Within-Neighbourhood Changes of Collective Efficacy—Longitudinal Measurement Invariance and Association with Changing Ethnic Diversity

Abstract In neighbourhood research, the concept of collective efficacy has been particularly successful in capturing key dimensions of the social fabric of communities, i.e. social cohesion and expectations for social control actions amongst residents. Yet, very few scholars have studied collective efficacy over time, looking at the social dynamics of neighbourhood development, and none has tested for longitudinal measurement invariance. We use a repeated cross-sectional survey of residents conducted in 2014 and 2020 in 139 neighbourhoods in Cologne and Essen (Germany) to analyse the measurement invariance of the scale collective efficacy. Applying multilevel CFA, previous analyses have shown that this scale has two separate latent dimensions on the individual level of respondents – social cohesion and informal social control – but just one latent dimension on the collective level of neighbourhoods (Gerstner et al., Social Indicators Research 144:1151–1177, 2019). We test for longitudinal measurement invariance of collective efficacy in the framework of multilevel CFA, and in a second step, explore the covarying changes of collective efficacy and sociodemographic indicators. We find scalar measurement invariance and only small changes of collective efficacy over time which, however, are significantly related to changes in ethnic diversity: increases in ethnic diversity are associated with slight decreases in collective efficacy, supporting both classic social disorganization theory as well as Robert Putnam’s ‘hunkering down’ hypothesis.