Abstract
Linked article:
2. Background for the Review
The importance of interventions for serious juvenile offenders cannot be overstated as this group pose a significant challenge to criminal justice agencies both in terms of frequency and seriousness of their offending and their later behavior as adults. Authorities are increasingly incarcerating these young people, however, doubt remains over the effectiveness of such an approach.
In this review “serious” includes violent and chronic (persistent) offenders, and “juvenile” or “delinquent” refers to young people aged 12 to 21 years.
This systematic review does not include community-based interventions as these are reported in related systematic reviews that include interventions in non-secure corrections in juveniles “Community-based alternatives to prison”, (Martin Killias), “Juvenile aftercare programs”, (Ken Adams) and “Re-entry programs for offenders”, (Christy Visher).
Relevance of reviewing this subcategory of offenders
In 1995, Thornberry, Huizinga and Loeber reported results from the Program of Research on the Causes and Correlates of Delinquency, which consists of three well co-ordinated longitudinal research projects: the Denver Youth Survey, the Pittsburgh Youth Study and the Roches ter Youth Development Study. In total these three projects involved 4,500 inner-city youths, ranging in age, at the beginning of data collection, from 7 to 15 years old.
Chronic violent offenders constituted only 15% of the total sample in Rochester and 14% of the adolescent sample in Denver, however, they committed 75% of all the violent offenses reported in Rochester and 82% of all the violent offenses reported in Denver. Data from the Rochester and Denver studies indicated the criminal versatility of these violent offenders i.e. they commit a wide array of other offenses including property crimes, public disorder, status offenses and drug sales. In conclusion the authors stated “If we do not successfully reach this small group, we will leave the vast majority of the violence problem untouched” (p. 220).
Similar conclusions can be drawn from many other studies i.e. that those juveniles responsible for violent offenses are at high risk of becoming chronic offenders, committing many types of offense and to receive an institutional sentence. For example, in the classic Cambridge (UK) study, which included a traditional white, urban and working class sample of 400 subjects, 40% of the males received convictions before the age of 40 (Farrington in press, a). The prevalence of offending increased up to age 17 and then decreased, but the mean age was 21, showing the skewness of the age-crime curve. The peak age of increase in the prevalence of offending was 14, while the peak age of decrease was 23. These times of maximum acceleration and deceleration are hypothesised to be times when important social influences are at work; at 14 male peers take precedence over parents and from male peers to female partners around age 23. Clearly, in a residential setting, the deceleration influence can hardly act, while the modelling on antisocial peers is present.
There is significant continuity between offending in one age range and offending in another. For example, 73% of males convicted in the Cambridge study as juveniles between the ages of 10 to 16 were reconvicted between ages 17 and 24, in comparison with only 16% of those not convicted as juveniles. Nearly half (45%) of the juvenile offenders were reconvicted between ages 25 and 32, in comparison with only 8% of those not convicted as juveniles (also studies of Krohn et al., 2001, and Stattin & Magnusson, 1991, as quoted by Farrington, in press, a). Effective interventions with juveniles should therefore affect later offending rates in adulthood.
As described above, violent juveniles are criminally versatile; 55 of the 65 males with a conviction for violence also received a conviction for a non-violent crime. To a large extent, the frequent offenders were versatile and sooner or later committed a violent offense. The probability of committing a violent offense increased steadily with the number of offenses committed, from 18% of one-time offenders to 82% of those with 12 or more convictions (Farrington in press, a).
Those juveniles with multiple convictions are more likely to receive further periods of incarceration. A twenty-state research program sponsored by the Office of Juvenile Justice and Delinquency Prevention ‘Juveniles Taken into Custody’ reported programs that shared age 18 as the upper age of juvenile jurisdiction, permitting readmission rates to be calculated over a reasonable time period. Of the 8057 youths released in 1992, 27% were readmitted within 1 year of their release. Male readmission rates were much higher than for females (28% and 16%, respectively), and there was a strong relationship between the number of prior correctional commitments and readmission rates (Krisberg & Howell, 1998).
This review relates mainly to males as they pose a greater challenge both in terms of frequency and severity of their offending (Loeber, Farrington & Waschbusch, 1998). For example, in the Uniform Crime Reports of 1995 males under 18 comprised 86% of the arrests for violent crime (Hawkins, Laub & Lauritsen, 1998).
Farrington stressed the importance of targeting chronic offenders for crime prevention and treatment; given that many violent juvenile offenders are also chronic/versatile offenders that receive institutionalised sentences, the effectiveness of the interventions becomes a critical issue.
Another explanation of the persistence of offending involves the consideration of a stable underlying construct such as criminal potential; an early age of onset is one marker of a high criminal potential, which is later manifested in persistent offending. Again, if many violent juvenile offenders are also general, non-specialised chronic offenders, the task of identifying successful interventions for these juveniles before they reach adulthood is of up most importance.
Multiple problem violent (MPV) boys are boys who are both violent and show multiple problems (non-violent offending, drug use, self-reported delinquency, impulsivity, anti-establishment attitude, heavy drinking and gambling). Recently, Farrington (in press, b), in the context of the Cambridge Study, investigated the risk factors for MPV boys. He reported that the predictors of convictions for violence up to age 32 and of frequent non-violent offenders were very similar (also see Capaldi and Patterson, 1996, as quoted by Farrington, in press, b). The implication is that violent offenders were generally frequent offenders. A review focusing on the effectiveness of interventions for incarcerated violent juvenile offenders therefore has a bearing on the treatment of offenders in general.
