Working at the social-clinical-community-criminology interface: The GMU Inmate Study

Abstract

This paper describes our attempt to import social-personality theory and research on moral emotions and moral cognitions to applied problems of crime, substance abuse, and HIV risk behavior. Thus far, in an inmate sample, we have evidence that criminogenic beliefs and proneness to guilt are each predictive of re-offense after release from jail. In addition, we have evidence that jail programs and services may reduce criminogenic beliefs and enhance adaptive feelings of guilt. As our sample size increases, our next step is to test the full mediational model, examining the degree to which programs and services impact post-release desistance via their effect on moral emotions and cognitions. In addition to highlighting some of the key findings from our longitudinal study of jail inmates over the period of incarceration and post-release, we describe the origins and development of this interdisciplinary project, highlighting the challenges and rewards of such endeavors.

Criminal recidivism is a serious social problem. Countless tax dollars sustain the many activities of the criminal justice system as it polices, prosecutes, and punishes repeat offenders. Although numerous academic disciplines and subdisciplines pursue questions relevant to recidivism, historically there has been relatively little cross-fertilization of ideas and a surprising provincialism in terms of foci. For instance, social psychologists might focus on the socialization of inmates' crime-relevant beliefs while incarcerated, whereas a clinician might address the challenges of providing mental health services to inmates. A sociologist might elucidate the role of poverty in the commission of crime, whereas a criminologist might focus on best practices in tracking parolees. Although these varied areas of interest are understandable given the types of questions and methodologies that drive each discipline, we argue that great gains are afforded when one deliberately considers the intersections among these areas.

We describe here our attempt to import social-personality theory and research on moral emotions and cognitions to applied problems of crime, substance abuse, and HIV risk behavior. Specifically, we seek to understand how to more effectively intervene with incarcerated adults to foster post-release desistance from crime and other risky behaviors. Moreover, we seek to strengthen our theoretical frameworks by considering and testing theories in real world contexts.

Criminal Recidivism in the U.S.: Magnitude of the problem and a history of despair

Without question, this is an area where the stakes are extremely high. Crime remains one of our nation's leading problems. Criminal activity costs America $105 billion annually – and this figure only captures the monetary dimension of cost (Miller, Cohen, and Wierama, 1996). Each one of the 16,204 murders, the 95,136 rapes, the 894,348 assaults, and the 2,151,875 burglaries in 2002 left indelible marks on the lives of the victims and on the network of relationships around them (Bureau of Justice Statistics, 2003b). As many victims attest, the psychological costs of crime are often far higher than the steep economic costs.

The cost of crime does not end there. Americans end up paying a second time, in those cases where offenders are apprehended, convicted and sentenced to serve time. The United States incarcerates a larger percentage of our population than any other country in the world (International Centre for Prison Studies, n.d.). In 2003, 2.1 million Americans were behind bars (Bureau of Justice Statistics, 2003a) -- a 73% increase in one decade (Bureau of Justice Statistics, 1996). It costs 2.5 times more to send someone to jail than to college for a year (Western Interstate Commission for Higher Education, 2002). In FY 2001, the average cost of housing an inmate in a U.S. prison was $22,650 per year (Bureau of Justice Statistics, 2004c). Housing in local jails costs upwards of $29,000 per year, due to high costs in maintaining smaller facilities.

Upon release, inmates have a two in three chance of being re-incarcerated, either through re-offense or a violation of probation or parole. Of U.S. prisoners released in 1994, 67.5 % were re-incarcerated within three years. Moreover, the rate of re-incarceration in the U.S. is steadily on the rise (Bureau of Justice Statistics, 2004b).

As the dollar cost mounts, so too does the human cost. More than half of all incarcerated Americans are parents of minors, resulting in 2.3 million children affected by parental imprisonment (an increase of 500,000 since 1991; Bureau of Justice Statistics, 2000). Compared to their peers, those 2.3 million children of prisoners are up to six times more likely to become incarcerated themselves, with 1 in 10 being incarcerated before even reaching adulthood (Kleiner, 2002; Brenner, 1998).

