Table 4.
Diabetes
| Statistical Discrimination | ||||||
|---|---|---|---|---|---|---|
| Dep. variable: Diagnosis of Diabetes | Traditional Disparities (1) | Bayesian Model (2) | Excluding X* Effects from (2) (3) | Adding Race Effects to (3) (4) | Adding Gender to (3) (5) | Adding Age to (3) (6) |
| Signal | 2.780 | 2.851 | 2.887 | 2.882 | 2.897 | 2.862 |
| (27.87)** | (12.44)** | (13.97)** | (13.92)** | (13.60)** | (12.81)** | |
| Prior | 0.984 | 5.878 | 5.962 | 6.124 | 1.830 | |
| (0.22) | (4.49)** | (3.95)** | (4.45)** | (0.73) | ||
| Prior X signal | −0.948 | −1.136 | −1.233 | −1.255 | −0.947 | |
| (0.41) | (0.59) | (0.60) | (0.63) | (0.44) | ||
| Race X signal | 0.048 | 0.072 | ||||
| (0.20) | (0.29) | |||||
| Race (minority) | 0.170 | 0.125 | −0.053 | |||
| (1.47) | (0.51) | (0.23) | ||||
| Female | −0.210 | −0.212 | −0.249 | |||
| (1.96)* | (1.73) | (2.36)* | ||||
| Age: 31–41 years | 0.221 | 0.206 | 0.141 | |||
| (0.92) | (0.79) | (0.55) | ||||
| Age: 42–60 years | 0.659 | 0.637 | 0.540 | |||
| (2.64)** | (1.78) | (1.72) | ||||
| Age: 60+ years | 0.807 | 0.767 | 0.621 | |||
| (3.43)** | (1.41) | (1.61) | ||||
| Other controls | ||||||
| Health status | Yes | Yes | Yes | Yes | Yes | Yes |
| Patient's SES | Yes | Yes | Yes | Yes | Yes | Yes |
| Insurance, type of visit | Yes | Yes | Yes | Yes | Yes | Yes |
| Geography | Yes | Yes | Yes | Yes | Yes | Yes |
| Clinician character | Yes | Yes | Yes | Yes | Yes | Yes |
| Observations | 8,084 | 8,084 | 8,084 | 8,084 | 8,084 | 8,084 |
| Bayesian information criterion | −1,550 | −1,532 | −1,564 | |||
*
Significant at 5 percent;
**
Significant at 1 percent.
Robust z statistics in parentheses.
X*=(race, gender, age); dep.=dependent; SES=socioeconomic status.