Researchers have started to argue that part-time (PT) employees should not be viewed as one undifferentiated group, leading Martin and Sinclair to develop a PT typology based on employee demographics and personal attachments. Attitudinal and behavioral differences have been found among their groups. This prior research has been limited by classifying work role involvement through descriptions of roles outside of work, rather than assessing it empirically. We use more precise, direct measures of involvement, including time spent on, flexibility of, and psychological involvement in, roles outside work to predict differences in work role involvement, attitudes, and turnover intentions. Moreover, we test the PT typology on a geographically and occupationally diverse sample, overcoming previous generalizability concerns. We found that part-timers with more psychological involvement in roles and those with roles outside work that are less flexible and more time intensive have less work role involvement, higher turnover intentions, and generally less positive attitudes. In addition, the results provide evidence that the more direct measures better predict work role involvement and turnover intent than the indirect PT typology measure. Implications for management and research are discussed. Copyright © 2010 John Wiley & Sons, Ltd.