social capital measurement

general introduction

All references on these pages are simultaneously listed - see the link in the main menu on the left.

During the last decades, social capital appeared as a promising research perspective in the social sciences, resulting in many interesting empirical findings. Progress in the field has been limited by several theoretical problems however, one of which is its measurement (Flap, 1999, 2002, 2004; Lin, 1999, 2001b; Snijders, 1999). Due to both a variety of definitions of social capital on several levels of analysis, and a lack of appropriate data sets, the measurement of social capital has seen many ad hoc constructions of instruments, and has been far from standardised (links to a wide variety of operationalisations and measurement methods can for instance be found on the measurement page of Fabio Sabatini's social capital research directory).

Currently, whereas most operationalisations of social capital are discussed and constructed for the concept as seen on the collective level, there is much more conceptual clarity when it is viewed from a perspective in which it is seen as an individual pool of resources, embedded in personal networks. Therefore, together with Tom Snijders and Henk Flap, I have been working on a research project on the measurement of individual social capital - most of it carried out at the sociology department of the University of Groningen, within the ICS graduate school.


A first aim of this research project was to provide a thorough discussion of the theoretical issues that precede the measurement of social capital, and the resulting choices we could and should make in the design of instruments (Van der Gaag & Snijders, 2004)). This resulted in a specific approach in which we concentrate on measuring individual level social capital: general, positive social resources embedded in personal social networks in the general population. Working within the framework of Lin's 'Network theory of social capital' (1999, 2001a/b), an emphasis is put on an individual's access to social resources: keeping an eye to the possible mobilisation of the measured social resources, but not on their actual use. To do this two currently available measurement techniques, the Name Generator / interpreter (McCallister & Fischer, 1978), and the Position Generator (Lin & Dumin, 1986; Lin, Fu, and Hsung, 2001) are investigated for a representative population sample. In addition, one newly developed social capital measurement instrument, the Resource Generator (Snijders, 1999; Van der Gaag & Snijders, 2005) is tested. (For new developments and overviews of studies using these instruments also see the Position Generator and Resource Generator pages.)

In several parallel investigations, the measurement properties of each of these three instruments are investigated in terms of their internal validity, reliability, parsimony, and effiency in use. An emphasis is put on the valid representation of the three dimensions of social capital: alters, resources, and alters' willingness to give the individual access to their resources (Flap, 2002). Two main types of measures are constructed from Name Generator, Position Generator, and Resource Generator data. First, 'deductive' measures, which are directly based on theoretical notions about a beneficial morphology of social capital - such as its volume or its diversity. However, in the absence of more specific knowledge about the goal specificity and productivity of specific parts of social capital, such measures are not sufficient. In addition, we therefore propose an inductive approach to compose multiple, resource-domain specific measures that are more suitable to answer such research questions (Van der Gaag & Snijders, 2004), building on earlier, preliminary work (Lubbers, 1998). Finally, the intercorrelations between measures from all three instruments are evaluated, and we propose an approach how to use all these various indicators in comparative analyses of returns to social capital (Van der Gaag & Snijders, 2003; Van der Gaag, Snijders & Flap, 2004).

Although not part of this project, the same data also allow for investigations into the link between the above mentioned micro-level measures and at least one indicator often used for the operationalisation of collective level social capital: membership of voluntary organisations. This study is further described in Bekkers et al (2004).

related papers

data and instruments

All investigations in this project are based on data collected in the 'Social Survey on the Networks of the Dutch' (SSND), a large scale face-to-face survey conducted among a representative population sample in the Netherlands in 1999-2000 (N = 1,007) which included Name Generator, Position Generator, and Resource Generator questions. You can download these questionnaire items in English.