Alternative Sentencing Policies for Drug Offenders

Introduction: Linda Truitt, National Institute of Justice

Linda Truitt: Good afternoon. Does everyone hear me OK?

Good afternoon — there we go. My name is Linda Truitt. I'm with the National Institute of Justice, which is part of the Office of Justice Programs and the U.S. Department of Justice. NIJ is the development, research and evaluation department. Our job is to study objectively, independently and using evidence based knowledge to develop tools and information particularly for state and local interests.

Today's panel is on alternative sentencing policies for drug offenders. Its based on an NIJ grant to evaluate the effectiveness of Kansas Senate Bill 123, which — as you'll learn — mandates drug abuse treatment for nonviolent drug possession offenders in lieu of prison.

Very briefly, I just want to describe NIJ's portfolio of drugs and crime research. Here, we develop and evaluate effective law enforcement, court and corrections responses to criminal behavior related to alcohol and other drugs. This research informs crime reduction through several approaches listed here, including epidemiology, prevention/intervention, drug markets, market disruption and technology.

This important policy analysis presented today is a genuine researcher-practitioner relationship and partnership, and I want to acknowledge a couple of our audience members who are the executive director and the research director of the Kansas Sentencing Commission. Thank you very much for coming today.

Our panel begins with Roger Werholtz. He is the secretary of corrections for the state of Kansas and a member of the Kansas Sentencing Commission. They work closely with NIJ's researchers on this project for quite a while now. Mr. Werholtz chairs also the Kansas criminal justice coordinating council and is a member of several councils concerned with several things, including re-entry, sex offenders, rehabilitation and restoration, information technology, mental health services, substance abuse prevention, and community corrections.

Next, NIJ's grantees will present findings from their research on criminal justice processes and offender performance.

Dr. Andres Rengifo is an assistant professor in the department of criminology and criminal justice at the University of Missouri, St. Louis. His interests include communities and crime that emphasize neighborhoods, social networks and informal controls.

Dr. Don Stemen is an assistant professor in the criminal justice department at Loyola University, Chicago. His interests focus on sentencing and corrections, in particular innovation in policies and their impact on agencies and other systems. They began this grant at Vera Institute of Justice, which graciously allowed the researchers to continue working independently on this project with the Sentencing Commission.

Please note that Ill leave about 20 minutes at the end of this for Q and A after hearing final comments from Mr. Werholtz. Ill advise the presenters to take some questions but mostly for points clarification. And I think we also have some handouts that you may want to get at the end of this. If we don't have enough to go around — which I don't think we do — please feel free to contact the researchers at the end of this to request more information.

Lastly, I just want to point out some of the information resources we have at NIJ. All of the work products that result from our cooperative agreements and grants and contracts are made available through several archives. The first is NCJRS, which houses all of the work products that are written — reports and other. Also NACJD, where we house all of the data archives — whether its qualitative or quantitative research results. And also for this panel, I wanted to note the commissions Web site, which provides information on SB123 and a host of other information.

So I'll start with Mr. Werholtz.

Roger Werholtz, Secretary of Corrections, Kansas Department of Corrections, Topeka, Kan.

Roger Werholtz: When we were talking about how to organize this, my role really devolved to talking about the history and implementation of Senate Bill 123. If you want to examine the mechanics of that bill and the implementation processes, go to that Web site; its a 135-page document that offers incredible detail on what the content of the legislation is and how the bill was actually implemented over time and how its evolved in our state. The other thing that I have become aware of is that Ill have to be careful not to jump into any hyperbole here. I have to be very careful about stating the facts because as Linda pointed out, Helen Pedigo, who's the executive director of the Sentencing Commission, is here so I've got a fact checker in place, and I don't want her to get up and say, What the secretary meant to say was this, that or the other. One more disclaimer I'm not an attorney; Helen is an attorney, and if you have some interest in the actual mechanics of the legislation ... Were you working for the legislature at the time that this was passed or were you still with JJA?

Helen Pedigo: [Inaudible.]

Werholtz: At any rate, there's the expert for a lot of the legal mechanics and the legislative mechanics for this particular piece of legislation. Now to get to the point of why I'm here.

Senate Bill 123 was one of the products of the Kansas Sentencing Commission, which has a statutory requirement that any time our states prison population reaches 95 percent of capacity or is projected to be reaching 95 percent of capacity within the coming year, it is incumbent upon the Sentencing Commission to propose one or more alternatives to construction of prison beds. One of the duties assigned to the Sentencing Commission was to create a nonpartisan, objective mechanism to depoliticize the process of coming up with alternatives to incarceration. And in the 2003 legislative session, this was the product put forward by the Sentencing Commission. In going over some of the context of what was happening at that time — this came two years after the passage of Senate Bill 323, which had two components to it — it increased the amount of good time that individuals could earn, and it shortened the period of post-incarceration supervision that offenders would serve after they were released from prison. It had the effect of very rapidly reducing the prison population and the parole population because of folks who suddenly had their periods of post-incarceration supervision shortened. And one of the byproducts of the passage of that legislation for one of the politicians who supported it was losing an attorney generals primary that everybody in the state assumed that he would win. Obviously folks in the legislature — particularly folks who were supportive of policies that would help control the prison population — were a little bit gun shy about supporting various kinds of alternatives to incarceration. But two years later we were facing the prospect of having to add capacity to the prison system in the very near term, and the Sentencing Commission — in fulfillment of its statutory duties — proposed Senate Bill 123.

What that piece of legislation proposed were that individuals with certain low-level criminal histories who were convicted of — essentially possession only — drug offenses could not be sentenced to prison but instead were sentenced to community-based supervision supplemented by intensive community-based substance abuse treatment.

Is that us that's doing that? I don't know.

The stated intent ... I'm just having way too much fun up here. I suppose its the chocolate mousse that's kicking in.

[Laughter.]

