The measurement of organizational characteristics in empirical studies that focus on inter‑group, knowledge‑ based collaboration requires research methods different from those used for measuring the characteristics of individuals. As an answer to that, key‑informant methodology is a frequently adopted approach that has been associated with qualitative methods. However, recently organizational researchers have used the technique to obtain quantifiable information on organizational structure, internal power distribution, within the group, and external relationships among groups that base their collaboration on the knowledge they share. This paper focuses on the threats to validity, which are inherent in empirical studies that adopt the key‑informant methodology as a social science tool. In particular, the paper thoroughly examines the effects that the Bagozzi and the Cook and Campbell construct validity criteria as well as the Huber and Power key‑informant validity criteria have during the two important phases of a research, i.e. developing valid measures of the theoretical constructs, and testing the relationships between theoretical constructs. The empirical results used in this paper stem from an investigation that aimed to evaluate the contribution of Shared Knowledge and Information Technology to Manufacturing Performance. Mutual Trust and Mutual Influence, among the collaborating groups (in this case manufacturing, quality and R&D), which were used in our study, as the two antecedents of shared knowledge. For the purpose of this research, an evaluation model was developed and survey data was collected from 51 medium to large size industrial companies with a total of 112 manufacturing groups, representing five industrial sectors (alimentation, automotive, chemical and pharmaceutical, electro‑mechanical, and textile), were analyzed to test the model. The key‑informant methodology that has been used for the selection of research responders was tested against threats to validity. As a conclusion, the paper exhibits the implications of the above widely accepted construct validity criteria and specific key‑informant validity criteria, building upon the results of the above industrial empirical research. The lessons learned are presented in a way that may lead future organizational researchers to error preventive measures.
Keywords: inter-group collaboration, key-informant methodology, threats to validity, lessons learned