When Less is More: The Proportion of Creative Members and R&D Team Innovative Performance

ABSTRACT

Researchers have long assumed that a high proportion of creative members in a team is associated with high team innovative performance. This study proposed that the relationship between the proportion of creative members in a research and development (R&D) team and the team’s innovative performance is an inverted U-shape. Using a sample of 120 R&D teams with time-lagged and multiple-source data, we found support for this argument. Furthermore, we theorized that the human resource management practices of participative management and pay dispersion moderate that U-shaped curvilinear relationship. Our data show that the curvilinear relationship holds when high participative management was implemented and when pay dispersion is high. We derive implications for management theory and practice.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Our arguments relate to task conflict over ideas, territory and resources as task conflict is associated with the pattern of relationships that we hypothesize (Xie et al., 2014). Although researchers have differentiated between task and relationship conflict, research shows that lower levels of task conflict may be beneficial to innovation and higher levels is detrimental to innovation (De Dreu, 2006). We acknowledge that task conflict may also be harmful at higher levels also because it degenerates into relationship conflict (Simons & Peterson, 2000). However, for simplicity we rely on arguments relating to task conflict and use the terms team conflict and task conflict interchangeability.

² Supervisors rated only permanent employees, so temporary employees (e.g., part-time research assistants and project assistants) were not included in this measure. For the firms with single R&D teams–that is, those included in our sample–members made up over 80% of the firm’s entire R&D personnel, as reported by the firms and verified by our independent count.

³ We were unable to identify a logical decision rule regarding which team to include for firms with more than one team. Taking the average of multiple teams did not seem justifiable because our research focused on the role of creative members and their relationships with other members within a single team, rather than across multiple teams. It was, moreover, not possible to take the average of all R&D teams as representing the organization, because the HCCP database does not include all R&D teams of each company with multiple R&D teams.

⁴ We acknowledge that team supervisors may have rated team members with an assumption, which may hold true in the context of R&D teams, that creative members are more competent than less creative individuals. However, our measure of the proportion of creative team members is not based on this assumption–rather we construct our measure from the aggregate number of members in each category. Supervisors are likely to be accurate in differentiating between creative and less creative personnel in R&D teams are creative activities are salient in that context.

⁵ The results remained largely consistent even without the inclusion of the control variables.

⁶ In addition, we used Cook’s distance to examine the impact of outliers on the result; values larger than 1 are thought to signify potential outliers (Cook & Weisberg, 1982). We found a similar pattern of results after we removed one outlier on this basis. None of our moderated results were significantly affected by the removal of this outlier. These results are available from the authors.

⁷ Our sample of 120 R&D teams includes teams where supervisors opted not to label any member as a creator as well as teams where every member was labeled as a creator. Therefore, the range for proportion of creative team members is 0% to 100%, in Figures 1, 2, and figure 3.