The stated intent of Senate Bill 123 was to provide offenders with the needed level of treatment, coupled with the appropriate level of supervision. Obviously, the underlying intent was to prevent growth in the prison population. The initial proponents of this legislation were the — one of Helens predecessors as executive director of the Sentencing Commission and the chair and the vice chair of the Sentencing Commission — which were a district court judge and a district attorney from one of the more populous counties in our state. As a brand new secretary of corrections who was working for a democrat in Kansas — and democrats in Kansas are kind of an endangered species — working for a minority-party governor, my official position was neutral on the legislation, but the bulk of the testimony that I offered was how it could be successfully implemented. And as the session continued to progress, the governor declared her support for the legislation, and it was otherwise supported by treatment professionals and advocates throughout the committee testimony. The primary opponents of the legislation — as one might expect — would be prosecutors and in some instances some law enforcement folks. But the interesting piece of history about this was having the district attorney from one of our most populous counties — who was also the vice chair of the Sentencing Commission — able to take on his peers in that committee testimony and actually challenge a lot of the assumptions that they put forth. But that gives you kind of a sense of what the political atmosphere was like in our state at the time that this legislation was initially proposed.

The primary issues in terms of how to implement the legislation actually may be of greater interest if you consider a policy like this in your own state. One of the fundamental questions was where to place it. If you're familiar with Kansas, our field services structure is chaotic and extremely fragmented. We have 105 counties in Kansas; every country has three adult field services agencies that supervise felony offenders. They are configured in 31 judicial districts, 30 community corrections programs and two parole regions, and none of those boundaries — that's a little hyperbole; I'm sorry, Helen — some of those boundaries don't overlap.

Traditional probation is supervised by the administrative judges of 31 judicial districts. The employees are state employees, and the operational funding comes from county governments. Community corrections, which supervises high-risk probationers, are county employees whose operational funding comes primarily from the state. They are configured into 30 geographic organizational areas, which sometimes follow judicial district lines, sometimes combine judicial districts, and sometimes divide judicial districts. And then parole is configured into two regions, which are operated by the Department of Corrections; the employees are state employees, and the operational funding comes through the state.

So one of the fundamental questions is who's going to supervise these folks? Well the analysis that was done by the Sentencing Commission research folks indicated that the bulk of people who were picking up what would become Senate Bill 123 sentences were supervised by traditional probation or court services. But court services was actively lobbying not to supervise these individuals. What eventually evolved was that the community corrections program stepped forward and said, if we are adequately funded with treatment resources and if we are given adequate financial resources to hire the additional supervisory personnel, we will take on — what was at that time estimated to be I think — about an additional 800 individuals — which essentially would be transferred from court services to community corrections.

Now this is one of the mistakes that we made that as researchers I think most of you would say, Well of course, you should have considered this, but we made an assumption that because most of these individuals were on court services they must be low-risk people. Nobody was doing any kind of risk evaluation at that time. We are now, as a state, heavily committed to the LSIR, but nobody was doing it at that time. And what we ultimately learned was that instead of our assumption of 20 percent of them being high risk, and 80 percent of them being moderate or low risk, it turned out that 80 percent of them were high risk and 20 percent were moderate or low risk. So [this is] one thing to consider in your own implementation story if you decide to embark on replicating our experience. At any rate, the decision was made to place this particular responsibility for supervision with the community corrections programs in our state.

The second responsibility that came with Senate Bill 123 was certification of drug treatment, and that was assigned to my agency. The rationale for assigning that responsibility to us, I think, primarily had to do with the level of confidence that the legislature had with the Department of Corrections to see that a program was properly implemented, but it created some interesting dynamics for us in that we are not the state agency that's responsible for licensing substance abuse treatment programs. And in essence, it gave us credentialing and oversight responsibility for programs that were traditionally outside our bailiwick and created an opportunity for us to come into conflict with a peer agency. So the particular approach that we took was to say that we are not going to try and duplicate either the credentialing responsibilities for individual practitioners nor the licensure of substance abuse treatment programs. Those responsibilities will remain with the department of social and rehabilitation services. But what we will do to establish [a] certification process is to try and ensure that the agencies wishing to be certified as an S.B. 123 recipient — which means that they would have access to that funding — can demonstrate some competency and expertise in working with offenders. The approach that we took to trying to put that component into place was to say that you must demonstrate competence in delivering cognitive behavioral treatment. You either need to show that that's integrated within the curriculum of your particular treatment program, or you need to go through training that will be provided by the Department of Corrections — not to teach you how to do substance abuse treatment but to teach you how to work with criminal populations.

The third thing that occurred — and this is one that I never ceased to be amazed that it happened and the first part of this ... I've already started saying it so its going to come out wrong — but what we did was work with treatment providers to develop a set of services and a set of funding caps and costs for unit of service for each of the S.B. 123 treatment components. Now that part seems like an obvious thing to do, and that's not the part that surprises me — we ought to do business this way. The part that surprises me is that when we came back to the Sentencing Commission with the fee structure for this particular set of treatment components, it was significantly higher than what the state had typically paid for its other treatment services. And the Sentencing Commission signed off on that, and I thought, Well, OK; I'm a little surprised by that, but the legislature will never let this happen. When it went to the legislature, the legislature signed off on it, too. So we then had two different sets of treatment menus with two different sets of unit costs attached to them, and the set that was tied to Senate Bill 123 was significantly higher than the set that was tied to traditional state-funded substance abuse treatment. It was one of the hopes of this legislation that we would, by virtue of providing this additional revenue stream, create additional substance abuse treatment capacity in our state — that was the second incorrect assumption.

What happened instead was that the Senate Bill 123 people got into treatment faster — which is a good thing for criminal populations — but they displaced the people who were getting access to treatment capacity coming through the SRS, our traditional substance abuse services funding streams. So we displaced the population instead of growing treatment capacity throughout the state; something for you to consider if you embark on this particular solution.

The other thing that I think we learned in implementing 123 was the importance of providing regular oversight. Initially it was intended that the community corrections programs be the gatekeepers, both in the sense of connecting offenders to treatment programs — and that is a role that they did and continued to perform — but they were also supposed to be the performance monitors. The role of the Sentencing Commission was initially to simply evaluate the performance of the programs over time. But what we learned were two things: The community corrections programs felt they didn't have the capacity to manage the billing processes (that defaulted to the Sentencing Commission), and a number of the community corrections programs felt that they could not call their substance abuse treatment providers into account for their performance or lack of it — and this is a function of being predominantly a rural state. What the directors of the programs were telling us is, I don't want to tell my treatment provider that they're not doing a very good job because they're the only one I've got in my district, and I cant afford to alienate them. So the responsibility for doing the performance accountability piece again has defaulted to either the Sentencing Commission or to my agency to carry out. But that's been one of the other experiences that we've had that I think has potentially been problematic.

