Social Support Typologies: Different Approaches for Reducing Social Support Data

Filip Agneessens, Hans Waege, and John Lievens

Abstract

Social support refers to certain qualitative aspects in social contacts. Information about social support in surveys is usually collected through name generators and name interpreters. Here we use a faster way for asking questions on ego-centred social support-networks and propose a compact way for reducing this information. Our aim is to measure diversity of role relations and diversity of support content in social support networks. We take role relations, rather than individuals, as the unit of analysis for support networks (i.e. the aggregated contribution of persons with the same role relation). A random sample of 623 Belgians living in the Flemish region were asked to name the role relations they can rely on for each of five specific sorts of support. Using latent class analysis (LCA) within log-linear analysis, we focus on aggregated content diversity and aggregated role diversity. We explore whether a limited number of types of support for role relations and of role relations for items of support can be found in the sample. A large number of alternative models are found. We explore several alternative models and subsequently we evaluate the results. Our evaluation is based on the fit of the model (significance of the likelihood-ratio) and the stability of the parameters (identifiable solution). We find that in many cases the kinds of support (that a role relation gives to the respondent) can be represented by a latent variable with a limited number of classes. On the other hand, the types of relations that give a specific kind of support can hardly ever be reduced to an underlying categorical variable..