A brief motivational interviewing intervention with prisoners: when you lead a horse to water, can it drink for itself?

Abstract

Offenders’ readiness to engage in changes that will reduce their risk of reoffending is now recognized to be as important as the design and delivery of programmes that support such change. Interest is growing in both how to increase engagement in change processes, and how to measure any improvements in engagement. This study evaluated the effects of a brief offending-focused motivational interviewing (MI) intervention on reconviction in male prisoners serving sentences for diverse crimes. Men who undertook MI were significantly less likely to be reconvicted than those who did not. The results also served to validate a stage-based measure of readiness to change derived from Prochaska and DiClemente's Transtheoretical Model. Prisoners who were offered MI increased their readiness to change by an average of one stage, while the scores of men who were not remained unchanged. Furthermore – whether men undertook MI or not – change in stage of change predicted reconviction. This was a high-risk sample, making the results intriguing for at least two reasons. First, reductions in recidivism are usually achieved only with much more intensive programmes for high-risk men. Second, according to ‘traditional’ cognitive–behavioural rehabilitation theory, programmes need to target change in dynamic risk factors directly to reduce reconviction risk. That these results were obtained with men whose initial motivation was low, and in the absence of any ‘traditional’ criminogenic rehabilitation, raises questions about whether there is more than one mechanism involved in desistance.

Keywords:

• motivational interviewing
• Transtheoretical Model
• stages of change
• motivation
• prisoner rehabilitation
• criminogenic programmes

Notes

1. The New Zealand male prison population comprises approximately 49% New Zealand Maori, 38% New Zealand European and 11% men from other Pacific Islands (e.g. Samoa, Tonga).

2. The manual is available from the first author.

3. Data for ANOVAs were not normally distributed. However, repeating analyses with transformed data did not change the significance of findings, as would be expected given the large effect sizes. Statistical assumptions for other analyses were met.

4. In these analyses, both MI and TAU samples include a mix of men who did and did not also complete a criminogenic programme.

5. The difference between the Time 1 RTC means here and in Table 1 arises because these analyses compare only those men with Time 2 data. No information was available on why these data were missing. It is most likely that the men were not reassessed because prison staff failed to follow the required procedures.

6. We did not conduct parallel analyses for re-imprisonment because of small sample sizes.

7. We had planned to measure self-efficacy in this research, but in the first study from this project, the data from our chosen measure (the GSE scale, Schwarzer & Jerusalem, 1995) had a median score of 40 (out of 40, n = 260), so they had to be discarded. A more sensitive measure of self-efficacy would be a valuable addition.