Until recently, the prospects for successfully intervening with criminal offenders looked grim. Beginning in the 1970's, a sense of hopelessness descended on criminologists and policy makers, as they concluded that “nothing works” with incarcerated offenders (Martinson, 1974). This conclusion came as the result of several influential reviews of treatment studies showing no measurable impact on offenders. Funding for programs and interventions in correctional settings plummeted. Today, we are in a very different place. Further reviews have shown that many of those non-significant treatment studies had low power, substandard designs, or both. Criminologists, sociologists and psychologists have generated a wealth of new data using state-of-the-art methods and empirically informed treatments. Research consistently shows that well-conceived treatment does work with offenders. (In fact, punishment -- boot camps, longer sentences, harsher conditions -- is just about the only thing that consistently does not work)

Less clear is why these programs work. Most evaluation studies focus exclusively on outcome measures such as rates of re-arrest and re-incarceration. Little research has directly evaluated possible “mechanisms of action.” Evaluation of these hypothesized mediators of program effectiveness would greatly increase our understanding of how and why certain interventions work. Most important, such information would aid in designing more efficacious treatments that have maximum impact on those mechanisms of action.

How can social-clinical psychology inform the fight against crime?

To date, the vast majority of research on adult offenders has been conducted by criminologists and sociologists whose fields emphasize macro-level sources of influence (e.g., prisonization, social consequences of racial discrimination, social control inherent in the fabric of society). Far less attention has been paid to internal psychological factors, such as moral emotions and moral cognitions, and to individual differences in people's experiences of their relations to significant others and to the community. Yet, it is precisely these sorts of psychological factors that likely serve as “mechanisms of action.”

Why Consider Shame and guilt?

Emotions form the core of our motivational system. They provide the press – the driving force – and the direction for our behavior. In the context of crime and recidivism, experiences of shame and guilt are particularly germane. These moral emotions are presumed to play a key role in deterring immoral and antisocial behavior (Ausubel, 1955; Damon, 1988; Eisenberg, 1986), while also fostering corrective change following a transgression.

It is surprising, then, that research on criminality and recidivism has devoted little attention to these “moral” emotions, even though an absence of guilt or remorse is cited as one of the hallmarks of psychopathy (Hare, 1991; Rogers & Bagby, 1994; Samenow, 1984, 1989). Braithwaite's (1989, 2000) influential work on restorative justice and “reintegrative shaming” provides strong a theoretical argument for considering these emotions in work with offender populations. (Braithwaite distinguishes between “reintegrative” shame and “disintegrative” shame, which in many ways paralleling psychologists' distinction between guilt and shame, respectively.) However, with few exceptions (Harris, 2003) ¹, if shame and guilt appear at all in the empirical literature on criminal behavior and recidivism, the consideration is brief or superficial, or both. The current study systematically examined offenders' capacity for shame and guilt and the implications of these emotions for recidivism.

What is the Difference Between Shame and Guilt?

Whereas scientists and laypersons alike often use the terms “shame” and “guilt” interchangeably, a decade of research with non-offender samples indicates that these are distinct emotions (Lewis, 1971; Lindsay-Hartz, 1984; Niedenthal, Tangney & Gavanski., 1996; Tangney, 1993; Tangney & Dearing, 2002; Weiner, 1985; Wicker, Payne & Morgan, 1983) with very different implications for recidivism and reform. Feelings of shame involve a negative evaluation of the global self; feelings of guilt involve a negative evaluation of a specific behavior. Although subtle, this differential emphasis on self (“I did that horrible thing”) versus. behavior (“I did that horrible thing”) sets the stage for very different emotional experiences and very different patterns of motivations and subsequent behavior.

Shame is an acutely painful emotion that is typically accompanied by a sense of shrinking or of “being small,” and by a sense of worthlessness and powerlessness. Not surprisingly, shame often leads to a desire to escape or to hide -- to sink into the floor and disappear. In contrast, guilt is typically less painful and threatening because the primary concern is with a specific behavior, not the self. Guilt is typified by a sense of tension, remorse and regret over the “bad thing done.” Rather than motivating an avoidance response, guilt motivates confession and repair. Simply stated, shame motivates a desire to “duck the heat.” Guilt motivates a desire to “face the music.”

What are Criminogenic Beliefs?

