Table 1. Characteristics of the four datasets that we considered.
BR, base rate.
COMPAS balanced BR | COMPAS low BR | LSI-R balanced BR | LSI-R low BR | |
Number of cases | 1000 | 1000 | 311 | 1954 |
BR of recidivism | 48% | 11% | 29% | 9% |
Features | Streamlined | Streamlined | Streamlined/enriched | Streamlined/enriched |
Number of responses (no feedback) |
2700 | 2400 | 2850/2500 | 2850/2900 |
Number of responses (feedback) |
2400 | 3000 | 2700/2400 | 3000/2550 |