The investigators will randomize 300 cocaine-dependent methadone patients to 1 of 6 conditions: (a) a control group, (b) a contingency management condition that arranges a 100% probability of winning a prize with each draw and has 3 prize categories, (c) a contingency management condition that arranges a 31% probability of winning and has 3 prize categories, (d) a contingency management condition that arranges a 100% probability of winning and has 7 prize categories, (e) a contingency management condition that arranges a 31% probability of winning and has 7 prize categories, or (f) usual prize contingency management with a 50% probability of winning from 3 prize categories. Magnitudes of reinforcement will be identical across conditions, but lower overall probability conditions arrange for greater chances of winning larger magnitude prizes. The investigators expect that the new contingency management conditions will reduce cocaine use relative to the control condition, that 31% probability conditions will decrease drug use relative to 100% conditions, and that 7-prize category conditions will reduce drug use compared to 3-prize category conditions. In addition, the 31%/7-category condition is expected to be most efficacious. Results will be instrumental for further developing prize contingency management to improve outcomes of cocaine-dependent methadone patients.

Inclusion Criteria:

• age > 18 years
• current DSM-IV diagnosis of cocaine dependence
• enrolled at the clinic for >3 months
• on a stable dose of methadone (no changes) for >1 month and not requesting a dose alteration
• submitted >1 clinic cocaine positive sample in the last 6 months
• English speaking
• pass a quiz related to understanding the informed consent form

Exclusion Criteria:

• serious, uncontrolled psychiatric illness
• significant cognitive impairment
• in recovery from pathological gambling