Criminals who persist in a life of crime often hold a distinct set of beliefs – (im)moral cognitions -- that serve to rationalize and perpetuate criminal activity. Our project was enriched early on by collaboration with clinicians who have extensive experience working with offenders at the Fairfax County ADC. In focus group sessions, the clinicians identified key beliefs and cognitive distortions that they aim to address in treatment with repeat offenders. For example, it is not unusual for inmates to make external attributions for the cause of their conviction. More than a few offenders genuinely perceive that the primary reason they are in jail is an overzealous cop, an associate's betrayal, or society's failure to provide adequate employment opportunities. Another common cognitive distortion among inmates centers on the experiences of a victim. Many offenders view a broad range of crimes as “victimless.” They may believe that a victim (e.g., of burglary, fraud, even rape) is not really harmed unless there is concrete physical injury. They may be oblivious to the reality of psychological pain.

Based on the insights of clinical caseworkers at the frontlines of rehabilitation,² we developed the Criminogenic Beliefs and Assumptions Scale (CBAS) to assess the presence and magnitude of such beliefs. The CBAS is a 25 item self-report measure designed to tap 5 dimensions: (a) Failure to Accept Responsibility (“Bad childhood experiences are partly to blame for my current situation”); (b) Notions of entitlement (“When I want something, I expect people to deliver”); (c) Negative attitudes toward authority (“People in positions of authority generally take advantage of others”); (d) Short-term orientation (“The future is unpredictable and there is no point planning for it”); and (d) Insensitivity to impact of crime (“A theft is all right as long as the victim is not physically injured”). Several dimensions identified by the clinicians appear in previous efforts to conceptualize cognitions associated with criminal activity (Barriga, et al., 2000; Gendreau, Grant, Leipciger, & Collins, 1979; Shields & Simourd, 1991; Walters, 1995; Yochelson & Samenow, 1976). The CBAS, however, is unique in its incorporation of restorative justice theory, most clearly exemplified by the insensitivity to the impact of crime dimension.

Moral Emotions and Moral Cognitions as Mechanisms of Action

Moral emotions and moral cognitions are especially promising avenues for intervention for two reasons. First, moral emotions are dynamic, as opposed to static, factors. The majority of documented predictors of recidivism (Andrews & Bonta, 1994; Blackburn, 1993; Gendreau, Little & Goggin, 1996; Harris, Rice & Quinsey, 1993; Zamble & Quinsey, 1997) represent “water under the bridge” – background factors rooted in past history (unstable family life, elementary school adjustment, age of first arrest etc.) or enduring aspects of the person (intelligence, etc.). These factors may suggest avenues of broad and difficult social change that may benefit future generations. But, as Zamble and Quinsey (1997) have observed, such static or “tombstone” factors do not provide points of intervention for the 2.1 million inmates currently in US prisons and jails, nor for the millions of Americans who will be newly incarcerated in the next 10 years. Their history is already written.

Second, moral emotions and cognitions are malleable factors that should be amenable to intervention. As underscored in the literature on anxiety and depression, a number of empirically-supported interventions are highly effective in targeting and modifying disorder-specific cognitions and emotion. Techniques from social-cognitive, cognitive-behavioral, and interpersonal therapies can be readily modified to address the moral cognitive and emotional factors discussed here. In short, these are precisely the factors that may function as “mechanisms of action” and explain how and why extant treatments reduce recidivism.

Integrating Social Psychological Theory on Moral Emotions with Restorative Justice Theory

Prior to embarking on this project, the first author's work on shame and guilt had been conducted exclusively on community (non-incarcerated) samples. How did we get here? It was the result of the sharp thinking of an undergraduate research assistant, Conrad Loprete, who was volunteering with restorative justice programs at the jail at the same time that he was assisting us at the university with a study of shame and guilt. Conrad remarked that, although we did not use the same words, we and the clinicians at the jail were dealing with many of the same concepts. Conrad set up an “intellectual blind date” with the director of OAR, Fairfax, David Manning.

