Open Access Open Badges Research article

Psychometric evaluation of a short measure of social capital at work

Anne Kouvonen1*, Mika Kivimäki23, Jussi Vahtera3, Tuula Oksanen3, Marko Elovainio4, Tom Cox1, Marianna Virtanen3, Jaana Pentti3, Sara J Cox1 and Richard G Wilkinson5

Author Affiliations

1 Institute of Work, Health & Organisations, University of Nottingham, 8 William Lee Buildings, Nottingham Science and Technology Park, University Boulevard, Nottingham NG7 2RQ, UK

2 International Centre for Health and Society, Department of Epidemiology and Public Health, University College London Medical School, 1-19 Torrington Place, London WC1E 6BT, UK

3 Finnish Institute of Occupational Health, Topeliuksenkatu 42 a A, FIN-00250 Helsinki, Finland

4 National Research and Development Centre for Welfare and Health (STAKES), POB 220, FIN-00531 Helsinki, Finland

5 Division of Epidemiology and Public Health, Community Health Sciences, University of Nottingham, QMC, Nottingham NG7 2RD, UK

For all author emails, please log on.

BMC Public Health 2006, 6:251  doi:10.1186/1471-2458-6-251

Published: 13 October 2006



Prior studies on social capital and health have assessed social capital in residential neighbourhoods and communities, but the question whether the concept should also be applicable in workplaces has been raised. The present study reports on the psychometric properties of an 8-item measure of social capital at work.


Data were derived from the Finnish Public Sector Study (N = 48,592) collected in 2000–2002. Based on face validity, an expert unfamiliar with the data selected 8 questionnaire items from the available items for a scale of social capital. Reliability analysis included tests of internal consistency, item-total correlations, and within-unit (interrater) agreement by rwg index. The associations with theoretically related and unrelated constructs were examined to assess convergent and divergent validity (construct validity). Criterion-related validity was explored with respect to self-rated health using multilevel logistic regression models. The effects of individual level and work unit level social capital were modelled on self-rated health.


The internal consistency of the scale was good (Cronbach's alpha = 0.88). The rwg index was 0.88, which indicates a significant within-unit agreement. The scale was associated with, but not redundant to, conceptually close constructs such as procedural justice, job control, and effort-reward imbalance. Its associations with conceptually more distant concepts, such as trait anxiety and magnitude of change in work, were weaker. In multilevel models, significantly elevated age adjusted odds ratios (ORs) of poor self-rated health (OR = 2.42, 95% confidence interval (CI): 2.24–2.61 for the women and OR = 2.99, 95% CI: 2.56–3.50 for the men) were observed for the employees in the lowest vs. highest quartile of individual level social capital. In addition, low social capital at the work unit level was associated with a higher likelihood of poor self-rated health.


Psychometric techniques show our 8-item measure of social capital to be a valid tool reflecting the construct and displaying the postulated links with other variables.