How Cognitive Diversity and Attention Change Group Discussions?

A group of 129 university students tested a new online discussion system where a conversational agent (a kind of chatbot) suggested ideas. Researchers split participants into four groups: cognitive diversity (very different ideas) or homogeneity (similar ideas), and with or without instructions on how to pay attention to the bot's proposals. There was also a control group that discussed in pairs without a bot.

The findings are striking. When the bot offered diverse viewpoints, discussions were broader: students explored more topics and generated more original ideas. Conversely, when opinions were similar, conversations became deeper, with more detailed and well‑developed arguments.

Attention played a double role. Those who received instructions on how to listen to the bot saw its ideas as an obstacle, reducing the number of valid ideas. Those without instructions produced more correct ideas and understood the topic better.

In practice, cognitive diversity pushes thinking outward, while homogeneity favors deeper reflection. And attention, if mishandled, can stifle creativity. These results help design better virtual assistants for schools, making group discussions more effective and fun.

The concept of cognitive diversity stems from the theory that groups composed of members with varied knowledge, experiences, and thinking styles can generate a broader range of solutions than homogeneous groups. In HCI (Human‑Computer Interaction), conversational agents can be programmed to deliberately introduce contrasting viewpoints, acting as "thought catalysts."

In this study, the independent variable was manipulated with a 2×2 design: (1) cognitive diversity (diverse vs. homogeneous) and (2) attention (with instructions vs. without). The control condition allowed isolation of the bot's intervention effect. Discussion "breadth" was operationalized by counting distinct topics that emerged, while "depth" was assessed based on argumentative complexity and the number of references to sources or concrete examples.

Statistical analysis employed a two‑way ANOVA with Bonferroni correction for multiple comparisons. Results showed a significant interaction between diversity and attention: the "diversity‑with‑instruction" condition yielded the greatest thematic breadth but lower depth compared to "homogeneity‑without‑instruction." This suggests that guided attention can inhibit the ability to integrate new ideas, likely due to added cognitive load.

From a neurocognitive perspective, diversity stimulates the fronto‑parietal network responsible for creative thinking, whereas homogeneity favors activation of the dorsolateral prefrontal cortex, involved in detailed processing. Instructional attention may activate the executive control system, reducing cognitive flexibility.

Practical implications are significant for designing intelligent educational systems: it is essential to balance the amount of diversity introduced by the bot with scaffolding strategies that do not overload learners. Future research could explore adaptive feedback based on real‑time engagement monitoring, such as eye‑tracking or language analysis, to dynamically optimize cognitive diversity and the required level of attention.