We soon learned that although not explicitly addressed in criminal justice theory, the distinction between shame and guilt is consistent with restorative justice approaches to rehabilitation.³ Restorative justice is a philosophical framework proposed as an alternative to traditional, punishment-focused ways of thinking about crime and criminal justice. Restorative justice theory emphasizes the ways in which crime harms relationships in the community. Crime is viewed as a violation of the victim and the community rather than a violation of the state. Accountability is defined in terms of taking responsibility for actions and taking action to repair harm caused to the victim and community. In effect, restoration, or making things right, replaces the imposition of punishment for its own sake as the highest priority of the system. Most intriguing to us, restorative justice is essentially a “guilt-inducing, shame reducing” approach to rehabilitation. Offenders are encouraged to take responsibility for their behavior, acknowledge the negative consequences to others, empathize with the distress of their victims, feel guilt for having done the wrong thing, and act on the consequent press to repair. But they are actively discouraged from feeling shame about themselves. The resulting inmate study, jointly conceived by a clinical-social psychologist and an expert in corrections, seemed the most logical extension of our work on moral emotions.

The GMU Inmate Study

The GMU Inmate Project is a multi-phase longitudinal study of inmates, focusing on factors that may reduce criminal recidivism, substance abuse relapse, and HIV risk behavior, and more generally enhance the post-release adjustment of criminal offenders. Because a key interest of this project was the effectiveness of short-term interventions with relatively serious offenders, selection criteria were developed to identify incoming inmates likely to serve at least four months (i.e., long enough to complete the 4-6 session baseline assessment and to have the opportunity to request and engage in at least some jail programs and services.⁴

Shortly after assignment to the general population, eligible inmates were presented with a description of the study and asked to participate. All participants were over 18 years of age and were assured of the voluntary and confidential nature of the project. In particular, it was emphasized that the decision to participate or not would have no bearing on their status at the jail or their release date. Regarding confidentiality, interviews were conducted in the privacy of professional visiting rooms or secure classrooms. In addition, data are protected by a Certificate of Confidentiality from the Department of Health and Human Services.

The focus of the initial assessment is to collect historical information about the participant (e.g., pre-incarceration levels of HIV risk, drug abuse history, familial background, etc.) and to also evaluate baseline levels of key constructs, including moral emotions, criminogenic beliefs, etc. Over the course of incarceration, we collect jail records to monitor the individual's use of services and disciplinary infractions. Then, just prior to release from jail, either to the community or to another facility, we re-assess key constructs and evaluate post-release plans. One year after release, we contact participants again, obtaining reports of both detected and undetected crime, drug use and abuse, HIV risk behaviors, moral emotions, cognitions, and other measures of interest. We also evaluate the presence of positive indicators of post-release adjustment (e.g., employment, living situation, relationship status, payment of child support, community service).

Initial Findings on Inmates' Moral Emotions

Although a primary focus of this study is the implication of shame and guilt for post-release adjustment several basic questions needed to be addressed from the outset. First, can we measure shame and guilt reliably in an adult inmate population? Second, is there substantial variation in offenders' capacity to experience these moral emotions (vs. the argument that offenders are simply deficient in this regard)? Third, if there is variation, do previous findings regarding shame and guilt replicate with an incarcerated sample? The answer to each question seems to be yes!

We assessed shame-proneness and guilt-proneness with the Test of Self-Conscious Affect for Socially Deviant Populations (TOSCA-SD; Hanson & Tangney, 1996). This version of the TOSCA was designed specifically for use with incarcerated individuals and other “socially deviant” populations. As in preliminary studies of prison inmates (Hanson, 1996), results from the current study indicate that shame and guilt can be measured reliably and that inmates show substantial variability in their propensity to experience these moral emotions. Moreover, concurrent psychological and behavioral correlates indicate that inmates' shame and guilt-proneness as measured by the TOSCA-SD serves many of the same functions as shame and guilt-proneness measured by the TOSCA in community samples. For example, researchers have found that in non-incarcerated, community samples of children, adolescents, and adults, shame-proneness is associated with a variety of measures of psychopathology including PTSD, obsessive-compulsiveness, psychoticism, anxiety, and depression (Andrews, Brewin, Rose, & Kirk, 2000; Ferguson, Stegge, Eyre, Vollmer, & Ashbaker, 2000; Harder, Cutler, & Rockart, 1992; Tangney, Wagner & Gramzow, 1992). “Shame-free” guilt-proneness on the other hand shows either no relation or a negative relation to such psychological symptoms (Bybee, Zigler, Berliner & Merisca, 1996; Gramzow & Tangney, 1992; Quiles & Bybee, 1997; Tangney, 1999). We have found similar results in the current study of offenders, with shame-proneness being significantly positively correlated with inmates' anxiety, depression, and low self-esteem. Conversely, guilt-proneness was significantly negatively related to a range of psychological symptoms. In addition, parallel to studies of community samples, inmates' propensity to experience guilt about specific behaviors was positively related to their capacity for empathic concern and other-oriented perspective taking.