The only other comment that Id make at this point is what has happened to the bill over time, and for some of those details — during the Q and A piece — we may want to defer some of the questions to Helen. But what I liked so much about the initial version of Senate bill 123 was that for the treatment population, it acknowledged the reality of relapse and the expected occurrence of relapse, and so one of the attractive features of that legislation, from my point of view, is that if somebody came up with a dirty U.A., they could not be revoked and sent to prison. That was simply not an option. Repeated treatment episodes were an option. They way that you got out of Senate Bill 123 behaviorally was committing a new offense or simply becoming so uncooperative with the treatment process that your behavior was disruptive to the treatment program. But relapse in and of itself was not a gateway to get somebody back into prison. That particular feature has eroded a little bit over time; the limitless chances don't exist anymore. And the breadth of the target population has been narrowed over time. There are still attacks that are made on 123, typically by prosecutors. When we get into an analysis of some of the data that Don and Andres have to share with you, there's been some discussion about the cost benefit analysis that has been done. Rather than jumping into that now and anticipating what they have to share with you, I think it would be much more interesting to you to go through the material that they have. And then we can talk about how this debate has matured and evolved in Kansas and what the prospects for the future of Senate Bill 123 look like, so let me defer to the experts here.

Don Stemen, Assistant Professor, Department of Criminal Justice, Loyola University Chicago, Chicago, Ill.

Don Stemen: You'll have to defer beyond me. So, let me make sure Roger erased this for me.

Roger Werholtz: I was having so much fun with that.

Stemen: It looked like it; it did make me nervous, too.

Audience Member 1: Changing all the results.

Stemen: Yes, that's right. Andres and I had a long discussion about what we should title this. I wanted to call it Nothings the Matter With Kansas, maybe The Road From El Dorado. Andres chose this one. Its exciting, straightforward and functional, which is fine. Linda pointed out at the beginning that much of the research that NIJ funds — in fact all of it — is a partnership between researchers and government and this project in particular has been a strong partnership between ourselves and, and the folks in Kansas — both Roger and his staff at the Department of Corrections, who have been patient with us and provided us with much data after repeated requests and changes to what we wanted. The Sentencing Commission has been great and helped us a lot with advice, with assistance in understanding what's going on in Kansas, with accessing data for us and helping us understand how to work with the data, and NIJ with their continued support of this project and advice early on about the way to approach what we were doing there.

I just want to say one thing about S.B. 123 — its unique and not unique in different ways. There have been a number of states in the last 10 years or so who've taken on an initiative of mandatory drug treatment in some way in the community in lieu of incarceration. California's Proposition 36 did it; Arizona's Prop 200 attempted the same thing and Kansas — both through voter initiatives, though. Kansas's initiative is the first legislatively adopted statewide initiative to provide mandatory community-based drug treatment in lieu of incarceration. Other states have done it on a smaller scale through legislation. Texas has a small program for people who are state jail based or who would be directed to state jail in the state for low-level drug possession. Indiana has what they call a forensic diversion program, which attempted to do the same thing but was watered down soon after its passage.

Kansas is unique though, in a statewide effort to provide mandatory drug treatment in the community in lieu of incarceration for those convicted of low-level drug possession. Roger touched on several of the things that were going to talk about. The goals of the program as we see them were twofold. One, to affect the state at a systemic level — right — to affect practices within the criminal justice system. Sentencing practices to change the way that people were being sentenced in the state for low-level drug possession to have them sentenced to the community-based corrections rather than to prison. It was intended to affect supervision practices in the state. Not just who was supervising drug possessors in the community but how they were supervising drug possessors in the community — the point about understanding that relapse is part of treatment, understanding that the focus on partnerships with drug treatment providers was important to the success of the program. And what we found in early conversations with folks in the community corrections districts who were responsible for supervising offenders was that these treatment partnerships were happening, that changes in focus or approaches to supervising drug offenders in the community was changing. And it was also intended to have an impact on treatment provision in the state — increasing availability of treatment, the number of treatment providers in the community. So at the systemic level it was intended to have an impact — reducing reliance on prison, changing practices in the community. And it was intended to have an individual-level impact as well, to actually change recidivism rates, to change offender behavior, to change substance abuse patterns among offenders.

What Ill talk about is the systemic level and how successful the state has been in actually changing practice in the state; and its been very successful. Andres is going to talk about the individual-level impact where there's success but not at the same scale as at the systemic level. And well talk a little bit at the end about how these things go together and how it may, you know, how you may evaluate the overall success of the program in lieu of these two kinds of separate goals that the program is intended to have.

As Roger pointed out early on in his talk, the point of S.B. 123 was to change sentencing practices. People, prior to implementation of S.B. 123, those convicted of ... This is S.B. 123-eligible cases; these are offenders convicted of a first or second offense of simple drug possession. Andres and I had a long conversation about what was complicated drug possession. We couldn't come up with what it was. Nonetheless, simple drug possession, who have low-level criminal histories — criminal history score in Kansas, E through I, which means they have no prior person offenses. They're also ineligible if they have any prior drug sale or manufacturer offenses. So when we talk about S.B. 123-eligible cases, we mean people convicted of simple drug possession, a first or second offense, who have no prior person offenses, drug or manufacturer or sale offenses.

