The Effect of Total-Identification on the Performance of Computer-Mediated Group Idea Generation

아이디어 실명제가 컴퓨터 기반 그룹의 퍼포먼스에 미치는 영향에 관한 연구

컴퓨터 기반 브레인스토밍에 전통적으로 적용된 아이디어 익명제는 생산성 향상을 위한 새로운 방법을 모색하는 연구가들에 의해 문제점으로 지적되어 왔다. 익명성이 유도하는 프로세스 손실 (social loafing, downward productivity matching 등)을 극복하기 위해 시도된 실시간 생산성 피드백기능 조차 시간이 경과함에 따라 생산성이 질에서 양으로 변질되는 경향을 보여왔다. 본 연구는 이러한 관찰된 단점을 극복하기 위해 실시간 생산성 피드백기능에 실명제를 추가하여 통제 효과를 실험하였다. 익명제에 비해 월등한 생산성향상은 유도되지 않았지만 실명제의 효과는 양과 질에서 분명한 차이를 보였다. 따라서 앞으로의 연구에서는 실명제를 보완할 새로운 요소를 찾아내는 기법에 대한 고찰이 필요하다고 보여진다.

Results of more than three decades of research have provided evidence that computer-based idea generation is more than a management fashion. However, given that the performance of computer-based groups when compared to that of (paper-and-penciled-based) nominal groups did not yield substantial productivity differences regardless of group sizes and other factors, computer-based idea generation appears to be another management fad. One reason for such productivity paradox stems from the absence of clear task performance information because all contributions are generically anonymous, inducing social loafing and downward productivity matching – a tendency of productivity matching to a perceived baseline. Prior studies introduced a real-time performance feedback mechanism as an external intervention to combat process losses due to anonymity of comments. Despite a significant performance improvement, we observed a consistent limitation that is the tendency of performance competitiveness from quality to quantity toward later stages of the idea generation session. In this study, we introduced a situational factor, i.e., individuals’ real names as an additional external intervention variable to increase the degree of identifiability and evaluability of individual outputs, while maintaining the same quantity feedback. The results showed that (1) expected statistical significances were yielded for both quantity and quality of ideas and (2) performance competitiveness did not diminish throughout the idea generation session in the total-Identification condition. However, given that the magnitude of the effect of total-identification when compared to the effect of pseudonymity on ideation performance doesn’t seem large enough, a pursuit of further research to identify factors that could reinforce the effect of total-identification is necessary. The results also showed that individuals’ real names in conjunction with performance feedback did not have an effect on reducing the number of junk comments. The notions of face-saving and comments filtering are suggested to fine-tune the interface of computer-based idea generation.