Many researchers assume that, although detrimental to a person's own mental health, the profound pain associated with shame is useful in that it motivates people to avoid “doing wrong,” decreasing the likelihood of transgression and impropriety (Barrett, 1995; Ferguson & Stegge, 1995; Kahan, 1997; Zahn-Waxler & Robinson, 1995). As it turns out, results from the few studies that have examined the relation of shame and guilt-proneness to transgressive and risky behavior are quite consistent. Shame-proneness has shown either no relation or a positive relation to aggression, delinquency, and substance use while guilt-proneness has been negatively related to these same behaviors (Dearing, Stuewig, & Tangney, 2005; Ferguson, et. al 1999; Stuewig & McCloskey, in press; Tangney, Wagner, Barlow, Marschall, & Gramzow, 1996; Tibbetts, 1997).

The current study offers the most extensive data yet on the relation of the moral emotions to different forms of risk and transgressive behavior. Baseline measures of inmates' shame-proneness showed no relation to self-reports of physical aggression but was positively correlated with externalization of blame. Guilt-proneness was substantially negatively related to both. Similarly, baseline levels of shame-proneness were positively related to pre-incarceration alcohol problems, drug problems, and polydrug use, and to clinical symptoms of alcohol, marijuana, and cocaine dependency. In contrast, guilt-proneness was negatively related to substance abuse problems and to risky sexual behavior. Importantly, whereas baseline guilt-proneness was negatively related to clinicians' assessments of psychopathy and clinicians' independent assessments of violence risk, shame-proneness was related to neither.

While we have just begun to examine offenders' behavior and adjustment at one year post-release, initial results suggest that the moral emotions may be useful in predicting criminal re-offense. In a sample of about 132 cases, baseline shame (assessed at the beginning of incarceration) did not predict post-release re-arrest, commission of one or more undetected felonies, nor the number of different kinds of undetected felonies (an indicator of criminal versatility). In contrast, guilt appeared to function as a protective factor, in that high guilt scores predicted lower criminal versatility, as measured by number of different kinds of undetected felonies. Taken together, our findings thus far suggest that shame and guilt have the same psychological and behavioral implications among offenders as in the community at large.

Initial Findings on Criminogenic Beliefs

Our data reveal that, as expected, criminogenic beliefs and deviant behavior go hand-in-hand. Not only are criminogenic beliefs strongly linked to concurrent measures of aggression, antisocial personality, and a history of criminal activity, such beliefs are also related to clinicians' ratings of psychopathy and risk for violence. In addition, criminogenic beliefs are negatively correlated with measures of empathy. Taken together the pattern of correlations attests to the construct validity and usefulness of the CBAS.

But the big news is that criminogenic beliefs appear to be prospective predictors of a number of critical outcomes. The magnitude of criminogenic beliefs assessed at the outset of incarceration predicted official reports of inmate misconduct during incarceration. Additionally, criminogenic beliefs significantly predicted post-release offense (recidivism). In fact, using a test for of the difference between dependent correlations, this simple 25-item self-report scale predicted as well as the PCL:SV (Hare, Cox & Hare, 1995), the gold standard for assessing psychopathy. These finding are important in two ways. First, the time and expertise required to conduct and score a PCL:SV interview prohibits its use as a screening measure for incoming inmates. Second, whereas psychopathy is a static construct based substantially on past history, the CBAS measures beliefs and assumptions that are dynamic, presumably amenable to intervention.

Did Criminogenic Beliefs Change During Incarceration?

We next examined how dynamic these beliefs and assumptions actually are. First, we looked at the impact of incarceration on criminogenic beliefs – considering both the degree to which there are mean changes in criminogenic thinking across the period of incarceration, and the degree to which there is individual variability in change during that same time period.

Differential Association Theory strongly predicts that time incarcerated should result in a detrimental “prisonization” effect (Sutherland, 1947). From this perspective, inmates should become increasingly socialized into the criminal world via interactions and experiences with other deviant individuals during their incarceration (Walters, 2003). We were surprised to find no evidence of prisonization in our sample. In fact, the data actually suggest a slight decrease, on average, of criminogenic beliefs over the course of incarceration. This decrease was especially apparent on items assessing negative attitudes toward authority.

