by Prof. Andrew L. Spivak

The incarceration rate, which from the 1920s to the early 1970s hovered between about 100 to 120 state and federal prisoners per 100,000 Americans, has risen nearly fourfold. While the rate of increase has slowed substantially in recent years, the raw numbers continue to climb and many prisons are operating at or above capacity. Along with increasing correctional populations, per capita state prison expenditures have likewise expanded, but spending for building new prisons or renovating existing facilities has not been commensurate with the overall level of growth.

Under these pressures, the country increasingly turned to the private sector to deal with burgeoning prison populations. Americans’ affinity for market economy solutions and a corresponding distrust of government in general led to a belief that the private sector might outperform government agencies in providing incarceration services. Although the proportion of state correctional budgets spent on private prisons is relatively low, about six percent nationally, some states spend a much higher proportion than others.

As of the last available data (about 3 years ago), Oklahoma ranked fourth in the nation with 30% of its funds allocated for private prison expenditures, and was sixth in the percentage of total prisoner population housed in private prisons. Thus, several years ago, the Sooner state seemed like a good place in which to conduct an evaluation of the effectiveness of private prisons, with prisoner recidivism as the comparison measure.

Until last year the only academic studies of private and public prisons using recidivism as a benchmark had taken place in Florida. I had worked for the Oklahoma Department of Corrections since 1997 and regularly dealt with reception and release data, which I had used previously to study recidivism in general. Since inter-facility movement data also was available, I realized that it would be possible to replicate an analysis originally designed by William Bales, a criminal justice professor at Florida State University. Bales had been the research bureau chief at the Florida Department of Corrections (FDOC) in the 1990’s and I met him at a Justice Research and Statistics Association (JRSA) meeting in 2005, where he was presenting a paper that he’d recently published in the journal Criminology & Public Policy.

Bales and his colleagues had used a quasi-experimental design to determine whether spending more time in private prisons affected prisoners’ post-release performance. It didn’t, and the results were exciting because 1) they contradicted some previous findings which suggested that private prisons had a positive effect on prisoners’ post-release performance, and 2) his work was of much higher quality than what had been done in the past. He accounted for more controls (things like age, gender, offense type and prior criminal record), and matched comparison groups more precisely than any of the previous studies. When he explained his methodology to me in greater detail after the presentation, I knew that I had access to roughly the same type of data in Oklahoma, and was determined to find out if that state would show similar results.

To provide some background, privatization of correctional services began to increase during the War on Drugs in the 1980’s and the resultant rapid expansion of prison populations around the country, which led to a search for more economical and effective ways to house prisoners. This movement towards privatization also created concerns about the quality of services provided by private prison companies. The compatibility of incarceration’s goal of societal responsibility with the private prison industry’s inherent profit-seeking motive incited fears that profit-conscious budgets would cut corners more than would allocated public budgets, and these kinds of concerns led to calls for judicial review and an ongoing ethical debate.

Amid this debate, two important topics arose over the past two decades regarding the effectiveness of private prisons in meeting the needs of the government and the prisoner: cost and prisoner post-release performance. Private prison companies purported several ways to reduce the costs of incarceration. One set of researchers, in a 2003 report funded by Corrections Corporation of America (CCA), proposed that private prison contracting leads to cost savings in the public sector due to increased accountability that is fostered by competition. In other words, private prisons supposedly save money because “their existence helps control the cost of public prisons.” Other researchers claimed that private companies can build prisons more quickly and cheaply, although subsequent research at the Bureau of Prisons contradicted that assertion. The debate over whether private prisons are cheaper, the same or more expensive than public prisons is far from resolved, and like many privatization issues, a great deal of additional research effort and subsequent ink will be spilt over it.

The other evaluative aspect of an emergent private prison industry, prisoner post-release performance, is a matter of the degree to which the private industry’s services compare in quality to those of the public sector. The idea is that better ser-vices will lead to prisoners who are better prepared to succeed after release. As early as 1992, an article in The Journal of Criminal Law and Criminology, a relatively high-ranking and prestigious academic publication, claimed that the quality of confinement, using an index with measures of security, safety, order, care, activity, justice, conditions and management, were all better in private prisons than in either state or federal prisons. Interestingly, however, this finding of superior performance held true only when examining staff opinions. From prisoners’ perspectives, the public prisons were actually preferable. Other research claimed that public prisons were better run, and additional comparative studies raised issues involving many different measures, including institutional operations, safety, staff turnover, gang activity, drug use and programming.

