This paper examines the relationship between consumers' health plan ch
oices and health plan performance ratings. We make rn,se of an initiat
ive at a large firm to collect aggregate, and disseminate to employees
plan performance ratings. We estimate several statistical models, inc
luding share equations-which allow for the presence of important unobs
erved plan attributes-and logit models. Although report card ratings a
ppear to be related to enrollment choices, the relationship is not uni
form. For some dimensions of performance, the results are consistent w
ith the hypothesis that employees respond to the performance ratings.
For other dimensions, the ratings seem less influential than other pla
n attributes that employees likely observed without the data release.