These surprising findings are not a function of problematic reliability and validity. The CBAS is reliable (internal consistency was high at both baseline, alpha = .85, and at pre-release, alpha = .83) and the concurrent correlations described above support construct validity. We attribute the modest decrease in criminogenic thinking to the quality of our host jail, an unusually well-run facility, and the large number of programs offered. These findings offer real hope for rehabilitation, but they also remind us to be cautious in generalizing to other correctional facilities. When considering such descriptive means, our results may be more representative of what could be in the world of corrections given progressive leadership and reasonable resources, not necessarily what is happening across the country.

Considering group means, there was a modest decrease in average criminogenic beliefs over the course of incarceration. The correlation between criminogenic beliefs measured at intake vs. those measured just prior to release, however, was moderate (r =.6). In other words, there was considerable individual variability in changes in criminogenic beliefs. Some inmates become more criminogenic in their thinking over the period of incarceration, while others become less so.

We anticipated that certain personality and background factors would moderate the direction of change in criminogenic beliefs over the course of incarceration. In other words, it seemed likely that characteristics brought to the situation by inmates would in part determine who ultimately showed signs of “rehabilitation” versus detrimental prisonization effects. Surprisingly, we were unable to detect moderating effects of any magnitude across a range of individual difference factors including psychopathy, violence risk assessment, age, prior jail experience, baseline shame, and baseline guilt.

Impact of Existing Jail Programs on Moral Emotions and Cognitions

Next in the search for moderators, we conducted preliminary analyses to determine whether involvement in programs and services is related to changes in criminogenic beliefs and moral emotions. They are. In particular, involvement in psychoeducational programs (e.g. impact of crime, anger management, conflict management) and alcohol and drug services (most often 12-step programs such as Alcoholics Anonymous) was associated with reductions in criminogenic beliefs. Moral emotions seemed likewise amenable to intervention. Inmates who were involved in religious programs (e.g.,scripture studies, discipleship classes) and services (e.g., worship services, Friday prayers) showed an increase in adaptive feelings of guilt.

Of course, even with a longitudinal component, the interpretation of correlational data is difficult. Only experiments can examine causality. Experiments are relatively rare in the offender literature, in part because they are expensive and in part because they notoriously difficult to conduct behind prison walls. We are currently involved in evaluating one such program (Impact of Crime Workshop), which focuses on individuals' cognitions.

Summary and Future Directions

Thus far, we have evidence that criminogenic beliefs and proneness to guilt are each predictive of post-release re-offense. In addition, we have evidence that jail programs and services may reduce criminogenic beliefs and enhance adaptive feelings of guilt. As our sample size increases, our next step in this line of inquiry is to test the full mediational model, examining the degree to which programs and services (particularly the IOC workshop) impact post-release desistance via their effect on moral emotions and cognitions.

Initial Findings on Substance Abuse and HIV Risk Behavior

A large proportion of inmates reported frequent substance use and substance use problems in the year prior to incarceration (Dearing, Stuewig, & Tangney, 2005). Shame-proneness was positively correlated with substance use problems, whereas guilt-proneness was inversely related (or unrelated) to substance use problems suggesting that shame and guilt should be considered separately in the prevention and treatment of substance misuse.

On average, participants reported less frequent drug use in the first year post-release than during the year prior to incarceration. Pre-post differences were statistically significant for marijuana, cocaine, opiates, and polydrug use. Frequency of alcohol use did not show a significant decline. Similarly, there was a significant decline in symptoms of dependency for marijuana and cocaine but not for alcohol or opiates. Our next step in this line of inquiry is to evaluate moderators of the substance abuse trajectories of individuals after release from jail. Why are some individuals able to remain abstinent following incarceration while others fall back into this cycle of self-destruction? We hypothesize that participation in jail programs and services will facilitate recovery and post-release abstinence. For example, prescription of psychotropic medication should help reduce psychological symptoms, thereby reducing the need for individuals to self-medicate. Similarly, we anticipate that psychosocial interventions, especially those based on restorative justice principles, will facilitate recovery via a reduction in the propensity to experience shame, in favor of adaptive experiences of guilt.

