The link between educational attainment and socio-political values is well-established (1). In Britain today, university graduates, on average, have considerably more liberal cultural attitudes and somewhat less liberal economic attitudes than non-graduates (2). But does studying at university really cause graduates to develop distinctive attitudinal profiles?
This is a question that has preoccupied social scientists for decades and is notoriously difficult to answer, because the relationship between educational attainment and socio-political attitudes is extremely complex. While it is certainly plausible that graduates might develop distinctively (il)liberal attitudes as a direct effect of their experiences of studying at university, it may simply seem this way because graduates, or those who will later go on to become graduates, are disproportionately exposed to other kinds of experiences which stimulate the development of these attitudes. One obvious example of such confounding is the fact that British graduates typically earn considerably more than non-graduates over their life courses (3), and that income is an important determinant of attitudes (4).
Only by accounting for all alternative pre- and post-university explanations of individuals’ socio-political values (illustrated by the black lines in Figure 1) in statistical analysis can we identify the independent causal effect of university study in determining these attitudes (see the orange line in Figure 1). This is problematic because most existing data do not contain sufficient information on these kinds of confounding variables.
My recent study, “Demystifying the link between higher education and liberal values: A within-sibship analysis of British individuals’ attitudes from 1994–2020”(5), published in the British Journal of Sociology, leverages the unique household structure of the British Household Panel study and Understanding Society data (6) to better identify the causal effect of university study on British individuals’ attitudes, in the period 1994-2020. It does so by estimating this effect only within groups of siblings who grew up in the same household, so that all unmeasured family-invariant pre-adult experiences can be controlled, in addition to the measured sources of confounding which have been accounted for in most existing studies of this association.
I estimate the effects of university study on British individuals’ cultural and economic attitudes – using gender egalitarianism and environmentalism as indicators of the former and left-right orientation to measure the latter.
My study shows that, initially, university study has a strong effect on British individuals’ attitudes. Before controlling for any confounding variables, those who have studied at university were indeed considerably more culturally liberal and more economically conservative than their less educated counterparts. Clearly then, there is a strong educational divide in attitudes in Britain today. Yet, my study also shows that once we control for all measured pre- and post-university confounders of the education-attitudes association, as existing studies have done, estimates of this university effect shrink substantially in size. After doing so, the effect of university study on economic attitudes ceases to be statistically significant, and while estimates of university study’s effect on cultural attitudes are reduced by at least 60%, they remain fairly large and highly statistically significant. It still appears that graduates are considerably more culturally liberal than non-graduates.
However, once I use my within-sibling design to control for additional unmeasured confounding, I find these estimates of university study’s effect on British individuals’ cultural attitudes not only get even smaller but also cease to be statistically significant. Under this more robust statistical test, the estimated effect of university study on British individuals’ cultural attitudes was so small that I found there to be a considerable chance that this effect could indeed be zero (once we account for the statistical uncertainty around estimation). It therefore seemed most accurate to conclude, on the basis of my findings, that university study is only likely to have a very modest direct causal effect on British individuals’ cultural attitudes, and that this effect is not always liberalising.
My study provides evidence to suggest, contrary to the claims of many right-leaning commentators (7), that studying at university actually does little to make graduates more liberal and can also have the opposite effect. Rather, the association of educational attainment with attitudes is largely spurious. It materialises predominantly because liberal young people disproportionately choose to enrol at universities in Britain today.
Elizabeth Simon was formerly a PhD candidate in the Social Statistics and Demography department at the University of Southampton. She is currently a Postdoctoral Researcher in British Politics within the Mile End Institute at Queen Mary, University of London. Her research focuses on how educational attainment shapes public opinion, electoral behaviour and wider society.