Berkeley Dietvorst’s main stream of research investigates when and why forecasters fail to use algorithms and explores prescriptions that increase consumers’ and managers’ willingness to use them. His other streams of research address such topics as order effects on consumer choice, choice architecture, and consumers’ reactions to corporate experiments.
In his first article, published in the Journal of Experimental Psychology: General and titled “Algorithm Aversion: People Erroneously Avoid Algorithms After Seeing Them Err,” he demonstrates that people are substantially less likely to choose to use an algorithm in favor of a human forecaster after seeing the algorithm err, even if they have also seen it outperform the human forecaster. Dietvorst found that this preference for human forecasters stems from a quicker loss in confidence in algorithms, even when human forecasters make the same or more numerous mistakes. His most recent article, forthcoming in Management Science and titled “Overcoming Algorithm Aversion: People Will Use Imperfect Algorithms If They Can (Even Slightly) Modify Them,” suggests that one may be able to overcome algorithm aversion by giving people just a slight amount of control over the algorithm’s forecasts.
Dietvorst received a BS in economics and a PhD in decision processes from the Wharton School of the University of Pennsylvania.