Level of analysis considerations in organizational citizenship behaviour research: an empirical investigation of individual and work group effects among hospital employees.

Abstract

An implicit assumption in previous research is that the relationship between job satisfaction and organizational citizenship behaviour is an individual level phenomenon. However, due to the use of raw score correlation-based or related analyses, previous investigations have not shown empirical support for the individual level of analysis. This study empirically tested several relationships between job satisfaction (including facets) and a specific type of citizenship behaviour in order to determine whether such relationships were relevant for individuals or for the entire work group. Results indicate an overall lack of group-based effects. Instead, individual-difference effects represent the significant relationships found in the data. Several null effects were also obtained. These results go beyond the traditional approach to organizational citizenship behaviour research because group-based relationships have been explicitly rejected rather than simply assumed to be unimportant. Evidence for individual-differences was provided by testing for levels of analysis effects in terms of individuals and work groups. Future research should assess the generalizability of these results by including tests for levels of analysis.