Case-population sample extracted from Fang and Gong (2020). The case-population sample is extracted from Fang and Gong (2020) by the following procedure. The case sample is composed of 2,261 flagged physicians (that is, those who billed for more than 100 hours per week). The control sample of equal size is randomly drawn without replacement from all observations and its flagged status is coded missing. The sample is composed of 4,522 physicians who billed at least 20 hours per week.
FG_CPA data frame with 4,522 rows and 5 variables:
indicator: physician is male
indicator: physician has a MD degree
experience in years
1 if an observation belongs to the case sample; NA otherwise
indicator: number of group practice members less than 6
Fang, H. and Gong, Q. (2017). Detecting Potential Overbilling in Medicare Reimbursement via Hours Worked. American Economic Review, 107(2), 562-91.
Matsumoto, B. (2020). Detecting Potential Overbilling in Medicare Reimbursement via Hours Worked: Comment. American Economic Review, 110(12), 3991-4003.
Fang, H. and Gong, Q. (2020). Detecting Potential Overbilling in Medicare Reimbursement via Hours Worked: Reply. American Economic Review, 110(12): 4004-10.