Current doubts in the intervention with violent juvenile offenders
The challenges involved in the treatment of the violent delinquents have been widely reported. As Thornberry, Huizinga and Loeber (1995) point out, by the time most serious delinquents are identified and receive intensive treatment from the juvenile justice system, they are well into their delinquent careers. For example, the National Youth Survey in United States (Elliot, 1994; Elliot, Huizinga and Morse, 1986, quoted by Thornberry, Huizinga and Loeber, 1995) found a substantial gap between the peak ages of involvement in serious violence and processing by the juvenile justice system. In addition, the offenders enrolled on treatment programs have a host of negative characteristics that reduce the likelihood of successful intervention. “These offenders are older; are heavily involved in delinquent careers, and are likely to be progressed along overt, covert and authority conflict pathways. They are likely to be involved in other forms of delinquency, to use drugs, and to exhibit other related problem behaviors. They are likely to have multiple risks factors and social deficits […] The consequence is a spiralling behavioral trajectory that is exceedingly hard for prosocial forces to penetrate, but this is precisely what we ask treatment programs to do, often with inadequate program resources and no after care services. Given these limitations, our expectations of treatment programs should be modest” (Thornberry, Huizinga and Loeber, 1995, p. 233).
Lipsey and Wilson in 1998 highlighted the paucity of systematic reviews of interventions with different types of offenders, especially those most serious offenders who might be presumed among the most resistant to treatment. This includes juvenile offenders.
An underlying problem is the dearth of primary intervention research conducted specifically with serious juvenile offenders, most of the samples are mixed including less serious offenders and not separately identified and analyzed. In an attempt to clarify the situation in serious and violent juvenile offenders, Lipsey and Wilson (1998) conducted a meta-analysis (not in the context of a systematic review) focusing on two basic questions:
Does the evidence indicate that intervention programs generally are capable of reducing the reoffending rates for serious delinquents? And if so, what types of programs are most effective?
Lipsey and Wilson included 200 experimental or quasi-experimental studies (published between 1950 and 1995) that involved serious juvenile offenders to some degree (more stringent inclusion criteria produced a very small number of studies). The juveniles finally selected were those “reported to be adjudicated delinquents. In addition, most, or all, of the juveniles had a record of prior offenses and those offenses involved person or property crimes, or an aggregate of all offenses, but not primarily substance abuse, status offenses or traffic offenses” (p. 315). The juvenile samples were largely male and with an average age of 14 to 17 years old. They cateogorized the studies into non-institutionalized (N=117), and institutionalized (N=83).
With non-institutionalized juveniles, treatment effects were larger for juvenile samples with mixed priors (i.e., including some proportion of person offenses) than those with mostly property priors. The most effective interventions were a group composed of interpersonal skills training, individual counselling and behavioural programs. While the less effective interventions were wilderness/challenge programs, early release from probation or parole, deterrence programs (shock incarceration), and vocational programs (distinct from employment related programs).
The results with institutionalized juveniles, contrasted markedly with those for non-institutionalized juveniles: with the offenders in institutions, the treatment effects are much the same for a given program whatever the sample characteristics such as age, gender, ethnic mix and history of prior offenses. Again, the most successful intervention was interpersonal skills training, followed by the teaching family home program (Achievement Place project). The least effective interventions were wilderness/challenge programs, drug abstinence; employment related programs and milieu therapy.
The mean effect sizes were similar for both non-institutional (.14) and institutional interventions (.10) which was not statistically significant. Specifically, the intervention with institutionalized juveniles showed mean effect sizes of .35-.40 for the most effective treatment. In terms of the equivalent recidivism rate differentials, these techniques had an impact on recidivism that was equivalent to reducing a .50 control group baseline to around .30-.35, which is a substantial reduction considering the challenge presented by this category of offender.
Although Lipsey and Wilson categorized interventions as either institutional or non-institutional, they included in the institutionalized category many programs that were, in fact, residential community-based interventions, such as Achievement Place.
According to Andrews et al. (1990), treatment for delinquent behaviour is most effective when the juveniles to whom that treatment is administered have appreciable risk of actually reoffending (the ‘risk principle’). The contrary view, however, is often expressed: that it is the most serious cases that will be least amenable to treatment. The authors’ meta-analysis supported the risk principle: for both groups of offenders, the average intervention program produced a positive effect equivalent to about as 12% reduction in subsequent reoffense rates.
In spite of these results, it remains to be demonstrated what specific strategies are really promising in rehabilitating incarcerated juvenile o ffenders, and, as a subgroup, the serious incarcerated juvenile offender. Presently, we have some preliminary results which suggest that the efforts directed at juveniles are more promising that the ones directed at adults, in addition community treatment is more effective than treatment in prison or residential facilities (see Leschied, Bernfeld & Farrington, 2001; McGuire, 2001). The correctional intervention can be an economically efficient strategy for reducing reoffending in the community (Welsh & Farrington, 2001). Furthermore, some studies have shown that the community-based programs often cost less than their traditional counterparts (Empey and Lubeck, 1971;Henggeler, 1999; Krisberg, Austin & Steele, 1989, quoted by Krisberg, Currie, Onek, & Wiebush, 1995).
