Can We Measure Social Capital in Academia?

Figure 1, above, shows the differences in collaboration habits across the six UCPH Faculties. The Gantt percentage marks along the horizontal axes illustrate where negative (i.e., rarely to never) to positive (i.e., often to always) response tendencies lie, with the response 'sometimes' divided at the 0% mark. In part 3 of the questionnaire, we refer to and measure social capital in terms of 'preferences'. Here we have implemented an exploratory principal components analysis to validate the questionnaire dimensions, as well as a Chronbach's alpha test of reliability. Figure 2, above, shows the scree plot and Table 2 shows the factor loadings for the 27 questionnaire items (varimax rotation). The Kaiser-Meyer-Olkin measure verified sampling adequacy (KMO = .881). Bartlett's test of sphericity X2(351) = 7774.86, p <.001 indicates that correlations between the items were sufficiently large for a principal components analysis. The scree plot shows inflections that would justify retaining 3 to 4 components. A Chronbach's alpha test for the Likert-scale questions about preferences resulted in a range of acceptable to uncertain values. The subscale for the cognitive dimension of social capital consisted of 10 items (α =0.712), the second subscale for the relational dimension of social capital also consisted of 10 items (α =0.736), and the third subscale for the structural dimension of social capital consisted of 7 items (α =0.638). Results also indicated that slight improvements to these alpha values would be obtained by removing questions 8 (from the cognitive subscale), 18 and 21 (from the relational subscale) and 23 (from the structural subscale)