All of these performance standards were really peripheral, however, to the key indicator of correctional performance – the proverbial holy grail of prison research – which is recidivism. Saying that a prison, or system of prisons, is more beneficial than another for prisoners due to the quality of its services is a directly testable claim; you just have to determine whether prisoners do better after they are released. It sounds simple, but the methodological issues can be daunting. First, recidivism can be defined in a number of ways, including re-arrest, re-conviction, re-incarceration and even absconding from parole, and the timing of events between release and the recidivistic event can confuse the interpretation of data.

Then consider the fact that prisoners spend varying amounts of time in different types of facilities; they are not randomly assigned as in an experiment and their outcomes are subject to a host of possible confounding factors such as age, race, gender, time served, offense history and level of post-release supervision. Not surprisingly, only four studies prior to mine had attempted to assess the recidivism of prisoners according to their having been incarcerated in public versus private prisons. All four of those studies had been conducted in Florida and used data from the FDOC. The first three, conducted in the late 1990’s and early 2000’s, found some degree of support for a lower rate of recidivism among prisoners released from private facilities, but then along came William Bales and his colleagues, claiming that this was not really the case at all. Who was right?

The first study compared 12-month recidivism rates for 396 male prisoners released from public and private prisons. Prisoners from each group were matched for age, race and number of prior incarcerations, and were tracked for one year after release. Although private prison releases tended to result in better post-release performance (lower recidivism rates), the FDOC rightly expressed concern with the authors’ methodology, citing the small sample size and limited criteria for matching prisoners. An additional study then conducted further analyses on a subset of the original data, using more precise control-group matching and a longer follow-up period of four years. The authors confirmed the previous finding of a lower rate of re-imprisonment among prisoners released from private prison; however, the Florida Ethics Commission raised the issue of a conflict of interest due to the receipt of private corrections industry consulting fees by the group that undertook the project.

In 2002, a new team of researchers in Florida improved on the previous studies’ methods by using a significantly larger sample of prisoners and matching the groups across more criteria (adding offense type, custody level, education and time served). Most importantly, though, they expanded the categorical definition of private and public prisoner groups. Instead of release facility, they used the facility in which the prisoners spent the final six months of incarceration – a huge improvement, since the prior studies would count prisoners as “private” or “public” simply on the basis of the facility from which they were released.

Consequently, if a prisoner was transferred to a private or public facility mere days before his release, after spending years in the other type of prison, a gross misidentification would be represented in the data. In the new study, a regression model indicated there was no difference in recidivism among male prisoners from private and public prison groups, but did find a difference for female prisoners. Women in the private prison group were 25% less likely to re-offend and 34% less likely to be re-incarcerated than those in the public prison group.

Thus was the state of scientific research on private/public prisons and recidivism when William Bales and his col-leagues entered the scene in 2005. What they did was both original and far more sophisticated than any of the previous work. They designed new measures that would capture the degree of prisoners’ exposure to private and public prisons, using categories defined by the amount of time prisoners spent in each type of facility. This quasi-experimental design included multiple treatment and control groups coded by type of release facility, whether the prisoner had spent any time in public and/or private prisons, and the amount and proportion of time spent in each.

Therefore, prisoners who had served a certain amount of time in one type of prison, but no more than a certain amount in the other, would be compared only with those who had symmetrically opposing characteristics. Additionally, the study con-trolled for a variety of possible confounding variables. For Florida prisoners released over a six-year period and tracked for up to 60 months, the researchers found no significant difference between any of the treatment and control groups for adult male, adult female or juvenile male offenders. Spending more time in private facilities, as opposed to public prisons, didn’t seem to improve or worsen prisoners’ likelihood of returning to prison following their release.

But would these results hold up in other states? One thing about Oklahoma that made it such a good place to replicate the Florida studies is that all of the private prisons in Oklahoma (aside from a few halfway houses) were medium security. There were six privately-operated facilities that housed about one quarter of the state’s 24,000 prisoners. They could be directly compared with public medium-security prisons, most of which were about the same size, so the study wouldn’t have to account for differences in security-level.