Finally, preliminary data suggest that some violent offenders are more amenable to treatment than chronic property offenders (Redondo, Sánchez-Meca and Garrido, 1999). These data, however, are far from conclusive, especially in Europe, in part as a result of the paucity of programmes that can be averaged to extract different conclusions in terms of the moderator variables.
Redondo et al. (1997), in the first European meta-analysis study reported that in terms of crime typology, the most effective interventions (criterion: general improvement) were obtained with offenders against persons (r = .419), and the least with sexual offenders (r = .085), and that juvenile centres (r = .257) and juvenile prisons (r = .193), were more effective than adult prisons (r = .119). Behavioural (r = .279) and cognitive-behavioural programmes (r = .273) were most effective and retribution programmes (r = .039) the least. Concerning just the outcome of recidivism, the mean effect size (ES) was r = .12. Cognitive-behavioural (r = .265) and behavioural programmes (r = .232), appeared twice as successful as the mean of the programmes.
In a second systematic review, Redondo, Sánchez-Meca and Garrido (1999) analysed the specific influence of 32 European treatment programmes (applied during the eighties) on recidivism. As before, important findings included 1) behavioural and cognitive-behavioural programmes were the most effective, 2) treatments were more successful with juvenile offenders. The reason for this probably reflected the use of most successful techniques (behavioural and cognitive-behavioural) with juveniles; and 3) the greatest effectiveness was achieved with violent offenders (not sex offenders), which seems to confirm the risk principle (Andrews et al., 1990).
In an update of the European meta-analysis, Redondo, Sánchez-Meca and Garrido (2002), found that the largest effect sizes were obtained with adolescents (r + = 0.35), although all of the age categories achieved significant positive results.
Outcome measures in this area of “violent offending” also pose a challenge to an investigator and reviewer and must therefore be considered in this review. Serin and Preston (2001, p.219) describe it clearly: “For violent offenders it seems most probable that reductions in violent reoffending would be viewed as the most desirable outcome, yet even then there could be debate, since violence defined by conviction is a poor proxy to actual behaviour. Also, if a violent offender recommits a violent crime, but relative to his history it involves a less serious incident, less victim injury or longer time to reoffense, is this a clear indication of program failure? Defining outcome only dichotomously, then, limits our understanding about program effectiveness”.
The definition of “violent offender” and the issue of measures of recidivism have yet to be clarified. It is necessary to specify with more detail the characteristics of the offenders enrolled into programs and the quality of the reoffending, separating the new violent offenses from the general recidivism rate.
Juvenile offending is a global concern and thus any systematic review should include studies in languages other than English.
Community-based graduated sanctions programs seem to be at least as successful as traditional incarceration programs in reducing recidivism. In addition, community-based programs often cost significantly less than their traditional counterparts. More research must be conducted to give us a clearer picture of what works and what does not work with violent and chronic juvenile offenders (Krisberg, Currie, Onek, & Wiebush, 1995).
The effectiveness of secure corrections treatment, remains to be demonstrated, and this includes traditional juvenile prison, borstal and training schools as well as modern small units for some kinds of offenders (with individualised treatment as a philosophy in the program intervention).
The role played by different moderating variables (for example: prior offense history; chronicity of violent offending; age at intervention, booster programs and gender) requires further investigation, as described by Lipsey and Wilson and the European meta-analysis.
In summary, many gaps remain in our knowledge about the treatment of serious delinquents: Community-based graduated sanctions programs seem to be al least as successful as traditional incarceration programs in reducing recidivism. In addition, community-based programs often cost significantly less than their traditional counterparts. More research must be conducted to give us a clearer picture of what works and what does not work with violent and chronic juvenile offenders (Krisberg, Currie, Onek, & Wiebush, 1995). The Lipsey and Wilson (1998) meta-analysis compared institutionalized and noninstitutionalized treatment for serious youth, but they included in the institutionalized category many programs that were in fact residential community-based interventions, like Achievement Place. It lacks to know which is the contribution of secure corrections treatment, meaning for that traditional juvenile prison, borstal and training schools as well as modern small units for some kinds of offenders (with individualised treatment as a philosophy in the program intervention). The role played by different moderating variables (for example: prior offense history versus non prior history; violent non chronic offenders versus violent chronic offenders; intervention in a early age versus a later age; programs that count with booster treatment after leaving versus programs do not count with it; male delinquents versus female delinquents) has to be more investigated, as we have presented above in the case of Lipsey and Wilson and the European meta-analysis. In our opinion, the theoretical idea of saying that non secure interventions is superior to secure corrections has not been proved empirically in the case of the serious juvenile offenders. The most recent antecedent we have (Lipsey & Wison study) do not support this conclusion clearly. The measurement of “violent offender” and the issue of the recidivism has not been clear and unequivocal. It is necessary to specify with more detail who are the subjects treated and the quality of the reoffending, separating the new violent offenses from the general recidivism rate. Finally, there has been very few studies coming from Europe in the analysis of this topic. It is necessary, in the systematic review, to address all studies we can reach made in other languages different to English.
