Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper studies the effectiveness of treatment for substance abuse with data on more than 10,000 treatment episodes from Maine. We measure effectiveness as the reduction in the rate of drug use between admission and discharge. In a nonexperimental setting we use instrumental variables to estimate the effect of treatment, measured as number of visits, in an ordered logit model framework. After controlling for selection bias, treatment appears to be effective for moderate and heavy drug users. The marginal productivity of treatment increases then decreases. We estimate a treatment "cutoff point" at which marginal productivity becomes zero for both moderate and heavy users.