The economic crisis has hit some parts of the Spanish population harder than others. We assume that this inequality is amplified by social networks, in the sense that (a) the personal networks of some people are more affected than those of others, (b) the degree to which networks are affected covaries with individuals’ own socio-economic situation, and (c) how people react to the crisis and how they cope with it is not only determined by their own economic situation, but also by the situations of others in their broader social environment.
With the aim to test these assumptions, we designed a special module of the National Barometer in 2014-’15 in collaboration with the Centro de Investigaciones Sociológicas (CIS). Following among others DiPrete et al. (2011) and McCormick et al. (2010), we included the Network Scale Up Method to estimate the size of acquaintanceship networks. Also, we included questions about the number of people respondents knew in various subpopulations (e.g., people who had lost their jobs, who had been evicted or who had migrated). Other questions referred to respondents’ socio-economic attributes (assumption b), and their trust in public institutions, interpersonal trust, and the availability of social support (c). Interviews were performed by the CIS using CAPI with a nationally representative sample of approximately 2,500 inhabitants of Spain. We used the R package NSUM for the estimations of total network size and overdispersion of population sizes in networks (assumption a), and regression analysis to relate the residuals to individual attributes (b) and outcomes (c).
Results indicated that acquaintanceship networks were of similar size in Spain as reported for the US (DiPrete et al., 2011). Indeed, the number of people respondents knew from the different subpopulations showed variable degrees of overdispersion, indicating that some subpopulations were more unequally distributed over networks than others. The degree of over- or underrepresented was related to individual characteristics, mostly income, education and age. For example, the young and highly educated knew proportionally more people who had emigrated or who had found new jobs in the past three years, but they did not differ in the proportion of network members who lost their jobs. Last, trust in institutions and social support availability were not only related to individual characteristics but also by the clustering in networks, in line with our general expectation, but not in interpersonal trust.
DIPRETE, T. A., GELMAN, A., MCCORMICK, T., TEITLER, J., & ZHENG, T. (2011). Segregation in social networks based on acquaintanceship and trust. American Journal of Sociology, 116 (4), 1234-1283.
MCCORMICK, T. H., SALGANIK, M. J., & ZHENG, T. (2010). How many people do you know? Efficiently estimating personal network size. Journal of the American Statistical Association, 105 (489), 59-70.
Acknowledgements: The authors would like to thank the CIS, and in particular Félix Requena, Berta Álvarez, Carmen Lence and Araceli Mateos.