# Dynamic groups in complex task environments: To change or not to change a winning team?

**Authors:** Darío Blanco-Fernández, Stephan Leitner, Alexandra Rausch  
**arXiv:** [2203.09157](https://arxiv.org/abs/2203.09157)  
**Date:** 2022-03-17  
**Summary page:** https://www.summarizepaper.com/en/paper/2203.09157v1/

## Key points

- Organizations rely on group formation to solve complex tasks
- Groups adapt to the demands of the task by changing their composition periodically
- Previous research on the effects of group adaptation on task performance has produced ambiguous results
- Researchers conducted an agent-based model study using the NK-framework to better understand this relationship
- Reorganizing well-performing groups may be beneficial, but only if individual learning is restricted
- Centralizing decision-making led to better performance than random or decentralized decision-making for low-complexity tasks (K=3)
- Decentralized decision-making was more effective than centralization or randomness for moderate complexity tasks (K=5)
- Group adaptation had a positive effect on performance for both low and moderate complexity tasks when individual learning was restricted
- Group adaptation had a negative impact on performance for moderate complexity tasks when individual learning was not restricted
- The study highlights the importance of considering both individual learning and task complexity when examining the impact of group adaptation on performance in complex environments.
- Organizations should carefully consider whether to change their winning teams depending on these factors.

## Summary

Organizations often rely on group formation to solve complex tasks, and groups frequently adapt to the demands of the task by periodically changing their composition. Previous research has produced ambiguous results regarding the effects of group adaptation on task performance. To understand this relationship better, researchers conducted an agent-based model study using the NK-framework. The results showed that reorganizing well-performing groups may be beneficial, but only if individual learning is restricted. Additionally, when dealing with low-complexity tasks (K=3), centralizing decision-making led to better performance than random or decentralized decision-making. However, for moderate complexity tasks (K=5), decentralized decision-making was more effective than centralization or randomness. When individual learning was restricted, group adaptation had a positive effect on performance for both low and moderate complexity tasks. But when individual learning was not restricted, group adaptation had a negative impact on performance for moderate complexity tasks. This study highlights the importance of considering both individual learning and task complexity when examining the impact of group adaptation on performance in complex environments. It also suggests that organizations should carefully consider whether to change their winning teams depending on these factors.