Is There a Causal Relationship Between Occupations and Ideology? The Case of the British Third Sector

In recent decades, quantitative researchers have debated the relationship between employment and ideology. Some of this work develops typologies of occupations with distinct economic and cultural values (e.g. Kitschelt and Rehm, 2014), with occupations with technical and interpersonal task structures tending to be more economically left-wing and culturally liberal. Subsequently, researchers have elaborated the occupational foundations of left-liberal values with different samples (e.g. McAndrew et al., 2020). However, such work uses cross-sectional samples and, therefore, is unable to make causal inferences. In a review of the wider literature on employment and ideology, Selenko et al. (2025) conclude that 'One of the most serious limitations of our evidence base was the cross-sectional nature of most studies... Even of the longitudinal studies available, none tested for causality direction.' We address this gap. Using the British Election Study (BES), a longstanding panel study of the British electorate with items on ideology and employment, we examine the causal relationship between ideology and employment among third sector workers. On BES, this industry has sufficiently granular coding and, moreover, is theoretically interesting. Organizations in the third sector tend to have strong occupational missions – often these reflect left-liberal values – and BES data show that third sector workers tend to have more left-wing economic values and liberal cultural values than workers in other sectors. We use five measures of values: an economic left-right scale, a liberalism-authoritarianism scale, a left-right self-placement item, a redistribution self-placement item and an item which measures opinion of the Conservative Party. Examining pre-transition values with a logistic regression with person random effects, we discover a relationship between third sector transition and left-liberal values on all five measures, i.e. a selection effect. We investigate the influence of transition on values with the stacked difference-in-differences method (Wing et al, 2024). Addressing the problem of panel imbalance, this method splits the data into sub-experiments for specific ‘adoption years’, with control units adopting no earlier than the end of the event window. We find evidence of a causal relationship. With the exception of the liberalism-authoritarianism scale, transition is associated with increased left-liberal values on all of the measures. In conclusion, we discuss the extent to which one may generalize from our case and propose future studies. Notwithstanding challenges with existing datasets, there is scope for quantitative work. Qualitative researchers might also examine the mechanisms by which value change occurs among third sector workers. References Kitschelt, H., & Rehm, P. (2014). Occupations as a site of political preference formation. Comparative political studies, 47(12), 1670-1706. McAndrew, S., O’Brien, D., & Taylor, M. (2020). The values of culture? Social closure in the political identities, policy preferences, and social attitudes of cultural and creative workers. The Sociological Review, 68(1), 33-54. Selenko, E., Schilbach, M., Brieger, S. A., Van Hootegem, A., & De Witte, H. (2025). The political consequences of work: An integrative review. Journal of Management, 51(6), 2355-2388. Wing, C., Freedman, S. M., & Hollingsworth, A. (2024). Stacked difference-in-differences (No. w32054). National Bureau of Economic Research.