So prior to S.B. 123 most people were being sentenced to court services which is, as Roger said, was standard community ... standard probation in the community. Its minimal supervision; its minimal conditions; its for relatively a short period of time — about 12 months — and that's what most people were getting. Few people were getting community corrections; about 30 percent were going to that, which is somewhat more supervision, more conditions than court services, still supervision in the community. And about 10 percent, 8 percent were going to prison prior to implementation. And you can see afterwards, sentencing practices changed dramatically. Right now, most people are going to S.B. 123 — about 70 percent — and it plateaued fairly quickly. And the rest are dispersed through court services, community corrections and prison, which is, as Roger pointed out, the intent. These people would be shifted off of the court services caseload and onto the community corrections caseload. Now, initially, the impact on admissions directly to prison looks slight. It decreased from about 10 percent of all cases prior to implementation to about 4 percent afterwards. Later well talk about how many people that actually turns into over the first five years of the program — but sentencing practices changed quite quickly.

If we just look at this, which puts it in a different context rather than over time, you can see the shift. One thing to consider is ... well, there are two things. One is the burden placed on community corrections. Prior to implementation of this act, they supervised about 30 percent of drug possessors convicted in the state. They now supervise 88 percent of all drug possessors convicted in the state. As Roger pointed out, that was the intent. They were willing to take on the burden.

[Inaudible.]

Sorry about that; the middle one should be community corrections. Sorry, the light blue one is community corrections. I apologize. Andres was supposed to type ...

[Laughter.]

I was the conceptual person. Andres was the detail man. Nonetheless, so for community corrections, it had a strong impact on agencies in the state. They are now burdened with 88 percent of drug possessors. Prior to implementation, they were burdened with about 30 percent.

It has a profound impact on individuals, as well. Prior to implementation of S.B. 123, 62 percent of drug possessors in the state received minimal supervision, minimal conditions while in the community, a short period of time, about 12 months, that they were supervised. Now 88 percent of drug possessors in the state receive much more supervision, a lot more conditions under community corrections, and spend about four more months under supervision in the community than they would have prior to implementation — about 16 months. When we talk about at the end, the impact of the program and individual outcomes — matters a lot, right? When you talk about impacts on recidivism rates, the more you're supervised, the more conditions you're put under; the longer you're supervised, the more likely you are to fail — through revocations primarily. And we see that later when Andre — I don't mean to steal your thunder — later at the end of the presentation.

Well also see how community corrections has responded over time, though. When this first happened right after passage and they brought on all of these drug offenders onto the cases of community corrections officers, they continued to supervise these offenders much the way they had supervised anyone else in the community — monitoring, surveillance, enforcement — in the same way they had in the past, rather than, as Roger pointed out, looking at this as a treatment population, looking at relapse as part of treatment, as part of the process of treatment. And it took time for that change to kick in among probation officers. And we can see that in their use of supervision interventions and treatment interventions over time.

But you can see there is still 30 percent of what we deem eligible offenders not going to S.B. 123. And that's a questions for us is Why is that? which we haven't determined entirely yet. So there's something happening in the state, whether its about circumvention by judges or something else going on. But we can see that its across all different criminal history scores. Again this is criminal history E through I, which means people have no prior person offenses they've been convicted of. And you can see among this group that this particular group — criminal history score ... they either have no prior criminal history or they only have prior misdemeanor convictions. Of any of the groups on this chart, this is the group that should all be going to S.B. 123, but there's a sizeable number that haven't over the life of the program.

S.B. 123 is meant to be mandatory is two senses — one on judges, in that they had to send eligible offenders to the program. There was some ability to depart from that if they provided reasons, if they could sentence someone to another sanction. Its also mandatory on offenders, that they cant opt out of the program. This is also what makes it different than things like Proposition 36. Prop 36, you can be sentenced to Prop 36 and then decide you don't want to go and take whatever other sentence you're given. In Kansas that's not possible. You have to go to S.B. 123 and serve the sentence in the community with treatment. So we find some apparent circumvention here, but we also find it at the other end of the spectrum in that many offenders who aren't eligible for S.B. 123 are nonetheless getting sentenced to S.B. 123. These offenders with criminal history scores A through D are people who have prior person offenses in their histories and who would be ineligible for the program. Again, judges can depart and sentence someone to S.B. 123 who is otherwise ineligible. Nonetheless, there's some circumvention of the law going on among judges. So its chain sentencing practices almost entirely the way they expected, but there's still these instances in which its not happening.

What we found when we were — we were looking across the state is that these circumventions were rather concentrated in particular areas of the state. There was a divide between rural and urban counties in their departure from the mandatory nature of S.B. 123. There's also a difference between the rural and urban jurisdictions in their supervision practices of S.B. 123. So there's something going on in the state around implementation in these jurisdictions that were still trying to understand.

So this shift in offenders from court services to community corrections for their supervision has also led to a pretty big change, as I said before, in the length of time they're actually supervised in the community. You can see S.B. 123 offenders spend about 16 months in the community under supervision. S.B. 123 allows judges to sentence people up to 18 months of supervision and treatment in the community. There's no requirement that they sentence folks for up to 18 months, but that's generally the sentence that's imposed. But you can see its significantly longer than what they were getting prior to implementation — about four months. And that four months matters a lot given the conditions of supervision that they have under community corrections and how different they are from court services.

And this is partially due to the, as I said, to the structure of S.B. 123 rather than any change in practices. So these have several ... implications, sorry. The number of offenders sentenced under S.B. 123 has grown steadily throughout this period. The system itself has seemed to adapt fairly well to that, given ... even given the volume, the change in volume of offenders coming through, there's no real change or delay in processing, in assessments for offenders before they're sentenced to S.B. 123 or afterwards; the state has responded very well to that. Community corrections seems to have responded very well to that, to the increases in case loads. There has been a reduction in the number of people going to prison under S.B. 123. While it may appear slight in percentage forms, in the absolute numbers that well see over a five-year period, its actually quite large for a state the size of Kansas, and it has a profound impact on the costs of prison in the state.

And overall the system appears to be working the way it was intended. There are some issues and some challenges still though. This determination of eligibility for S.B. 123 is still a question: whether everyone understands eligibility even after five years of implementation. Whether everyone agrees with the eligibility criteria under S.B. 123 is another question. Whether judges are going around the law for certain groups of offenders who either aren't deemed the right population for community-based supervision and treatment. There's also these urban and rural differences in the practices in the state, and this overall increase in sentence lengths, none of which may have been intended during ... at the time of implementation.

