Validity and reliability
All of the questionnaire items were validated by nine healthcare scholars and senior nurse managers. Subsequently,construct validity was examined using confirmatory factor analysis (CFA). To avoid errors in relation to many items analysed in the CFA, the study exploited the
‘half reliability’ approach with respect to the average for each factor, by comparing odd numbers and even numbers (Weng et al. 2010). For instance, this study divided the eight items that were designed to measure the degree of knowledge creation into two composite indicators; the average value for the odd-numbered items was designated as Knowledge creation 1, and the average value for the even-numbered items was designated as Knowledge creation