29.04.2014 Views

0 - National Criminal Justice Reference Service

0 - National Criminal Justice Reference Service

0 - National Criminal Justice Reference Service

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Outpatient, Ancillary Groups or Drug Education only) due to a shortage of intensive treatment<br />

slots. Critical to the matching process is the use of matching criteria closely related to the<br />

outcome criteria (e.g., recidivism and drug use).<br />

Matching must be sensitive to the three principles of risk, need, and responsivity derived<br />

fiom empirical research. First, effective programs clearly differentiate between low-risk and highrisk<br />

clients (Andrews et al., 1990; Bonta, 1996; Gendreau, 1996; Jones, 1996). High-risk cases<br />

should receive high levels of intervention and services; low-risk cases should receive minimal<br />

intervention. Second, criminogenic needs are dynamic @e., changing) risk factors that are<br />

predictive of recidivism (e.g., antisocial cognitions and emotional states, association with<br />

antisocial peers, substance abuse, weak self-control and problem solving skills). Programs that<br />

0<br />

effectively target and reduce such individual needs accomplish larger decreases in re-offending.<br />

Third, programs that appropriarely target the specific needs and leaning styles of their clients are<br />

more effective.<br />

One of the drawbacks of matching is that many cases drop out as the number of matching<br />

variables increases. It is thus paramount to use a small number of relevant predictors to select<br />

candidates for the comparison group. Using aggregate matching, overall distributions in the<br />

experimental and control groups are made to correspond on the criterion variables. Although<br />

individual matching is more precise @e., a “partner” for each treated client is selected from the<br />

unexposed group of offenders), individual matching is far more expensive, time-consuming, and<br />

difFcult to execute for a large number of matched variables. Research has consistently indicated<br />

the priority of several static predictors: offender age, age at first arrest, number of previous<br />

convictions (property, person, or drug offenses), and prior history of drug use (Andrews et al.,<br />

40<br />

This document is a research report submitted to the U.S. Department of <strong>Justice</strong>. This report has not<br />

been published by the Department. Opinions or points of view expressed are those of the author(s)<br />

and do not necessarily reflect the official position or policies of the U.S. Department of <strong>Justice</strong>.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!