Тоня Самсонова
ноябрь 2016.

Is there a link between someone's levels of education and their propensity for liberal values?

1 ответ

This link is often brought up – most recently in relation to the analyses that show that, in the 2016 US presidential elections, the most consistent difference in voting preferences was between those educated at college (tertiary) level, who predominantly voted Clinton, and those who voted Trump – the majority of whom did not have college education. This is also related to the longer-standing, but now particularly relevant, discussions about the purported 'liberal bias' in American higher education. So let’s look at the evidence.

When trying to understand how education and values are connected, it is very important not to mistake correlation with causation. Seeing that two effects occur at the same time does not mean that one is necessarily causing the other. Statistical analyses that show that higher proportion of voters for Clinton were highly educated do not suggest it was their education that decided how they would vote: after all, a large portion (up to 43%) of college graduates did back Trump. In a related fashion, it makes sense to recall that the majority of Tea Party supporters were highly educated. Therefore, it would be wrong to assume that levels of education automatically translate into specific political preferences. In order to understand the link, we need to look deeper: to the level of causal mechanisms.

Social scientists have different understandings of how causality operates. For instance, do people adopt liberal views while they’re in education? Or does having liberal views make them more likely to end up at university in the first place? Historical materialists assume that it’s one’s relationship to the means of production – in other words, their class position – that influences their values. To the degree to which the dominant class is liberal, then, education institutions will both reflect and reproduce these values. Rational choice theorists, by contrast, assume people act in ways that maximise their personal gain in a given situation; however, their assessment of the ‘situation’ is also influenced by their position in social structure. This would mean that those coming from more privileged social backgrounds are more likely to pursue and value higher levels of education.

Most interpretations, of course, fall somewhere between these poles. Rather than aiming to assess whether education determines values (or vice versa), we can try to understand how they are related. Institutional environment, in this regard, certainly plays a role. To the degree to which the values such as pursuit of knowledge, openness to debate, and academic freedom are liberal, we could say that universities reflect this ethos; however, there is very little evidence of a systemic liberal 'bias' in any specific system of higher education. Furthermore, it is important to bear in mind that political preferences are influenced by a host of other factors – family, occupation, location, peers and friends, to name but a few – and, while some of these correlate with education, not all do.

One aspect in which higher education does seem to matter are attitudes towards social justice: namely, recent research across a number of political contexts suggests the highly educated are more in favour of redistributive social policies. When these policies are taken up by liberal political parties (as, for instance, was the case under Obama’s administration in the US), we could reasonably assume the highly educated are likely to back them. However, this has more to do with generational effects and class stratification, than with education alone. In sum, though education certainly matters, understanding how different aspects come together to form stable preferences is an ongoing and fascinating endeavour.


It looks like the paper about preferences for redistribution makes a different point than what's represented here. It's about the type of education received, not its quantity. To wit, 

"the main empirical finding of this article is that individuals in countries with high levels of private spending on education are less willing to support government-induced redistribution."

This makes sense in context of social program stickiness. These programs are supported by people who benefitted from them or expect to benefit from them at some point. Government-subsidized education is a nice way to get "elites" and their kids to endorse other kinds of redistribution.

But as you note, disentangling causality is really hard and really important. One paper that does it is here.

It finds that the causal effect of education goes the OTHER direction and makes people less supportive of redistributionary policies: 

"Utilizing variation in compulsory schooling laws (CSLs) in the US and Britain and multiple identification strategies, I find raising CSLs by a year causes a 2-5 percentage point swing toward the Republican or Conservative party per cohort." 

The interpretation is that better educated earn more and are more likely to be net contributors than net beneficiaries to the safety net.

In the case of redistribution, an earlier paper that makes the redistribution point clearer is here: isites.harvard.edu ; it finds that "younger, better educated and more left-leaning individuals support increases in public spending on education"; Neil Gross makes a broader point in "Why are the highly educated so liberal?"
In both cases, though, I would stress it is about correlation, not causation. Which does serve to underline the point that it is difficult to draw conclusions about the link between education and any specific political preference.