I decided to create roughly the same types of quasi-experimental categories and control groups as did the Bales team, with most of the same controls (age, race, education, prior incarceration, offense history, post-release supervision, sentence length, time served and proportion of sentence served). However, in addition to the matched groups, I also decided to analyze a regression model with proportion of time spent in private and public facilities as predictor variables, and used all adult prison releases between 1997 and 2001, tracked through 2004.

In both the matched categories and the time-proportion model, spending more time in private prisons was linked with slightly worse post-release performance. In other words, the likelihood of recidivating (returning to prison) increased the more time prisoners had spent in private as compared to public facilities. As stated in the study’s published findings, “the analyses indicate a significantly greater hazard of recidivism among private prison inmates in six of the eight models tested (four of the six exposure and comparison group models and both of the continuous models). In every categorical model (including the two that were nonsignificant), private prison inmate groups had a greater hazard of recidivism than did public inmate groups.”

The effect was modest, only just enough to achieve statistical significance, but it was a complete surprise. I thought I was going to be writing about how my study either confirmed the findings of Bales et al., or supported the older research. However, now I had to explain why spending more time in private prisons led to higher recidivism rates. As a statistician, I prefer crunching numbers to writing (it took me more than a month to write this article for Prison Legal News). So I found a colleague to be my co-author. Susan Sharp is a professor at the University of Oklahoma who writes extensively about prisons and had already authored two books on capital punishment. We set out to package my analysis into a report that a scholarly journal would accept for publication.

Some of my colleagues encouraged me to suggest, in the discussion of my findings, that public prisons were better-run than their private counterparts, a conclusion that would certainly be good news to those who were already skeptical of prison privatization and would deeply annoy private prison advocates. However, I didn’t have any particular axe to grind with the private prison industry. Before leaving to accept a position as a university professor last year, I had worked almost a decade in corrections, starting in the 1990’s as a security officer and later serving as a case manager, but only in state-run facilities. Therefore, other than what I had found in the academic literature and heard anecdotally from fellow correctional employees, I didn’t know much about how private prisons were run compared to public facilities, and nothing that I had discovered from previous research pointed toward definitive quality differences.

So when my co-author and I submitted the findings of the Oklahoma data to the journal Crime & Delinquency, I insisted on playing it safe and making the most intellectually honest conclusion possible. Most likely, we wrote, there is no private/public prison influence on recidivism. Spending more time in private or public facilities probably doesn’t cause sig-nificant differences in prisoners’ post-release performance. More likely, the surprising results of my research were due to an administrative caveat.

When I was a case manager at a medium-security prison in around 1998, we used to get a lot of requests from prisoners for transfer packets (applications) to go to private prisons. They wanted to transfer to the private facilities because the housing units there, unlike those in the state prisons, were relatively new and featured air-conditioning – a big deal when sweltering Oklahoma summers can heat the cells into the 90’s, even at night. Since the state correctional budget was so constrained, case loads were horrendous (I had a protracted double roster of 180 prisoners for several months), and I believe that we unconsciously favored our most difficult prisoners for transfers in order to get them out of our housing units. Those prisoners with time-consuming disciplinary issues, excessive complaints and grievances and so forth were the ones we especially couldn’t wait to transfer ... and they may be exactly the kind of prisoners who would have a harder time staying out of prison after release.

This was my own pet interpretation, and there was no way to verify or disconfirm it, but it got me out of having to conclude that one type of prison was better than the other. Perhaps the next social scientist who compares private and public correctional facilities will figure out a clever way to determine whether I’m right or not. In the meantime, I tend to agree with William Bales that the private prison controversy will probably focus more on cost issues over the next decade rather than alleged differences in quality of service.

*Andrew L. Spivak is an Assistant Professor in the Department of Sociology at the University of Nevada, Las Vegas, and wrote this article exclusively for PLN based on his research study titled “Prisoner Recidivism as a Measure of Private Prison Performance,” co-authored with Professor Susan F. Sharp (University of Oklahoma) and originally published in Crime & Delinquency (July 2008, Vol. 54, No. 3, pp.482-508).