So the second piece that we wanted to look at was whether supervision and treatment practices have changed as well. Sentencing practices clearly have. Supervision practices also have changed quite profoundly in the state. We looked at the types of interventions that community corrections officers were using. As Roger pointed out, there was supposed to be a change in their approach to the supervision of drug offenders in the state. We wanted to look and see how probation officers were using interventions that they could use — employment programming, education programming, mental health programming and reintegration programming. You can see over time, there's a huge increase in the number of mental health interventions that community corrections officers are using. This may be a nature of dual diagnosis in the state, that many of the people who are drug offenders also have mental health issues that the community corrections officers are trying to address. But you can certainly see over time — after about 18 months of program implementation — the number of interventions goes up significantly and has continued to rise over the study period.

The other interventions show less of a dramatic increase except for in unemployment programs, where the community corrections officers have also started to use those a lot more. So we see a change a bit in the use of these supportive — what we've called supportive supervision practices — and we've seen a decrease in what we've decided to call restrictive supervision interventions. Those things like curfew, increased surveillance, restrictions on movement, reporting practices, things like that. Particularly around issues of restrictions on movements and increased surveillance, probation officers have started to use these a lot less over time, so we do see some change in their approach to supervising this particular population, at least in the use of their restrictive supervision interventions that they've been using.

As far as treatment goes, we haven't seen a real change in treatment modalities imposed. For the most part they've remained fairly stable, particularly those around outpatient, individual or intensive treatments. There's some decrease in the use of outpatient group treatment and an increase in relapse treatment provision, which you would expect over time; you would expect the use of relapse modalities to increase, but generally its remained fairly stable. What has changed, and as I said before, is this approach to treatment as a team approach. Probation officers and counselors alike told us that they see themselves as working as a team now? That they're both invested in both supervision and treatment and understanding the needs of offenders in the community and working together to make sure those needs are met. Probation officers and treatment counselors meet regularly to devise plans, to reform those plans over time, to make sure that they're meeting all of the needs of offenders.

Two things that we have noticed in the state around supervision practices and treatment practices that ultimately affect the overall impact of the program are the concentration of S.B. 123 sentences. These counties — oh man, I cant count them that fast — 10 counties account for more than 50 percent of all S.B. 123 cases in the state. There are 105 counties in the state and more than half of all the S.B. 123 offenders come from just ten counties, which means that the success of the program is heavily dependent on the supervision practices of these 10 counties. In a similar way, they're heavily dependent on a certain group of drug treatment providers. The dark line here represents — the dark blue line — represents the total number of treatment interventions that have happened in the state. You see they've gone up dramatically since implementation. The red line represents the percentage of those interventions that are provided by the top 15 treatment providers in the state. There are about 120 drug treatment providers in the state ... Is that right? One-hundred, forty drug treatment providers in the state and just 15 of them account for more than half of all treatment provided in the state. So treatment is heavily concentrated in very few providers. The success of the program is heavily dependent on the success of these particular providers to provide treatment in the state.

What this means for the state is that the availability of treatment is very limited by the market. As Roger pointed out, the program was to rely on existing, private drug treatment providers in the community — non-state drug treatment providers. So its heavily dependent on the treatment that they provide, where they happen to provide it, the number who happens to be providing it. So what's been happening is, we found — to a large extent through our discussions with probation officers — was that treatment decisions were often made based on availability, access to providers rather than needs. And this concentration in a few providers has created this particular approach to treatment. But S.B. 123 also hasn't created incentives for more people to enter the market. The number of treatment providers in the state has not grown over the last five years as was initially expected. As Roger said, they pay more for treatment under S.B. 123 than they did for other state ... I'm sorry, I'm [drawing] a blank. Than other treatment paid for by the state. But that increase in the amount paid for treatment did not translate into an increase in the market. And what you find in the state, as Roger pointed out, in very rural parts of the state, there's one drug treatment provider, and if they're bad it doesn't matter, that's where the people have to go; that's the only treatment provider. Or you may find that the only person who provides residential treatment is hundreds of miles away from the person who needs residential treatment. So they either don't get residential treatment because its too burdensome to send them there or they're sent far away from their community and family to get residential treatment. So what happens is, people have changed their approach to treatment — that you may not get what you need, you may get what's available right now.

And what this is — the next question, this is where Ill end — is the impact on the prison system. As Roger pointed out at the beginning, this bill was passed at a time when the prison population was reaching capacity, was going to go over capacity. This bill was intended to reduce the number of drug possessors going to prison as a way to prevent the prison capacity, or the prison population from going over capacity. So the question is what impact had it had on the prison system. In the first five years of the program, S.B. 123 averted approximately 463 sentences to prison for drug possessors; of course, that's only part of it. This is directly from court. So that reduction from about 10 percent to four percent in the number of people going to prison has translated to about 463 people being averted from prison. That's only half the question of its actual impact on the prison system because, as Andres is going to talk about, its also about revocations to prison. And the shift in people from court services to community corrections has changed revocation rates; the revocation rates for this population have gone up. The next question is how much has that affected the impact on the prison system? And Ill leave it for that for Andres to talk about this individual-level impact of the program that we've also been looking at.

Andres Rengifo: So during the past couple years we've been looking at offender-level impacts of S.B. 123, trying to assess to what extent the passage of this law can be linked to changes in behavior for offenders going through S.B. 123 and to what extent those changes in behavior are of greater value to the state of Kansas compared to what other drug possessors are getting in the state, either because they're going to prison or because they're going through standard community corrections or court services. So today were going to be talking mostly about recidivism — looking at the effect of S.B. 123 on reconviction and revocation rates for S.B. 123 offenders but also comparing those rates to similar offenders going through alternative sanctions — again, mostly court services, prison or standard community corrections.

Now a separate question is about the content of treatment, and were at this point trying to work with the administrative data produced by the Department of Corrections to look at what predicts success within S.B. 123; what combinations of treatment modalities are related to low recidivism rates. Were also trying to expand on the list of outcomes to move beyond criminal recidivism to look at other aspects of offender performance that can be linked to participation in S.B. 123 in terms of compliance with treatment but also in terms of risk levels and addiction scores. For the most part today were going to be talking about recidivism.

