Results

In order to control for the potential single-source problem in this study and to control for common-method variance, a Harman’s one-factor test was used to address the issue of common-method variance (Schriesheim 1979). This common factor explains around 26 per cent of the variance of the included variables. This is slightly higher than similar studies (Li and Atuahene-Gima 2001; Scott and Bruce 1994) controlling for common-method. Common-method bias is thus not a significant problem for this study. However, large multivariate regression models as analysed in this study are conservatively affected by common-method variance. If the data of this model suffers from common-method variance, the slope estimates decrease as a larger number of measured variables suffering from common-method variance (CMV) are included (Siemsen et al. 2010). In addition, the interaction effects induced by the context are, in particular, central in our study. Interaction effects are found to be deflated because of common-method variance (Evans 1985; Siemens et al. 2010). Overall, the analyses of this article are not driven by common-method variance, and the estimates are considered as robust, taking into account that the estimates will be conservatively affected by potential common-method variance.