摘要：During the last decennia, social capital has been a popular topic in social sciences. However, the concept is often used as a “catch all” for divers social characteristics. Consensus on how the concept should be measured is lacking. This keeps health researchers from drawing firm conclusions on the influence of social capital on health and hampers the decision on which social capital indicators to use in health research. This study compares five ways to operationalise social capital (generalized trust, a social network index, an expanded social network index, bonding, bridging and linking social capital and a multidimensional social capital index) in their ability to explain self-rated health and pain. To evaluate the models’ capacity to explain health, two logistic regression models were built, resulting in Nagelkerke R2 measures. Data were collected in a cross-sectional study in eight neighbourhoods in the region of Ghent (Belgium) by randomly sampling 50 adult inhabitants per neighbourhood. Findings show that the explanation of the observed variance in health by the studied social capital models ranges from 1.9% to 23.1%, but is more pronounced for self-rated health than for pain. The multidimensional social capital index explains most of the variance in health, but poses an important strain on the respondents due to a large number of surveyquestions. With some prudence, we presume that the explanatory added value of the more extended social capital models is rather limited from a practical point of view as the addition in explained variance of the other models seems not in relation to the number of questions needed. Researchers should weigh up the pros and cons of different manners to measure social capital carefully, taking the goals and focus of their study into account.