HIV risk behavior represents another important focus of research and treatment with jail inmates. Research consistently shows high rates of HIV infection among criminal offenders, but far less is known about the profile of risk behaviors responsible for this high infection rate. Results from 421 participants (Kendall, Smith, Quigley & Tangney; 2005) indicate high levels of risky IV drug behaviors and unprotected sexual activity prior to incarceration. We examined the relation of shame, guilt, and pre-incarceration symptoms of alcohol dependence to pre-incarceration HIV risk behaviors (Stuewig, Tangney, Mashek, Forkner & Dearing, in press). Symptoms of alcohol dependence were associated with elevated levels of HIV risk behavior (risky needle use and unprotected sex) prior to incarceration. Guilt-proneness was negatively related to risky sexual behavior. In addition, there was a significant interaction between shame and alcohol. Specifically, among those who were low on symptoms of alcohol dependence, shame-proneness was negatively related to risky sexual behavior.

Our next step in this line of inquiry is to evaluate moderators of HIV risk trajectories. Why do some individuals persist in HIV risky behavior following release from jail, while others do not? We hypothesize that participation in jail programs and services (e.g., psychotropic medication, AA/NA) will result in a reduction of HIV risk behavior via their impact on substance abuse. In addition, we hypothesize that interventions based on restorative justice principles will result in reduced HIV risk behaviors via a reduction in shame-proneness, an increase in community connectedness, and an increase in self-esteem.

Relation of HIV Risk Behaviors to Personality Disorders and Substance Dependence

Both symptoms of Borderline Personality Disorder (BPD) and psychopathy, a more serious form of Antisocial Personality Disorder, were associated with elevated levels of HIV risk behavior prior to incarceration. Analyses indicated that the relation between risky sexual behavior and psychopathy is independent of any co-morbidity with BPD, and vice versa (Tangney, Stuewig, Kendall, Mashek, & Youman, 2005). That is, psychopathy and BPD symptoms each accounted for unique variance in risky sex. Results further suggested that these two personality disorders influence exposure to HIV infection via distinct routes through the more proximal factors of substance dependence and diversity of sexual activity. Specifically, substance dependence fully mediated the link between BPD and risky sex, whereas there remained a direct relationship between psychopathy and risky sex, even when taking into account the elevated levels of substance dependence associated with psychopathy. Subsequent analyses showed that psychopathy leads to exposure to HIV infection primarily due to sheer number of partners.

These results have direct clinical implications. First, in delivering HIV education and prevention programs – both in and outside of correctional settings – it is important to target individuals with significant features of Borderline Personality Disorder (about 30% of incarcerated individuals) and psychopathy (about 18%). Second, these disorders are likely to involve their own unique profile of motivations and concerns, types of risky sex, barriers to condom use, etc. For example, psychopathy appears related to HIV risk because those high on psychopathy frequently place themselves in situations requiring precautions. (Psychopathy uniquely predicted IV drug use and multiple sex partners.) Across many such situations, they do not always take preventative steps (i.e., bleaching needles, using condoms). Features of BPD, on the other hand, did not uniquely predict number of sexual partners in the 6 months prior to incarceration. But when having sex, individuals with BPD are particularly unlikely to engage in safe practices. In other words, the issue for individuals with BPD is some sex in conjunction with especially low rates of protection (i.e., condom use) and high risk partners.

Community Connectedness – Another Mechanism of Action?

Community connectedness – the extent to which one includes the community in the self – is another construct that may serve as a mechanism of action explaining treatment effects. In addition, our data indicate that connectedness to the criminal community is orthogonal to connectedness to the community at large (Mashek, Stuewig, Furukawa, and Tangney, in press). Interestingly, inmates who see themselves as part of both the criminal community and the community at large exhibit elevated levels of psychological distress, a pattern predicted by theories of psychological dissonance.

