This study examines the cross-national validity and invariance of the EO measure (Covin and Slevin, 1989) by analyzing a data set spanning 1,200 SMEs from seven countries. Results from the study can be summarized under three main headings: dimensionality, validity, and invariance. From a dimensionality perspective, study results suggest the three-factor representation of EO is most appropriate. This finding is in contrast to previous research which has operationalized the construct in a unidimensional fashion.

This study also examined the convergent and discriminate validity of the constructs. Of the eight items examined in this study, two caused violations of both convergent and discriminant validity and were excluded from the remaining analyses. The multi-dimensional model excluding these items produced significantly better fit statistics than did the model including them. Four of the six items retained in the final scale, those intended to measure innovativeness and risk-taking, were originally developed by Miller and Friesen (1982). The other two retained items, designed to measure proactiveness, were developed by Covin and Slevin (1989). Thus, measurement problems associated with the items may be the result of items coming from differing sources.