So the first set of questions have to do with offending and failure while on community supervision. For the most part, what were trying to do again is comparing S.B. 123 to the alternative sanctions that were available to the state of Kansas prior to the implementation of S.B. 123 and after implementation of S.B. 123 because — as Don pointed out — we have a substantial number of drug offenders still being sentenced to court services or standard community corrections or court services. We also, again, have a set of questions about other outcomes related to S.B. 123, and that's preliminary work; we can talk about that work in the Q and A.

So just to reiterate here, we are looking at S.B. 123 illegible offenders, so that's offenders who have been convicted of first or second simple drug possession who have nonperson offenses — that's criminal history scores E through I. Were looking at about 5,000 of those individuals convicted of drug possession and sentenced to S.B. 123 between implementation in November of 03 through November of 08. And we are looking at outcomes of a comparison group that we constructed looking at, also, low drug possessors who were sentenced to prison, community corrections and court services, both prior to implementation of S.B. 123 and then after implementation.

The way we look at recidivism is we measure recidivism by looking at two outcomes: reconviction for a new offense and revocation from supervision. For both of these outcomes A complete report looks at failures that are linked to a new prison stay and a more general measure of failure that does not include a prison stay. For example, you can be convicted of a new offense but be reinstated to community-based supervision. So here were going to be looking only at failure from supervision that are linked to a prison stay because those are the ones that have the most direct impact on the prison system in Kansas. We looked at failures at different thresholds; today I will be talking mostly about failures at 24 months of exposure in the community, but we also have available rates at 12 and 18 months.

So these are the recidivism rates by group. The darker blue shows the recidivism rates at 12 months of risk exposure and the lighter blue at 24 months. So we see that that about a third of drug possessors sentenced to prison fail within 24 months. That's closely followed by community-corrections folks and then by folks sentenced to S.B. 123 — again, this is at 24 months. And just to reiterate, were looking at S.B. 123 illegible drug offenders, so were looking at offenders who have a relatively minor criminal history score, and all of them have been convicted, obviously, of simple drug possession. So this is what were trying to predict: the variation in the rate of failure across these groups.

[For] the first set of analysis, we ran a set of logistic regressions, just creating a set of dummies for participation. So we would just be interested in associating participation in court services, S.B. 123 or prison with that variation and the likelihood of failure. Then we also implemented a set of models looking at the time to failure, trying to predict whether S.B. 123 clients would fail at a faster rate than other drug possessors sentenced to an alternative sanction. One important challenge with this set of models is that they do not control for selection. The fact that drug offenders — low-level drug offenders — sentenced to prison may still be different from drug offenders sentenced to court services or drug offenders sentenced to a community-based program, even though we are looking at low-level drug possessors. So in order to control for that selection process, in order to model that selection process, we implemented a relatively new approach that seeks to reproduce follows a quasi-experimental design. So were trying to reproduce the likelihood of individual offenders who received an S.B. 123 offense an S.B. 123 sentence. By doing that, we will be sort of matching offenders who eventually went through S.B. 123 with those who went through an alternative sanction and then looking at their outcomes. So its a procedure by which you realign the sample in order to make the elements in that sample be at least statistically equivalent to one another. I will be talking about that in a second.

This table presents the summary results for our logistic regression that is our analysis looking at unmatched samples. Now the coefficients in this table represent the likelihood of failure associated with each of the alternative sanctions compared to S.B. 123. We also split the results by cohort, thinking that during the first two years of program implementation, we would not be able to observe the full impact of the program, so we split the sample into two different cohorts: the first cohort of S.B. 123 offenders — those sentenced to S.B. 123 between November of 03 and November of 05 (so giving the program two years to realign treatment and supervision practices) — and then after those two years, from 2005 to 2008 is our second S.B. 123 cohort. So again, the bars represent the change in the likelihood of failing, and here were looking at reconviction and revocation.

For the first S.B. 123 cohort, participation in S.B. 123 is related to a negative change in the likelihood of failing when S.B. 123 is compared to community corrections. So offenders going through S.B. 123 are 36 percent less likely to fail than those drug offenders sentenced to community corrections. Now compared to court services we find a significant effect by which participation in S.B. 123 for that first cohort increases the likelihood of failure by about a third — so 36 percent. When looking again at this second cohort — so allowing the program to align services, allowing the staff of the Kansas Sentencing Commission and the DOC to disseminate practices associated with the new program — we see that there is no significant relationship between ... no significant difference between those who participated in S.B. 123 — that second cohort — and those going through court services. So we find a significant effect of S.B. 123 compared to community corrections and compared to prison. And by that we mean that offenders going through S.B. 123 are less likely to fail than those going through community corrections and prison — again, more specifically looking for that second cohort. This is looking at logistic regression models with unmatched samples; so at this point were not controlling for the selection process in the sample. But we are just looking at drug offenders sentenced of single drug possession with low-level criminal history scores.

Our main interest was to look at the effect of program participation — whether offenders went through S.B. 123 performed better than offenders going through a different community-based or prison-based program. We also had a series of covariants of recidivism. Some of these have been identified in the literature, also associated with recidivism and community supervision; a couple of them were interesting. First, the importance of context [on] this relationship between the county of conviction and the likelihood of offenders failing while in the community. So, on the one hand offenders who were residing in more urban locations were more likely to fail, according to these models. Sort of an interesting finding because in urban areas you would think there would be a greater number of treatment providers, but also there is more supervision; so its important to think that a number of patterns that we've detected here are linked mostly to revocation procedures that are linked to failures, rather than new offending.

Now the other important pattern that is crucial is the performance of the offenders that we aggregated in that second cohort — so those offenders participating in S.B. 123 starting in November of 2005 — were performing better in terms of recidivism rate compared to S.B. 123 offenders who went through the program earlier.

