10.2190/46U9-T06L-8M32-PEFM 46U9T06L8M32PEFM 211 STUDYING EMPLOYEE TURNOVER BY SPLITTING UP THE USUAL COMPARISON GROUP 211 227 20050916 20050916 20050916 20050916 46U9T06L8M32PEFM.pdf http://baywood.metapress.com/link.asp?target=contribution&id=46U9T06L8M32PEFM 3 HERMAN STEENSMA WIM VAN BREUKELEN MARISKA STURM Leiden University, The Netherlands Dutch Space, Leiden, The Netherlands The purpose of this study is to demonstrate how a mixed method of turnover research, combining advantages of traditional methods while reducing the disadvantages, may improve the fitness-for-use of turnover research for organizations. In a survey study, an exit group of former employees who had left an organization voluntarily was compared with a group of employees who were still working in the organization. Then, the comparison group was split into two subgroups. The true-comparison group was composed of employees who had a low intention of quitting. The potential-turnover group was formed by employees with a high intention of leaving the organization. Results show that, in general, no differences exist between the exit group and the potential-turnover group in mean scores on variables that (according to theory) determine turnover. But the exit group and the true-comparison group differ sharply. Implications for turnover research are discussed.