Teams are frequently treated as more than the sum of their parts in organisational psychology. There is a lot of focus on shared group properties and constructs such as team cohesion, psychological safety, and collective commitment to explain why some teams outperform others. However, a recent analysis says that what we refer to as team dynamics might actually be a statistical illusion known as inheritance.
The main problem with team research is that individual members’ responses are often averaged to create group level data. But individual level relationships shape these groupmean correlations. For example: if individual height and weight are correlated, the average height and weight of a group will also be correlated, even if the group was formed completely randomly. This phenomenon is called inheritance.
It suggests that a significant correlation between teamlevel variables does not automatically prove that a unique group process is at work. On the contrary, the relationship might simply be inherited from the individual level. Hence, to really understand it there is a need to separate the incremental effect.
The distinct contextual influence that exists above and beyond relationships at the individual level is referred to as an incremental effect. Since applying adjustments to the entire group mean correlation is a fundamental error, current team research practices frequently involve correcting grouplevel correlations for unreliability.
Furthermore, the majority of approaches produce greatly inflated estimates of teamlevel effects and fail to distinguish between inherited and incremental components. For instance, a grouplevel effect was initially adjusted upward to a moderate 0.29 in a metaanalysis on psychological safety. However, the true teamlevel contribution was discovered to be a much more marginal 0.13 when the inherited individuallevel effects were appropriately eliminated. This implies that a lot of well known teamlevel synergies might be much weaker than previously thought.
Hence, when the group mean correlation is stronger than the individual level (within group) correlation, a positive incremental effect exists which indicates that the team context is indeed amplifying a relationship. However, negative incremental effects can occur when a group process actually weakens or reverses an individual relationship. For example: in a struggling workgroup, poor performance might lead a team to send more members to training which might result in a negative grouplevel correlation between training hours and performance even if training is effective.
So, how can research related to teams be improved?
Maltarich et al. (2026) proposed a threestep adjustment method:
- Decompose the groupmean association into its inherited and incremental portions.
- Adjust only the incremental portion using teamspecific reliability.
- Reconstruct the population estimate by adding the adjusted incremental value back to the inherited effect.
In conclusion, interpreting teamlevel findings based solely on statistical significance may not be as accurate as we believed it to be. Instead, providing accurate theoretical rationales for why a specific team process such as a shared norm or collective belief would create a unique effect that is distinct from individual attitudes. Hence, by disentangling these levels we can move toward a more honest and effective understanding of how teams truly function.
Food for thought: If many of the “team effects” we celebrate are actually inherited from individual-level patterns, are we overestimating the power of teams in driving performance? Should organisations invest less in abstract team-building interventions and more in selecting and developing individuals? And how might this change the way we evaluate leadership effectiveness?
References
- Maltarich, M. A., Bliese, P. D., & Carpenter, N. C. (2026). Relationships between shared group properties: theory, measurement, estimation, and adjustment. Personnel Psychology. https://doi.org/10.1111/peps.70022