Journal of Accounting Research 56 (2): 581-615

With V.G. Narayanan.

In this field experiment on over 19,000 students enrolled in HarvardX, we analyzed whether the most effective benchmark for performance was the peer median or rather a higher level, the peer top-quartile. We found that the best benchmark to boost performance depended on the student’s starting point. Lower initial performers did better when they saw the more achievable peer-median benchmark. Higher initial performers did better when they saw the more challenging, peer-top-quartile benchmark. This provides evidence of whether and how the right target for performance depends on the individual.

Link: Publisher

Journal of Accounting Research 59 (2): 657-711

With Patrick J. Ferguson and Sebastian Koppers.

In this field experiment in professional European football (aka soccer), we provided football players with larger or smaller sets of performance measures. We found that less information yielded the best improvement as long as this information focused on peer comparison. This heightened self-reported levels of competitive thinking and led to higher performance.

Link: Publisher

Journal of Accounting Research 58 (4): 1023-1067

The University of Utah Health Care was the first academic health care system to publicly disclose patient ratings of individual physicians. This practice has since spread widely to many other large hospitals. I examined how this affects performance on the ratings and in terms of the quality of care. I found improvements by both measures and that these were driven partly by increased time that doctors spent with patients.

Link: Publisher