The next step in this line of inquiry is to evaluate whether connectedness to the criminal community is associated prospectively with “thinking like a criminal” and engaging in drug use and other risk behaviors post-release. This prediction is consistent with the mechanism of “prisonization” (Innes, 1997), whereby individuals entering a correctional setting integrate into the “criminal subculture,” taking on the ethics and standards of that subculture. In contrast, because connectedness to the community at large should be associated with more prosocial attitudes, beliefs, and values, it should predict decreased drug use and risk behavior post-release. Moreover, we hypothesize that dual connectedness to communities with opposing beliefs and values will be associated both concurrently and prospectively with self-medicating drug use and subsequent HIV risk behaviors. “Treatments” such as contact with family and friends and participation in religious services while incarcerated should help to establish and maintain connectedness with the community at large, while also eroding connectedness to the criminal community. Finally, post-release perceived social support and employment should facilitate community connectedness, whereas contact with deviate peers should erode community connectedness and enhance criminal connectedness.

Working at the Interface of Social-Clinical-Community Psychology and Criminology

Practical challenges abound when conducting research inside highly regulated spaces such as jails. Much effort and time was invested in building relationships with individuals and institutions, including the Sheriff's Office, the State Police, and our local district attorneys to gain access to a protected population and their records. Other challenges familiar to those collecting data in the field also arise. The population is a fluid one, while in jail “lock downs” often occur, participants bond out or are released or transferred, when released from jail many participants lead a transient lifestyle moving in and out of different institutions, or often do not want to be found. Measures and procedures often need to be adapted to account not only for this instability but also characteristics of this population (i.e. low literacy rates, stigmatized nature of the population, sensitivity of the data).

Perhaps the most daunting challenge concerns our academic identity. Where do we belong? Although all three authors are psychologist by training, we find that we do not really “fit” with any of APA's 56 divisions. If we worked with delinquent children, we would find a home in developmental psychopathology, but adult offenders just aren't on APA's map. This is not a trivial point. We are disheartened not because we do not belong, but because of special hurdles associated with dissemination of new knowledge. As a case in point, we recently submitted a symposium to an APA division. The symposium, which focused on recidivism, sang the praises of psychological interventions of central interest to that division. Because of participants were inmates, however, the submission was bounced to another division, who similarly felt the topic was a poor fit for their program.

Our experience with the APA conference is likely a reflection of a more general problem. Where does one publish research that so clearly spans multiple sub-disciplines of psychology, much less multiple disciplines of the social sciences? Our experience to date has been that social psychology journals refer us to community psychology journals because of the potential for our results to be applied in this domain. The community psychology journals refer us to corrections journals based on the nature of our sample rather than the relevance of the ideas. And corrections or criminology journals have referred us to psychological journals because of the psychological nature of our primary constructs. And of course, the same goes for the selection of conferences. In fact, we have found that researchers in “our” area are tucked away in different disciplines, attending different conferences. Rarely is there a “critical mass” of like-minded colleagues in one place, at one time, to share ideas and findings.

Nonetheless, the rewards of working at the interface of criminology and social, clinical, and community psychology far outweigh the challenges. We have been gratified by the level of interest and enthusiasm expressed by many practitioners working on the front lines with the chronically underserved population of adult offenders. Many administrators and clinicians are hungry for new insights based on theoretically sound psychological principles and empirically supported strategies. To date, researchers have enthusiastically embraced the promise of intervening with children and adolescents, but adult offenders remain largely abandoned, harkening back to the ingrained belief that when it comes to rehabilitation, “nothing works” post-adolescence. This is an ideal context in which to heed Zimbardo's charge of “giving psychology away” for the public good.

Beyond potentially benefiting the very population we study, our research program has likewise profited from the insights of clinicians, administrators, and participants at the jail. Instead of merely theorizing about the role of moral emotions and moral cognitions in the realm of moral behavior, we have been forced to step down from the ivory tower to dig into the complexities of both qualitative and quantitative data. Our theorizing, of course, has also benefited from the “reality check” afforded the many other researchers who conduct field-based studies – our theories are fortified when their external validity can be documented.

The benefits described above represent only a few examples of the many advantages of community-based participatory research (e.g., Agency for Healthcare Research and Quality, 2003). Evaluation and intervention should be an iterative process between clinicians and researchers. In addition to contributing to the development of the CBAS, jail clinicians informed other key aspects of the study design. For example, the clinicians emphasized the importance of evaluating participants both before transfer to other facilities and before release into the community. In return, our basic research on the moral emotions was utilized in helping shape the inmate workshops. In each case, our respective endeavors have been strengthened and enriched by a shared partnership – a partnership that we hope will continue for years to come. We are eager to see other such partnerships develop across the fields of clinical, social, community and correctional psychology.