Learning from complexity: Effects of prior accidents and incidents on airlines' learning

Using data on accidents and incidents experienced by U.S. commercial airlines from 1983 to 1997, we investigated variation in firm learning by examining whether firms learn more from errors with heterogeneous or homogeneous causes. We measured learning by a reduction in airline accident and incident rates, while controlling for other factors related to accidents and incidents. Our results show that heterogeneity is generally better for learning, as prior heterogeneity in the causes of errors decreases subsequent accident rates, producing a deeper, broader search for causality than simple explanations like blame the pilot. The benefits of heterogeneity, however, apply mainly to specialist airlines. Generalist airlines learn, instead, from outside factors such as the experience of others and general improvements in technology. These results suggest a theory of learning across organizational forms: complex forms benefit from simple information, and simple forms benefit from complex information. The implications of our study for learning theories and work on organizational errors are discussed.