In order to analyze all the possible varieties of cross cultural studies, we propose the following categories of interventions included in our review (Redondo et al., 1997; Redondo et al., 1999; Garrido et al., 1999):
(1) Interventions aimed at reducing emotional distress and non-behavioural psychological therapies
The belief that offenders commit crime as a result of emotional distress has a long tradition in corrections and according to this concept, the treatment of offenders has to be directed at treating these underlying psychological problems. Therefore successful psychological therapy leads to a reduction or disappearance of criminal behaviour. In this model a heterogeneous set of techniques are utilized including those founded on the psychodynamic model, on medical or pathological model of crime, or on client-centred counselling.
(2) Interventions aimed at rectifying poor educational attainment
Many offenders, especially those of marginal extraction, did not complete their schooling and consequently, have a large educational deficit. The theory is that by increasing educational attainment through intensive schooling programmes recidivism will reduce.
These programmes consist of courses, school activities, delivery of materials for reading, etc. In these educational programmes, the education on theoretical concepts (as grammar, mathematics, etc.) prevails over teaching of practical skills.
(3) Interventions aimed at the learning of criminal behaviour and behavioural intervention
Learning theories (developed in criminology by, among others, Edwin Sutherland, Albert Bandura and Ronald Akers) consider that criminal conduct, as any other human behaviour, is learnt. The objective of behavioural programmes is to employ learning mechanisms to reverse the learning process, so that subjects can learn to inhibit their criminal conduct and can put new socially admissible behaviour into practice.
Two paradigmatic applications of these models are token economy programmes and the environmental contingency programmes. Token economy programmes are well known to those interested in corrections. The so-called system of progressive phases is an environmental contingencies programme applied in some institutions for delinquents and offenders. The main components of the progressive phase system are (Redondo, 1993): A series of behaviour goals are established, with the aim to encompass all the individual's daily life within the institution. A series of life units, or phases are structured into the institution: these phases differ mutually in two fundamental aspects; first, in the level of positive behaviour demanded of the participants; higher in the superior levels than in the inferior and second, the units are different in the available rewards for each unit, more frequent or intense in the superior units. The subjects are periodically re -allotted into the phases according to their improvements.
All the staff of an institution must be involved led by a small group of experts, in charge of the design, supervision and the evaluation of the programme.
(4) Social interaction skills and cognitive -behavioural intervention
These techniques are based on the cognitive-behavioural model, that emphasizes the need to teach offenders skills that will make their interaction with other people easier, either within the family, in their jobs, or in any other social context. One of the most complete cognitive-behavioural programme is that which follows the model of “Reasoning and Rehabilitation” (Ross and Fabiano, 1985; see also Ross and Ross, 1995). The principal elements include: Evaluation of the subject's deficits in cognitive and interaction skills. Treatment is applied in small groups for several weekly sessions. The most often used strategies include: interpersonal cognitive problems solving, social skills training, emotional control of anger explosions; critical reasoning; values development, negotiation abilities and creative thinking. Currently, cognitive-behavioural programmes are the ones most used with all kind of offenders (see Ross and Ross, 1995).
(5) Deterrence theory and the hardening of prison regime
Toughening the prisoners’ living conditions cannot be considered a therapeutic technique, however, in recent years, some jurisdictions have devised centres with strict living and tough discipline, inspired by military regimens. The inmates, both youths and adults, are offered a choice between serving their sentences entirely in a standard prison or serving a shorter sentence in these “special” institutions.
This intervention is based on the classical doctrine of deterrence; the application of penal sanctions to criminals would remove or reduce their future criminal conduct per se and the stricter the regime the greater will be its deterrence effect. The basic tenets of this model include: Rigid discipline and supervision regime encompasses every aspect of the inmate's life. Binding actions; that include work (not always of useful character), gymnastics, marches and, sometimes, group sessions, are planned. An inflexible sanctions system is applied.
(6) Healthy institutional environments and therapeutic communities
Therapeutic communities attempt to encompass all of the subject's daily life within the institutions. Relationships between imprisoned people and institutional staff are assimilated to those of patient/nurses in a therapeutic context. The main theoretical assumption is that healthy and participative environments in custodial institutions will bring about greater psychological balance in inmates and will reduce their violent behaviour, both during their period of incarceration and on release. The principal features of therapeutic communities include: Rigid controls and sanctions systems, common in closed institutions, are suppressed. Control of inmates’ behaviour depends on the community, composed of inmates and staff. Periodic community assemblies are carried out to discuss the problems that arise in the institution.
For a therapeutic community to be effective all staff need to collaborate.
(7) Avoiding labelling through the diversion programmes
This paradigm is not included in our review due to our focus on residential settings.
3. Objectives of the Review
General Objectives
To collate and assess the quality, in a systematic way, the outcomes of empirical research regarding the effectiveness of treatment programs implemented in secure corrections in order to decrease the reoffense rate and quality (i.e., type of offence) of serious (chronic and violent), delinquents (12-21 years old).
Specifics Objectives
To identify empirical published and unpublished studies (in different languages) relating to the evaluation of correctional intervention programs for institutionalized serious (chronic and violent) juvenile offenders. To classify the outcomes and characteristics of the identified correctional programs into the following categories: what works, what does not work, what is promising and what is unknown (see below), in accordance with the scale proposed by Sherman et al. (1997). To analyse the effects of correctional intervention in three different groups of serious delinquents: (1) chronic (no violent), (2) chronic and violent, and (3) violent. To discuss the outcomes and to propose, with the support of available evidence, correctional policy recommendations to reduce the reoffense rates (in terms of rate and quality) in this subgroup of offenders.
