We undertook several procedures that Podsakoff, MacKenzie, Lee, and Podsakoff (2003) recommend to reduce and evaluate the magnitude of common method bias, as well to assess construct validity. First, our focal explanatory variables – namely the measures of relative performance – have been derived from archival sources. Second, to verify that common method variance is not a substantive concern for the capability-based variables, we adopted the two-step approach suggested by Anderson and Gerbing (1988) to estimate measurement models. First,