(w/ Alison Wood Brooks) During every turn of every conversation, individuals decide: Should we stay on this topic or switch to a different one? Across thousands of synchronous and asynchronous dyads, we study the heuristics and biases that people use to navigate topic selection in cooperative phatic conversation. First, we show that humans have room for improvement in topic preference detection - knowing whether people want to stay on topic or switch, based on what they have said about it. We use natural language processing algorithms as a performance benchmark against pairs of strangers and close others. Second, we show that humans have room for improvement in topic management - and that simple interventions that empower people to switch topics more frequently, or plan topics in advance, can lead to Pareto improvements in their conversation together.
Michael Yeomans is currently a post-doctoral fellow at Harvard Business School, studying computational social psychology. His research applies machine learning and natural language processing to understand and improve how people communicate with one another. He completed his Ph.D. in Behavioral Science at the University of Chicago Booth School of Business.