We will include in the category of serious delinquents both violent and chronic offenders. As explained earlier, chronic and violent juveniles have many characteristics in common. Additionally, the majority of violent juveniles have previous criminal convictions and it is reported that almost half of persistent delinquents had committed violent offenses, (the criteria utilised in our research to be considered a sample as “violent”).
We will complete three independent analyses: Chronic but not violent delinquents. Violent delinquents adjudicated for serious offenses either currently or previously. Violent and chronic juveniles, which means, studies where less than half of the sample are violent delinquents, but with the addition of chronic offenders is higher than 50% of the total population in the research.
Meta-analysis A: samples that include persistent offenders, no violent offenders Meta-analysis B: samples that include only violent offenders (they can be persistent or not) Meta-analysis C: samples that include persistent offenders, non-violent
This review will pay particular attention to the variability caused by moderating variables. These include: type of treatment (theoretical framework of the treatment, length and intensity of the programme), subjects or participants in the programs (age, type of offense, gender), the setting in which the intervention occurs (the stage in which the programme is applied, the regime of the participants, and the country), methodology (type of assignment of the participants to the groups, groups mortality, and the follow-up period), and extrinsic variables (date of publication of the article and source) (Lipsey, 1994; Sánchez-Meca, 1997).
4. Methodology
Criteria for inclusion and exclusion of studies in the review
Types of studies
This review will include experimental and quasi-experimental studies with control or comparison groups and with prior and later assessment of the intervention. Furthermore, the outcomes presented include relapse rates and offenses. The inclusion of non-randomised control groups is necessary due to the paucity of well-controlled studies in the area of correctional intervention. Lipsey and Wilson (1998, p. 314) recognised this and stated in their meta-analysis that “These are questions that are answered most convincingly by experimental and quasi-experimental studies in which the subsequent offending rate of juveniles given treatment is contrasted with that of an otherwise comparable control group not given treatment. Such research yields statistical findings that represent the magnitude of the treatment effect observed in each study”.
Participants will be under a judicial supervision system (criminal or juvenile). A sample of individuals, residential units, specific modules or even an entire prison may be regarded as treatment or control groups.
In this review we will use the scientific methods scale of Sherman et al. (1997) and we will include only those studies ranked 3, 4 or 5 (see Appendix 1). As consequence, we exclude qualitative and pre experimental quantitative research papers (rank 1 or 2 in the same scale) as well as single case design reports. However, data from this kind of reports will be used to identify bibliographical references and other important variables.
Types of participants
The program recipients will be juveniles either male or female, in secure corrections aged between 12 and 21 years old, under either the adult or juvenile jurisdictions, characterised as serious (chronic and violent) delinquents.
In general, juvenile offenders are considered a group of young people from 12 to 21 years old (Fuhrman, 1986; Tolan & Guerra, 1994; Rutter, Giller & Hagell, 1998). In accord with the results of longitudinal research about criminal careers, during these ages more people is implied in offending; we also know that the habitual delinquents commit the most part of their offenses during this period. Furthermore, if we take in account the legal context, the range of age for legal responsibility in different countries oscillates between 12 to 21 years of age (Garrido, Stangeland & Redondo, 2001).
With regard to the category of serious delinquents we will determine that the population in the selected studies belongs to this category by inspecting the type of offense committed and their previous convictions.
a) Violent delinquents. We define as juveniles that have committed violent offenses. These comprise offenses “those acts in which someone is hurt and resulted in serious injury (requiring medical treatment-cut, bleeding, unconscious, etc.) or in which a weapon is used” (Thornberry et al., 1995, p. 224 in reference to Denver Youth Survey). Furthermore, we will include delictive acts that involve threatening behavior by physical force.
We will include studies in which more than half of the sample have committed or have a history of offenses such as: murder (and attempted murder), homicide, kidnapping, assault (including aggravated), robbery (taking into account armed robbery), car theft, voluntary manslaughter, endangerment, arson of occupied building. Other offenses such as the ones against life and integrity of others with serious body injuries caused by physical force or weapons as firearms, cold steels, etc are included. (See Wiebush, Baird, Krisberg & Onex, 1994; Wiebush et al., 1995, about the category of “serious and violent” offenses in which is based the Annual Survey of the Office of Juvenile Justice and Delinquency Prevention (OJJDP) p. 176).
b) Chronic delinquents. We define “persistent offender” as those juveniles with three or more previous legal adjudications (as they had been defined by some Justice Organizations as the Department of Juvenile Justice –DJJ- of United States, 1998; or in some studies like those of Capaldy & Paterson, 1996; Hagell and Newburn, 1994).
We will include studies in which more than half of the sample corresponds with juveniles with three or more previous legal adjudications, or studies where the mean of the criminal history of the sample is three or more previous legal adjudications for any kind of offenses except violent ones.
c) Violent and chronic delinquents. We include studies where the samples do not reach half the violent delinquents, but the addition of chronic and violent individuals is higher than 50% of the total population in the research.
