Quantifying the Impact of Social Influence on the Information Technology Implementation Process by Physicians: A Hierarchical Bayesian Learning Approach

Abstract

Technology implementation at the individual level within an organization, after the organization has adopted the technology, has been an ongoing challenge in every field. In this study, we develop a hierarchical Bayesian learning model to examine the impact of social learning, through both targeted early adopter effects and general peer effects, and experiential learning on the information technology implementation process by physicians in a community health system. Our unique data allow us to disentangle the most common and challenging endogeneity issues associated with most social influence studies. We find that the experiential learning signal is more accurate than the social learning signals in the technology implementation process; and, between the two types of social learning signals studied here, targeted early adopter effects are much more informative than general peer effects. Furthermore, we experiment with several policy simulations to illustrate and quantify the two different types of social influence on this implementation process. The simulation results suggest that maintaining consistency in technology usage by targeted early adopters is more effective than increasing the frequency of their technology usage in reducing their colleagues’ perceptions of uncertainty about the new technology. More specifically, we find that technology implementation probability would increase: (a) by 15%, on average, by adding a targeted early adopter to a group without early adopters; (b) by 25% by adding peer effects to solo users; and (c) by 47% by adding early adopter effects to solo users. The model can be adapted and generalized to other similar settings that examine social influence on the technology implementation process and also provide quantifiable measures of the improvements that the interventions may produce.