In this paper we present experimental evidence suggesting that observational learning (i.e. learning by not doing but by observing) may outperform learning by doing.