Finally, we will exclude studies in which more than half of the sample are sexual offenders, as this is the focus of another Campbell Collaboration systematic review (lead author Friedrich Lösel). Studies that include juveniles committing minor offenses such as shoplifting, minor public order, traffic offenses and status offenses for the first time (see Wiebush et al., 1995, pp. 176, 210) will be excluded too.
The term “secure corrections” means, in this review, environment or secure institutions characterized by physical restraint measures as locked doors, walls, bars, fences, etc. We will include as secure corrections: prisons, borstals, training schools, camps, ranches, special hospitals and any other more modern facility that holds juveniles accountable for their delinquent acts and provides a structured treatment environment. We exclude community programs or programs such as foster care, foster home, group home, periodical detention and, in general, those in which delinquents are in contact every day with the community (as Achievement Place).
Because of the existence of institutionalised programs with the latter period spent in the community, we will include the studies in which more than the 50% of the treatment takes place in the institution. In those cases the treatment in the community will be registered as a moderating variable.
The systematic review will include any kind of violent juvenile offender institutionalised in a secure facility, with mental retardation or other pathology.
Types of interventions
We will include interventions aimed at decreasing post treatment relapse when the juveniles are returned into the community. These include psychological approaches, social and educational procedures and methods, as well as environmental conditions directed to support the learning of prosocial behaviours and attitudes (see background section). The classification of the interventions is as follows: Non-behavioural psychological therapies. Educational interventions. Behavioural interventions. Interpersonal skills and cognitive-behavioural interventions. The hardening of facility regime, increasing discipline and sanctions. Healthy institutional environments and therapeutic communities.
Specifically, this review excludes studies that correspond with other Systematic Reviews from the Campbell Crime and Justice Group such as boot camps or scared straight programs.
Types of outcome measures
Eligible studies must report subsequent delinquent or criminal offenses as an outcome. Outcome measures must be presented in a quantitative way. Where different outcomes are reported we will make a secondary analysis, specifying that the eligibility criteria will be the report of reoffense measures.
This review will consider the following types of measures: Offense Measures: General recidivism: Official records obtained from the police or adult/juvenile justice courts, that involve any kind of new offenses with any kind of court response (parole, prison, etc.). Self reports on criminal behaviour from the offender after leaving the program. Such outcomes identify violent behaviours that did result in an arrest. Serious recidivism: any new official serious registered offense that causes a new ingress in the secure facility Recidivism rate: time passed since release from facility until the first offense committed and number of offenses committed during the follow-up. When a study measures the recidivism at various times after release, all those different evaluation points are registered, and priority is given to the longest follow-up period. When a study measures the recidivism at various times after release, all those different evaluation points are registered, and priority is given to the longest follow-up period. However, as we explain in another apart of this protocol –Criteria for determination of independent findings (page 15)-: if the studies selected have presented several outcome measures, we will compute the effect size for each outcome measure in separated meta-analysis in each temporal measure, in this way we will include all the follow up in the studies and will avoid the dependency in the data. Quality of the recidivism: type of offense committed after release. Outcome measures to be considered in secondary analysis: Psychometric Measures: Outcomes of test and inventories related to variables such as self-control, locus of control, psychological adjustment, self-esteem, etc. Behaviour achievements: getting a job, drugs and alcohol abstinence, etc.
Search strategy for identification of relevant studies.
Several strategies will be used to identify studies meeting these criteria. In order to reduce potential biasing, we will search: Published and unpublished. Between 1970 – 2002 Studies in areas of criminology, psychology, sociology, social service, education and psychiatry. From any country and one of the following idioms: English, Spanish, French, Portuguese, German and Italian.
The keywords used are:
Delinquent (s), delinquency, criminal(s), convicted, antisocial behavior, offender.
Delincuencia, delincuentes, criminales, conducta antisocial, encarcelados.
Institution (alized, al, alization), detention, facility (ies), prison (s, ers), incarceration (ed), hospital (s), borstal (s), correctional (s), reformatories.
Institución, institucionalizados, detenidos, prisión (es, eros), encarcelamiento, hospitales, correccionales, reformatorios.
Boy (s), girl (s), adolescent (ce), juvenil (es), youth, young.
Jóvenes, juvenil, adolescentes.
Treatment (s), program (s), therapy (ies), rehabilitation, adjustment, intervention (s). Tratamiento, programa, terapia, rehabilitación, ajuste o intervención.
Agression (sive), anger, violence, violent, serious, chronic, persistent.
Agresión, ira, violencia, violento, serio, crónico o persistente.
Methods to search studies:
In accordance with Petrosino (1997) the methods that will be used to search studies are: We will hand search a selection of specialized relevant journal contents that are held in our universities. We will review only some journals for more detailed search, as many are indexed in the electronic databases – listed in the next part of this protocol-.