So here we ... As I said earlier, one problem with the analysis; the problem was we didn't control for the selection process — the fact that not all simple drug possessors with low[-level] criminal history had the same chance of being selected into S.B. 123, although the program is supposed to be mandatory. Some of them have a greater likelihood of being sentenced to S.B. 123. So we first tried to model that selection process in order to reconfigure the sample to make the cases be similar to one another. In order to reconfigure the sample we fit a product model that first estimates the likelihood of receiving an S.B. 123 sentence, and then with that matched sample, we estimate the recidivism rates and then compare those recidivism rates and the match sample to assess the treatment effect — in this case, the effect of participating in S.B. 123 compared to court services and community corrections.

What we found is that S.B. 123 increases the likelihood of recidivism at 12 and 24 months compared to court services — and again here were matching samples, I can get into the details on how we did that: We tried different matching algorithms, and the results were robust to the matching procedure, which is something ... Its kind of a weak point of the propensity score matching is that findings are usually sensitive to the different procedures. So we tried different algorithms, and we found basically the same results. We also found that there was no impact of S.B. 123 on recidivism rates compared to community corrections. So individuals going through S.B. 123 did not [do] better than equivalent offenders going through community corrections. And here's a table summarizing the results. The 24-month follow-up for the first two rows for court services, the last two rows for community corrections. So you will notice sort of the unmatched sample is the raw sample the way we get it. You see the recidivism rate: so 22.7 percent for S.B. 123, 16 percent for court services in the unmatched sample. Once we realign those cases in order to make them equivalent, we reduce our sample to about 660 cases and then we compare recidivism outcomes for both offenders going through S.B. 123 and offenders going through either court services [or] community corrections. So this table shows that individuals going through court services — all the things being equal — have a lower rate of recidivism compared to those going through S.B. 123. Well take questions later but ...

Audience Member 1: Yes, but its a question of clarity. Can you just define, again, what is recidivism?

Rengifo: Recidivism — we use two different measures — these tables here combine reconviction and revocation.

Audience Member 1: [Inaudible.]

Rengifo: Leading to a prison stay for any offense, yes. So reconviction for a new offense and revocation for a technical violation, revocation due to a new offense we counted as a reconviction, but here were combining them. But again, most of these are operating through the revocation failure, and we can talk about it in the Q and A.

Results at 12 months were relatively similar, although the path of the difference in the coefficient for the average treatment effect was larger when looking at 24 months. So we also looked at time to failure and tried to estimate the speed at which offenders going through S.B. 123 and going through other sentencing alternatives were failing in the community. And here, the models were presenting. Twenty-four-month threshold: we found that participation in S.B. 123 delays time to failure relative to community corrections and relative to court services — I'm sorry, related to prison — but accelerates that time to failure when compared to court services. So offenders going through court services take longer time to fail using unmatched samples but also fail less than those going through S.B. 123, when we looked at the propensity score matching and the samples of equivalent offenders.

This is the survival curve. Here well look at the fraction of offenders who remained free in the community by ... The horizontal axis has the time to failure in days, and we see that ... You don't see really the colors ... The blue line, sort of the second line from top to bottom, is the line for S.B. 123 offenders. So the first line — the green line — is court services who had greater fraction of individuals remain in the community — survive in the community to use the terminology that linked to this methodology — than offenders in the other three groups. So again, offenders going through S.B. 123 fell faster ... a greater fraction of them failed faster than those going through court services.

So what are our implications of the findings at the offender level? S.B. 123 has a slightly greater impact of revocations than reconvictions. When we look at unmatched samples — not controlling for the selection process — we find a positive effect of S.B. 123. The problem is once we control for the selection process, then impact of S.B. 123 is really produced by differences in the characteristics of the sample and in the attributes of offenders going through court services versus the attributes of offenders going through other alternatives.

We found, again, that later cohorts of S.B. 123 do better than earlier cohorts, which is important from an implementation point of view, and its important in terms of the timing of additional reforms to S.B. 123. And, again, we found that compared to standard community corrections and prison, offenders going through S.B. 123 do better than offenders going through those programs; but they don't do better than those going through court services, mostly because individuals in court services are not revoked as much. And when they are revoked, they are moved from court services to community corrections, so there is an increased level of supervision but not necessarily they are going straight to prison from court services — so that's an important difference to keep in mind.

So what we learned in terms of the system level impacts — as Don had mentioned earlier — is that S.B. 123 has effectively reduced the use of prison in different ways. First, the most direct and straightforward impact is by reducing the drug possessors sentenced to prison. This is a mandatory program; again, we have more than 70 percent of simple drug possessors going now to S.B. 123 instead of going to prison. But also the use of prison has been averted by a reduction in the number of people who received that S.B. 123 sentence and were then revoked and sent back to the community. Because we have fewer failures in the community, then we have decreased use in the use of prison. Mostly through revocations, so we see fewer technical violations being handled in terms of sending the offender back to prison and also through reconvictions, although in a smaller number.

Thinking about the overall cost of housing offenders in state prisons and working with the staff of the Sentencing Commission, we estimated the cost avoidance of S.B. 123 to about $30 million, and we can, again, go over the details of that analysis in the Q and A. But again we have fewer bodies in prison linked to simple drug possession, again, not only at the sentencing stage but also in terms of the handling of technical violations. From our process evaluation where we spend a significant [amount of] time in Kansas interviewing community corrections officers and interviewing providers, both providers and officers indicated that they were comfortable and they were addressing the needs of offenders by having team meetings, by spending more time with the offenders. We see some of that not only because they are telling us that that's what's going on but also in terms of the administrative data recorded on the case management information system.

Also, in terms of how S.B. 123 connects to a broader set of corrections reforms that are taking place in Kansas — and I think, Roger, it would be great to hear a little bit about that it in your follow-up intervention — in terms of how, for example, the handling of revocations has something to do with S.B. 123 but also has something to do with the handling of revocations at large in the Department of Corrections, and how probation officers, but also parole officers, are now trained to look at failures and to look at supervision processes from a different point of view by using the LSIR, by working closely with community-based organizations. So its not only about S.B. 123 but its also about a broader initiative in Kansas on risk reduction and public safety.

