Success-Biased Social Learning: Heterogeneity and Flexibility based on Perceived Similarity
Having a detailed understanding of social learning is necessary to understand how culture evolves and how behaviors and beliefs spread in a population, including within the firm. Unfortunately, we do not know how complex social learning strategies are and how they vary across individuals and situations. Yet, the details are critical as small differences at the individual level can significantly influence cultural evolutionary dynamics at the population level. So far, researchers often assume homogenous social learning strategies, that is that strategies shared among individuals. At one end of the spectrum, cultural evolutionary researchers often assume simple heuristics, while at the other end of the spectrum researchers assume unconstrained Bayesian updating. Both extremes are probably inadequate. Focusing on success-biased social learning, we study social learning complexity and heterogeneity at the individual level. With an experiment, we show that social learners do not blindly imitate a successful model but instead adjust their strategy based on perceived similarity and relevance of the model. However, these adjustments are not symmetric. Holding the value of social information constant, participants perform better when exposed to social information framed in some ways compared to others. Further, results demonstrate tremendous heterogeneity of social learning strategies at the individual level. Strategies vary between individuals in the same situation and across situations for a single individual. In a gene-culture coevolutionary companion model, we explore how this ability to switch between social learning strategies affect the commonly assumed evolutionary trade-offs when success-biased learning is rigidly applied.