Anuario de psicología jurídica / Colegio Oficial de Psicólogos/ Madrid Criminalia / Academia Mexicana de Ciencias Penales Criminoticias / Instituto Andaluz Interuniversitario de Criminología (I.A.I.C). Sección de la Universidad de Sevilla Delincuencia = Delinquency: a social sciences interdisciplinary journal: una revista interdisciplinar desde las ciencias sociales / Universidad de Valencia Dei delitti e delle Pene: Revista di Studi Sociali, Storici e Giuridico Sulle Questione Crimenelle / Edizione Scientifiche Italiane Papers d'estudis i formacio / Generalitat de Catalunya, Departament de Justicia, Centre d'Estudis i Formacio
Adolescence British Journal of Criminology Criminal Justice and Behavior: an international journal / American Association for Correctional Psychology Criminology: an interdisciplinary journal / American Society of Criminology, Columbus, Ohio Criminology and Penology Abstracts / Criminological Foundation Criminology, penology and police science abstracts / Criminological Foundation Developmental Psychology Journal of Adolescence Journal of Clinical Psychology Journal of Legal and Criminological Psychology Journal of Applied Behavior Analysis Personality and Individual Differences Keyword searching of electronic databases: Psyque (spanish database) Current Contents Criminal Justice Abstracts Dissertation Abstracts Education Gull Text. ERIC (Education Resource Information Clearinghouse). Humanities Abstracts Medline PAIS International (Public Affairs Information Service) PsycINFO (Psychological Abstracts) Serfile Sigle Social Sciences Citation Index Sociofile (Sociological Abstracts and Social Planning and Development Abstracts) (Dr. Kury will look for relevant databases in German) Requests to researchers known to the reviewers working in institutions and authors working in programs aimed to improve incarcerated serious juvenile behaviour or in similar topics by either electronic or postal mail. Books, articles, meta-analysis and other documents tracking. World Wide Web searching with two popular search engines: Altavista and Google. Searching through the “Broad searches of the Campbell Collaboration Social, Psychological, Educational & Criminological Trials Register (C2-SPECTR) development by the U.K. Cochrane Centre and supervised by Pennsylvania University (Graduate School of Education) (Petrosino et al. 2000).
Eligibility of the studies
Both reviewers will examine the title and abstract of all studies identified. A copy of all eligible studies will be collated. An instrument designed to assess reviewers’ reliability will be used.
The complete reports of the selected studies will be obtained and read. The reviewers will assess the eligibility of studies using a checklist. This instrument was based on the example of the Eligibility checklist (version 9/26/00) by MacKenzie, Wilson & Kilder (2001) for the Systematic Review of Bootcamps, and modified for the specific needs of this review.
The Checklist (Appendix 1) contains the description of eligibility criteria and was applied to ten studies as a pilot test. On the basis of this pilot some items were changed.
Appendices 2 and 3 contain some examples of included and excluded studies.
Description of methods used in primary research
Few studies in the area of juvenile offender treatment research focus on empirical evaluation studies. Many of the studies we identified are theoretical assessments about the quality of a particular approach or the best way to consider several issues of the facilities regimes. Many identified studies specifically excluded those young people whom this review is focussed, namely violent juveniles. For example, in the study of Kirigin et al (1982), that presents the teaching family model –shown as example 3 in Appendix 3- it is clearly stated that “none of the programs admitted youths with histories of violence such as a murder, forcible rape or armed robbery”. Other studies try to present empirical outcomes, but the design and the analysis completed do not allow well founded conclusions to be made.
Sample sizes are often very small (8 or 10 subjects), selected for convenience. The results are generally qualitative inferences coming from staff or inmates themselves, or are based in some activities that are improved as a consequence of the analysed program (for example school attendance, disciplinary reports or some psychological variables registered in inventories and questionnaires). Schlichter & Horan (1981) (example 1 in appendix 3), attempted to analyse the effects of stress inoculation in institutionalised juvenile offenders, however, the outcome measures were limited to self-report scales and laboratory role-played simulations.
Criteria for determination of independent findings
There are two possible problems of dependency in the data. First, it is probable that the selected empirical studies report multiple outcomes such as recidivism, social adjustment, psychological adjustment, etc. in the same sample of subjects. Second, it is possible that studies report outcome measures at different times points such as recidivism at 6 months, 12 months, 24 months, etc. The reviewers propose two strategies in course to avoid these dependency problems: To resolve the first problem, this review proposes to compute the effect size (ES) for each outcome measure, that means, an index about recidivism, social and/ or psychological adjustments, etc. The reviewers will do separate meta - analysis for each outcome measure, one for recidivism outcome measures, others for social adjustment outcome measures, and successively with others. In this way, each meta-analysis will avoid the dependency problem. To resolve the second problem, this review proposes to compute the effect size (ES) only for the longest follow-up period data. For instance, if one study reports outcome measures at six months, at 12 months and at 24 months, the reviewers will only compute the ES at 24 months, because it is the longest follow-up. In this way, we will include in each meta-analysis only one estimate of ES, thereby avoiding the dependency problem. This second strategy can be mixed with another one a little more complicated, but at the same time more complete. If many studies have presented several outcome measures, for instance, at 6 months and at 12 months after the implemented program, we could compute the ES for each outcome measure in separated meta-analysis in each temporal measure in order to avoid the data dependency problem. This strategy can only be decided “ad hoc”, that is to say, only when we have selected the studies and when we have proved that the most part of those reports correspond with the same time frames (e.g. all selected studies report outcome measures 6 months, 12 months and 24 months after release).
Coding categories
The reviewers have developed a specific preliminary coding protocol to register the characteristics of all the chosen studies. This second instrument was constructed, as the first one, on the basis of the literature reviewed about correctional intervention programs for serious institutionalised delinquents. We have also taken into account previous experience in others systematic reviews such as the Boot Camp review (MacKenzie, Wilson & Kilder, 2001) and previous meta-analyses, Marín-Martínez, Hidalgo, López, López, Moreno, Redondo, Rosa and Sánchez-Meca, (2002).
