Sticking it all together: A critical assessment of the group cohesion–performance literature

TLDR

Group cohesion is widely studied but suffers from inconsistent definitions and measures, leading to confusion, despite a general consensus that it is positively associated with performance, though the mechanisms and dimensionality remain unclear. The paper reviews the current state of cohesiveness research and calls for grounding it in broader theory and longitudinal, sophisticated analyses to resolve inconsistencies. The review focuses on the cohesion–performance relationship.

Abstract

Although group cohesion is a widely studied construct in the group dynamics literature, there is a considerable lack of consistency and agreement regarding the construct. This paper reviews the current state of the cohesiveness literature. Cohesion is now generally defined as the group members’ inclinations to forge social bonds, resulting in the group sticking together and remaining united. Unfortunately, the large number of definitions and measures used by researchers has created a literature that is inconsistent and confusing. This review focuses on the considerable research on the cohesion–performance relationship. The general consensus of this research is that there is a generally positive association between cohesion and performance, but the theoretical and operational mechanisms creating this association are unclear. This lack of clarity may be due to the wide range of measures and assessments of cohesion and the cohesion–performance association. Two underlying contentious issues within this aspect of the literature are whether cohesion is unidimensional or multidimensional and the appropriate level of analysis required for measuring the construct. The resolution of these inconsistencies requires researchers to take a step back toward grounding the research in broader theoretical frameworks, to determine whether the cohesion–performance association has an underlying common cause, and to use longitudinal studies that use more sophisticated analytical models.