Šafr, Jiří. 2009. „Position generator and measurement of social capital in an egocentric network.“ Data a výzkum – SDA Info 3 (2): 211-240.
The Position Generator (PG) represents one method of measuring egocentric social networks, and key facets of social capital. Respondents are asked if they know a person from a list of jobs that have different social status. The Social Distance Survey (2007) fielded a Czech version of the PG which examined 18 jobs and investigated the strength and duration of ties, and gender of contacts. In this article, we first compare distributions obtained from the PG with the same occupations in population (egos) and from the name generator. Second, measures of social capital were computed. These include extensity, upper reachability, range and an aggregate index called ‘Access Social Capital.’ There are also estimates of lower reachability, mean and total status in a network. In addition, new measures are introduced such as (a) ‘average status combined with status range’ which reflects the “double advantage in networks”, (b) gender and strength of tie diversity, (c) relative measures of gender/ status congruence, and (d) inductive scales measuring access to high and low status professions. Validity of selected social capital measures is assessed using regression models that are operationalised with key socio-demographic variables, and indicators that measure the ethnic and educational diversity within ego networks. These models reveal that differences in the stock of social capital are primarily influenced by education, ISEI (an occupation status), and employment status. The most important relation is in between an ego’s status and a mean network ISEI score, upper reachability, and their interaction. This finding implies that these network measures best capture the concept of hierarchically ordered social resources. The validity of the PG is also assessed using a correlation analysis of the effects or outputs of the social network, i.e. income, job mobility, social trust, life satisfaction, and tolerance of ethnic groups. The article concludes with a comparison with other egocentric social network techniques and recommendations for further work.