Building on the work of Lipsey (1994) and Sánchez-Meca (1997) the coding instrument is divided in three variables groups: Substantive variables: this category includes the characteristics that are object of this review, this means special features about the research participants, treatment and context. Methodological variables: this group is referred to the qualities of the design of the selected studies. Extrinsic variables: this type of the studies characteristics are not directly linked with the object of this review; however, these can affect its results.
The coding instrument is located in appendix 4.
Statistical procedures and conventions
Data analysis
Two software packages will be utilized to analyse the data:
“Comprehensive Meta-analysis” (Vers. 1.0.5) (Borenstein & Rothstein, 1999).
MetaWin (Vers. 2.0) (Rosenberg, Admas & Gurevitch, 2000).
The reviewers will use tables to describe the distribution of the moderating variables coded in the selected studies in order to present the data. Thus allowing the systematic presentation of the characteristics of the interventions (theoretical framework of the treatment, length, intensity of the program etc.), the subjects/participants in the programs (age, type of offense, gender), the setting in which the intervention occurs (the stage in which the program is applied, the regime of the participants, and the country), methodology (type of assignment of the participants to the groups, attrition, and the follow-up period), and extrinsic variables (date of publication of the article and source) (Lipsey, 1994; Sánchez-Meca, 1997).
We intend to use frequencies and percentages, means, standard deviations, etc. depending of the quantitative or qualitative nature of the moderating variables.
We will use different index of ES in accordance with the type of outcome measures (quantitative or dichotomous). With dichotomous outcomes (e.g. recidivism vs. not recidivism) and there are two groups (intervention and control group) the reviewers will make a 2X2 Contingency Table. On the basis of this table we will calculate the odds ratio and the transformation to standardized mean difference through the transformation formula development for Hasselblad and Hedges (1995; see also Rindskopf and Shadish, 1998).
In cases of studies with quantitative outcome measures the reviewers will calculate the standardised mean difference between the treatment and control group, the result will be divided by a within-group standard deviation estimate. This procedure will be applied when the studies report means and standard deviations or in situations which the outcomes of inferential statistical tests as T-tests, analysis of variance, etc. are presented.
Once the index in each study is calculated, this will be corrected by the correction factor of positive bias. This bias is frequent in small sample sizes, as is described by Hedges and Olkin (1985).
The handling of missing data is important, and we will maintain these as missing, meaning that these data will not be substituted by other estimation on the basis of the available data in the others studies of the meta-analyses. When the case permits and requires, the reviewers will add data missing calculating the median of the available data.
The objective of this review is to carry out a quantitative synthesis of the outcome measures of the selected studies. Thus, applying statistical techniques about the index of ES and about the coded moderator variables.
We will use the meta-analytic procedures developed by Hedges & Olkin (1985) and in Cooper & Hedges (1994). In accordance with these authors the mean effect size is computed by weighting each effect size in function of its reliability, that means, by the inverse of its variance. This procedure implies that the studies with the biggest sample size will have greater weight in the statistics analysis because those are more reliable.
The statistics analysis will be applied from the fixed and random effects model..
In each meta - analysis (remember that we will do one meta-analysis for each outcome measure - recidivism, psychological adjustment or social adjustment, etc. -) we will apply the following statistics analysis: Calculation of the global mean effect size of every study with its 95% confidence interval by the fixed and random effects model. Calculation of the homogeneity test to check if the ES of the selected studies are homogeneous. If the result of the homogeneity test is statistically significant, then the reviewers will explain this heterogeneity showing which qualitative or quantitative moderator variables are affecting the effect size. When the moderator variables are qualitative (e.g. theoretical model of intervention), we will apply a weighted analysis of variance. When the moderator variables are quantitative (e.g., age of the subjects in the sample or the study) will be apply a weighted regression analysis. Finally, in this review we propose a hierarchical weighted multiple regression mo del to determine which subgroup of analysed moderator variables can explain the most part of variability of the ES.
The reviewers will carry out several sensitivity analyses.
1. Using the fixed and random effects models is, in fact, a sensitivity analysis about the statistical analysis applied. If the results obtained with the two types of models match, that gives support to our findings. If not, it will be a better approach to base our conclusions on the results obtained from the random effects model, which is the one that corresponds better to the population characteristics of the meta-analysed studies.
2. Because both random and non-random assignment studies are included in this review, we will complete a comparative analysis of the ES obtained with both types of designs in order to determine if the studies methodological quality affects the results. With significant differences between both types of design, to present the effectiveness in a separate way will be preferable; stressing the results obtained from the most solid random assignment studies.
3. Another sensitivity analysis concerns publication bias. We intend to compare the ES obtained from the published with that from the non-published studies, analysing if there are differences between them. In the case of not counting with enough non-published studies, we will apply the ‘fail-safe index’ developed by Orwin (1983) to determinate the consistency of our effectiveness results.
5. Time Frame
6. Plans for Updating the Review
We will update annually by repeating the specific searches of databases using the same keywords than in the initial work.
7. Acknowledgements
Dr. Julio Sánchez-Meca (University of Murcia, Spain) contributed to this protocol as a consultant in the statistical procedures
8. Conflict of Interest
There is no declared conflict of interest.