We see some net widening taking place in terms of individuals who were sentenced initially to low levels of supervision and low levels or no treatment at all under court services that are now being funneled through S.B. 123 with higher levels of supervision and higher levels of treatment. So by that we have a greater share of the populations under state supervision; some of that linked to greater levels of failure, again, linked to higher revocation rates. And as Roger pointed out, there is more information on the Kansas Sentencing Commission Web site, and we have a preliminary report posted on the Web site. Please do not hesitate to contact us for additional information. Thank you.

[Applause.]

Question and Answer Session

Linda Truitt: We have about 15 minutes left in the session. We are recording today, and so I'm going to ask that if you have a question, if you'd raise your hand, state your name and your affiliation, and then as briefly and straightforward as you can, pose the question and to whom you'd like to answer that question. Any questions, please? Mike Finnigan?

Mike Finnigan: [Inaudible.]

Andres Rengifo: So the question is about sort of the intensity of supervision and whether the ... under S.B. 123 we see greater levels of accountability for offenders than standard community corrections or court services. And we think that the answer is yes. So we ... so the program was housed in community corrections, so that meant that already it was sort of a more intense level of supervision. Under court services, some people argue, there is no supervision — you're just pushing paper around — and no treatment. So if that's the baseline, then its sort of easy to move up from that. Looking at interventions — intervention data recorded on the case management system — we see that for individuals going through S.B. 123, there are more interventions associated to supervision practices than we have on the court services side. A big problem is that the data on the court services side — because its not centralized — there is very little of it. So LSIR scores, program implementation, program participation, supervision interventions — we don't have almost [any]thing on the court services, so our ability to see what's happening with that population is restricted. But from everything we have at the moment we think that yes, offenders are more accountable on the supervision side on S.B. 123 than on court services.

Roger Werholtz: Let me add one other thing and that is, we've debated this a little bit over breakfast this morning. There are no risk assessments done at court services, so we don't know the risk levels of those individuals. What we do know is when the group shifted initially from court services to community corrections in that first cohort, we had seriously underestimated the level of high-risk offenders in that group. The other thing that I have no idea how this plays into some of the research results, but one of the things that we know is that the defense bar is ...

[Break in recording.]

Rengifo: ... the textbook threshold for significance. So what ... so we did two things. We adjusted the caliber in order to make those cases look like one another more closely, right? So by adjusting the caliber, were being more demanding on the quality of the matching, but then we lose initial representativeness. But by reducing the sample size in such a way that the offenders in both groups are similar to one another but not necessarily similar to the groups they're supposed to represent — so we did that. And in one or two matching algorithms, that significant relationship went away. For the ones that were very demanding on the matching procedure, which I think was an important part for us to make sure we were getting rid of that selection effect, the findings were robust in terms of trying with different algorithms to make those ... to generate the match. We tried nearest neighbor; we tried kernel based; we tried with radius caliber. As part of our report then in the appendix, we list all these tables — the table with all the different matching algorithms. But yes, it is sensitive to these decisions that have to do with a trade off to how representative you want your results to be and to what extent you highlight more the fact that these have to be similar offenders on the observables. And that's a separate question, in terms of how many observables do we have. So we may be matching on age, criminal history, gender, but still we only have six or seven or eight administratively recorded variables, so that's also another limitation. Thank you.

Truitt: Other questions? Yes, please.

Audience Member 1: [Inaudible. Recording includes loud static.] ... Foundation of Baltimore. I was curious, when you did your cost estimate, did you use a variable rate for the prison stay? You know, a lot of times they say, If you don't close a prison wing then you really cant say that you're saving X amount of dollars ... [inaudible].

Werholtz: I'm not the researcher in the room; let me make that really clear. But I talk with the legislature about this a lot, and the way that we couch this is in cost avoidance. At the time that the bill was passed, we were looking at prison construction and we could honestly and directly talk about the beds that we did not have to build and operate and could create an average cost for them. We had a discussion very recently, now that what's happened in my state is prison populations decreased by 7 percent. Well, we don't save $25,000 a year for every person that we keep out of prison; now we save 2,000, and the cost of treatment is about 3,000. And the critics of the legislation want to make that comparison saying, Well, we don't need to pay for all of this expensive treatment because we've got cheaper prison space that's available. Now we took that argument away when I closed four prisons in the last year so now I don't have surplus beds again. But that has been probably the crux of the debate around S.B. 123 at this point. My personal opinion is its pretty well institutionalized in our state with the majority of the legislature and with the courts; the judges like it and now want to see it expanded. So, you know, if you want to talk about cost reduction, its probably not cost effective, but at the time that it was passed, I believe it was an effective cost avoidance bill for the state, and its what historical context you look at it in. Does that make sense?

Truitt: I think we have time for one or two more questions. Anything? Yes, sir.

Audience Member 2: [Inaudible. Recording includes loud static.]

Werholtz: Let me talk about it from kind of the political or the small P point of view, and then Don and Andres may want to talk about it from the research point of view. I think that the question that you pose is perfectly logical. You know, the crime of conviction doesn't necessarily relate at all to the responsivity to treatment. But the overarching political issue is the degree of seriousness that we've attached to the offense. Coming from the Midwest, you know the issues around meth, and meth is the boogie man in our state, and so we have disproportionate sentences for convictions around possession and manufacturing a methamphetamine, as compared to other sentences. That's still a political discussion that hasn't been overcome in terms of an alternative ... treatment as an alternative to prison. I think there's widespread support as treatment in prison. The other thing that I would say that we — my personal opinion — that we have not done a good job on in Kansas at this point around this legislation is we have not held the treatment providers accountable for their performance. We hold the community corrections programs very accountable for their performance. We track and publish their revocation rates. We've invested a great deal of time in training them on how to supervise, but we have done next to nothing in terms of publishing the performance of various treatment providers. And in part that's because some of our community corrections programs are very concerned about alienating the only game in town. I think that's something that we need to tackle because you could argue if you're not getting much change in offender behavior out of the treatment intervention that you're providing, what's the value that you're receiving for your investment? Obviously were saving some prison beds, but were in a heavy cut back mode in our state right now, and this may be a really good opportunity to start singling out treatment providers that aren't working well with offender populations. I don't know if that answers your question, but that's what it made me think of.