For the past four years, Darren Wheelock and I have collaborated with Charles Franklin and the Marquette Law School Poll on a series of surveys of public attitudes toward sentencing and corrections policy in Wisconsin. Our 2015 results, released last week, seem to show remarkably high levels of support for prisoner rehabilitation. Of those who were asked, more than 80% expressed support for each of the following:
- Expanding counseling programs for prisoners
- Expanding job training programs for prisoners
- Expanding educational programs for prisoners
- Helping released offenders find jobs
At the same time, there are also indications of substantial, if somewhat lower, levels of support for various punitive policies:
- About 47% supported making sentences more severe for all crimes
- About 45% supported locking up more juvenile offenders
- About 62% supported increasing the use of mandatory minimum sentences for repeat offenders
- About 45% supported trying more juvenile offenders as adults
It is puzzling that many respondents expressed support for both pro-rehabilitation and tough-on-crime policies. We have also seen this phenomenon in earlier rounds of our polling.
In this round, we decided to test whether the apparent inconsistency results from “acquiescence bias” — the tendency sometimes found in polling research for some respondents to agree with whatever they are asked, regardless of the content of the proposal. We thus randomly divided our sample in half and “flipped” some of the questions.
For instance, we asked half of the sample whether they support increasing the use of mandatory minimums for repeat offenders, and the other half whether they support reducing the use of mandatory minimums for repeat offenders. The results do suggest that acquiescence may be playing a role in our responses. As indicated above, 62% said yes to increasing mandatory minimums, but we found that nearly as large a majority (58%) said yes to reducing mandatory minimums. Since respondents were randomly assigned to one question or the other, we would expect the demographic characteristics of the two groups to be very similar, which would point to acquiescence as a likely cause for the seemingly inconsistent results. However, we will have to confirm that hypothesis with more rigorous statistical analysis of the data.
My preliminary analysis also suggests that acquiescence may be playing a role with respect to the pro-rehabilitation results. For instance, a whopping 84% supported expanded counseling programs for offenders. When we flipped the question, though, 40% agreed that counseling programs for prisoners should be reduced. Note, however, that while flipping the question diminished support for rehabilitation, a majority of respondents still took the pro-rehabilitation position even when acquiescence would push in the other direction: 53% said they opposed reducing the programs for prisoners.
In short, while the “true” level of support for rehabilitation — that is, the level of support without any inflating or deflating acquiescence effects — is probably less overwhelming than all of the 80+% results might suggest, it is still likely the majoritarian position.