Is it more effective to collaborate in large teams or in smaller ones? Should we work alongside people who are similar to us or those who have different perspectives? New findings shed light on these questions, revealing some surprising insights.
Every manager faces a similar dilemma, especially in large corporations: what is the best way to organize employees to spark innovative ideas?
Is it more advantageous to be part of bigger groups or smaller ones? Should we collaborate with people who share our viewpoints or with those who differ?
Recent studies from Binghamton University, State University of New York, aim to answer these questions, presenting some unexpected outcomes.
For almost two decades, Professor Shelley D. Dionne, currently the dean of Binghamton’s School of Management, along with Distinguished Professor Hiroki Sayama, have researched intricate group dynamics and their influence on organizational success. Their latest paper, published in the journal npj Complexity, examines how distinct groups of individuals approach creative tasks.
Joining them in this research are Yiding Cao, MA ’16, PhD ’23; Yingjun Dong, MA ’16, PhD ’22; Minjun Kim, PhD ’24; Neil G. MacLaren, MBA ’18, PhD ’21; Sriniwas Pandey, PhD ’23; and Distinguished Professor Emeritus Francis J. Yammarino.
The study involved 617 Binghamton University students engaging anonymously through a Twitter-like online platform, tackling one of two assignments: create a catchy marketing slogan for a new laptop or craft a fictional story. Participants were divided into groups of 20 to 25 and logged in for about 15 minutes each day over 10 working days to submit ideas and provide feedback on others’ contributions.
Before the experiments, language analysis of self-introduction essays helped categorize students based on shared opinions or backgrounds. Some participants could interact with the whole group, while others communicated only with their immediate “neighbors” in a ring-shaped organizational layout.
To assess the quality of the ideas generated, the researchers enlisted PhD candidates specializing in marketing or management for the slogan task, and staff from the University’s Division of Communications and Marketing for the story-writing assignment.
The researchers acknowledged that several of their findings may be counterintuitive.
“Connecting all participants in a social network so everyone can see each other’s ideas resulted in reduced idea diversity,” observed Sayama, a faculty member at the Thomas J. Watson College of Engineering and Applied Science’s School of Systems Science and Industrial Engineering.
However, having the opportunity to communicate with many others increased participants’ satisfaction. In contrast, those who interacted with fewer people felt more isolated but produced higher-quality ideas.
Moreover, when individuals from diverse backgrounds collaborated, their ideas tended to become more conservative as everyone rallied behind “safer” options based on their areas of expertise.
“Randomly connecting individuals was actually the most effective method for generating high-quality ideas,” Sayama noted.
He likens the idea generation process to evolution, where ideas function like organisms that either adapt and flourish in a biodiverse ecosystem or perish due to environmental challenges.
“When two people converse, one represents an island and the other another island. There’s a channel linking both, with the ideas exchanged like birds or fish moving between the two,” he explained.
“Each mind holds countless idea ‘organisms.’ These ideas can migrate from one brain to another; if they are welcomed, they replicate; if shunned, they die off. Thus, the entire process revolves around fostering biodiversity.”
Sayama acknowledges that applying similar experimental setups in traditional workplaces would be challenging. While the researchers utilized AI to evaluate backgrounds and text submissions, managers can simplify their focus: what is the primary goal? Ensuring an array of ideas demands a different approach compared to merely preserving a particular idea.
Dionne believes the research’s strength lies in its interdisciplinary approach, which explored creative processes in networks.
“Linking idea generation to evolutionary processes allowed the research team to establish a shared framework for testing our assumptions; however, this framework was merely a starting point,” she remarked. “Without each team member’s unique expertise, we might not have fully understood collaborative behaviors within social network structures.”
Sayama added, “Readers of our paper may derive various insights from these findings. This is far from the final conclusion of the project; rather, it opens up numerous new avenues for exploration.”
The experiments were conducted before the COVID-19 pandemic and during its initial months, but the team became involved with the University’s public health decision-making and the project was paused.
“While we had collected data, the pandemic left us without the mental and physical bandwidth to analyze our findings and compile a paper,” Sayama shared. “But now, it’s finally being published.”
Even while the project was on hold, Dionne felt inspired by the team’s progress: “Reflecting on 2018, ‘artificial intelligence’ was hardly discussed in management literature, let alone used as an analytical tool. I knew that once we could reconvene, we had a rare chance to advance the field.”
