were both calculated from the same data.

Conservative supporters are about as well represented within the top 5% of IQ as they are within the general population, Labour supporters are slightly underrepresented, UKIP supporters are underrepresented, Lib Dem supporters are overrepresented, and Green supporters are overrepresented. Overall, as Figure 2 illustrates, the distribution of left/right orientation within the top 5% of IQ is relatively similar to the distribution within the general population.

While intelligence may account for some of the underrepresentation of UKIP supporters among academics, and some of the overrepresentation of Green supporters (Deary et al. 2008), it cannot account for the substantial underrepresentation of Conservative supporters. To the extent that the Conservatives are a less socially conservative party than UKIP, the figures in Table 3 are consistent with Carl’s (2015b) finding that intelligence may contribute to explaining the underrepresentation of socially conservative views in American academia, but not necessarily the underrepresentation of economically right-wing views.

Somewhat surprising is the relative scarcity of Lib Dem supporters among academics, given their overrepresentation within the top 5% of IQ. This may be attributable to the fact that, as noted above, the Lib Dem party was until recently dominated by its classically liberal wing, which espoused comparatively more right-wing policies, which may not have been appealing to academics. On the other hand, it may simply be due to sampling error.

SOURCE