Innovation from Imitation: Error Correction Mechanisms in Social Networks can Reduce Accuracy and Encourage Innovation

• In Event: 584 - Regular Session. Social Networks, Cognition, and Meaning

Abstract

An enormous body of scholarship examines how social contagions spread over social networks. But despite the fact that humans make mistakes, diffusion through social networks is typically modeled as though information is relayed with perfect fidelity. As a result, little is known about how errors, the robustness of message formats, and attempts to correct errors, impact the spread of contagions. We find in an experiment using human subjects that compact (high entropy) message formats, such as text messaging pidgin, are more prone to error than more expansive (lower entropy) formats, such as standard English. We also find that while efforts to correct mistakes improve the accuracy of individual messages, over time they paradoxically generate a wider variety of mutant forms of the contagion than would result from a lack of correction. This indicates that the ability of messages to cross even “small-world” human social networks may be overestimated and that models of social contagion and diffusion are biased upwards. This research also suggests that human tendencies towards conformity can generate diversity; failed error corrections create new versions of a contagion that diffuse in competition with the original, allowing innovation to result from imitation.

Authors

• Matthew E. Brashears, Cornell University

• Eric Charles Gladstone, University of South Carolina