The concept of collaboration cosmopolitanism has referred to the institutional and geographic distance characteristics of academic researchers’ collaboration patterns. We study the effect of collaboration cosmopolitanism on doctoral level research personnel working in different sectors–government, industry, and academia. The study examines the impact of collaboration cosmopolitanism on an important aspect of career success: job satisfaction. We employ the 2006 and 2010 Survey of Doctorate Recipients (SDR) collected by the US National Science Foundation to evaluate the cross-sectional and longitudinal effect of collaboration cosmopolitanism on job satisfaction. We are particularly interested in doctoral level researchers’ job satisfaction related to sector of employment while controlling for demographic and work characteristics such as gender, minority status, salary, and work hours. Findings suggest that scientists working at a higher level of collaboration cosmopolitanism tend to report a higher level of job satisfaction. Furthermore, we find that academic scientists are more satisfied than those working in industry. This finding holds in the longitudinal model—industry scientists are less satisfied—but we find that over time, government scientists are more satisfied than academic scientists, and much more satisfied than industry scientists.

Introduction

We know that collaborative research is productive, both by perception and by measured gain (Abramo et al., 2009; Andrade et al., 2009; Bruneel et al., 2010; Hamann et al., 2011; Lee and Bozeman, 2005). Our interest is in the impacts of research collaboration, not only on science writ large but, especially, on the careers of STEM researchers. That is, our concern is much more personalistic than previous studies of collaboration’s effects on productivity: what is the relationship of collaboration to job satisfaction? Our research focuses not on research collaboration experiences of all types but rather one specific aspect of collaboration, one we refer to as “collaboration cosmopolitanism” (Bozeman and Corley, 2004; Lee and Bozeman, 2005). The basic idea of collaboration cosmopolitanism is the extent to which researchers work with persons who are distant from them either institutionally or geographically. We shall later provide specific measurement details but for the present let us say that a very low degree of collaboration cosmopolitanism would be signified if all of one’s collaborations are with people in the same laboratory. By contrast, the most cosmopolitan collaborators would be those who work with people in different laboratories, organizations, or nations than their own. The central research question then is this: To what extent, if any, do more cosmopolitan collaborators differ from less cosmopolitan collaborators with respect to job satisfaction? We look at different nuances and variations of this question, focusing especially on differences according to sector of employment − academic, government and industry. Today, research is a generative process as collaborative teams attract more collaborators, thus accelerating the growth of research teams (Parker and Hackett, 2012). Furthermore, research takes place in a variety of settings not exclusive to universities (Roach and Sauermann, 2010), and scientists exhibit different preferences for the sector in which they wish to be employed (Agarwal and Ohyama, 2012; Fox and Stephan, 2001; Janger and Nowotny, 2016). Collaboration across sector, disciplines, organizations, or countries becomes more prevalent in individuals’ daily work life with the flow of globalization as well as the emphasis on interdisciplinary efforts seeking innovative solutions to the complex social problems such as health care or environmental issues (Van Rijnsoever and Hessels, 2011).