Social networks : an examination of social network measurement, and the relationship between social networks and health among older adults in New Zealand : a thesis presented in partial fulfilment of the requirement for the degree of Master of Science in Psychology at Massey University, Auckland, New Zealand

Abstract

The initial aim of this study was to explore the relationship between social network types derived from Wenger’s PANT and health among adults aged 55-70 years old in New Zealand. More specifically, it tested the “Mezzo” level of the Conceptual Model of health derived by Berkman et al. (2000), where it was postulated that social contextual factors (gender, education, socioeconomic status and ethnicity) shape the type of social network an individual is embedded in, and these social contextual factors contribute to the relationship between social networks and health. This study used data obtained from the second wave of the New Zealand Longitudinal study of Health, Work and Retirement, which used responses to a questionnaire survey from a representative sample (N=2430). Bivariate correlations and multiple regression analyses were run to test these associations. However the initial results showed theoretically incorrect correlations between social network types, and a large proportion of the population sample were allocated into “inconclusive” and “borderline” categories. Furthermore, although the social contextual factors clearly influenced the variation in health, the relationship between social network types and health were rather weak, especially in comparison to those found in other studies. Therefore, the study took a turn to examine and rescore Wenger’s PANT. The modification of the measure ameliorated the peculiar relationships between network types, reduced the number of participants in the “inconclusive” and “borderline” categories, but did not significantly improve the relationship between social networks and health. The only network types that contributed significantly to the multiple regression equation were the Locally Integrated and Wider Community Focused network types. Therefore it was speculated that the possible “active ingredient” of social networks may be the aspect of social engagement, where those who have more ties to their friends and community are in better health than those in restricted and family focused networks. This study only showed moderate support for the Conceptual Model. However it did provide evidence towards the need to explore the notion of social engagement and integration, as well as the development of a social network measure that includes perception of network ties, and the functional roles within social networks, not just the structural aspects..