Background A considerable literature of ethnic density studies links mental health to ethnic minority status in one’s neighbourhood. This design has also been applied to other types of identity, such as sexual minorities and Brexit identities, but surprisingly little attention has been paid to class-based identities. Bourdieu decomposes class into several distinct, but mutually reinforcing, forms of capital. Here, we examine whether a ‘class density’ association with mental health can be found for economic capital – people’s material assets – and cultural capital – symbolic signifiers of class such as tastes and interests.

Methods Multi-level regression with post-stratification was used to make area-level estimates of cultural and economic capital for each middle super output area in Wales using data from the National Survey for Wales 2017/18 (N=11381). Mixed effects models were fitted to economic capital, operationalised using respondents’ income, house ownership, and material deprivation, and cultural capital, operationalised using items on attendance at a variety of artistic, cultural, and heritage activities. Terms for demographic and area-level variables, as well as random effects of middle super output area were estimated and post-stratified using census to create estimates of area-level economic and cultural capital for all Welsh middle super output areas. These estimates were linked to independent individual-level data from the 2018/19 edition of the survey (N=4058). Mixed effects models containing individual-level capital, area-level capital, and their interaction were fitted, predicting whether respondents reported a mental health problem. Models were fitted unadjusted, adjusted for age and gender, adjusted for the other form of capital on the individual and area levels, and adjusted for all these and the other capital’s cross-level interaction. Multiple imputation was used to account for missing data.

Results For cultural capital, a cross-level interaction was found where area-level cultural capital was protective in respondents reporting higher levels of individual-level cultural capital, but a risk factor for mental health problems in those with low individual-level cultural capital (odds ratio=.83, CI_95%=.74-.93). No such relationship was found for economic capital (odds ratio=.96, CI_95%=.88–1.06). These results remained robust in the adjusted models.

Conclusion The presence of a class density association with mental health for cultural capital but not economic capital suggests that class-based identities can also lead to density associations with mental health, but that it is the symbolic signifiers of class, and not material aspects, which activate class identities in this way.