Petersilia, J. and E.P. Deschenes. "What punishes? Inmates rank the severity of prison vs. intermediate sanctions."
A sample of 48 prisoners in two main receiving facilities in St. Cloud and Stillwater, Minnesota were identified because they were likely to be targeted for intermediate sanctions and would thus have perceptions about the severity of such sanctions that might be considered relevant. Data was collected by means of interviews with inmates whose official corrections file data was RAND staff coded for various demographic and criminal history data. The data was analyzed to test for differences using ordered logistic regression and various models compared by using the chi-square differences from likelihood ratio tests. The results supported the hypothesis that it is not necessary to equate criminal punishment solely with prison given that many inmates have a greater dread of intensive probation.
This research is based on questionable assumptions about the value of the opinion of inmates as to what would be the worse punishment, prison or intermediate sanctions. The results are affected first by asking only prisoners who might be considered for intermediate sanctions, though this is a logical decision to make and a good sample to consider. Means and standard deviations are developed for the rank orders of the 15 sanctions in the survey questionnaires, and the author uses this to suggest that inmate consensus is greatest at the highest and lowest levels, meaning either $100 fine or 5 years in prison. Similar results are found in the overall ranking order of the various sanctions. Ordered logistic regression is used to model the ordered categorical responses as a function of the type of sanction, and this was used to test for differences between individuals. Using this method, it was found that there were no differences by race, prior prison experience, employment history, drug dependence, or safety. The authors ...