From Juvenile Justice to Criminal Justice: Race and Psychosocial Profiles as Predictors

Da Justiça Juvenil à Justiça Criminal: Raça e Perfis Psicossociais como Preditores

De la Justicia Juvenil a la Justicia Penal: Raza y Perfiles Psicosociales como Predictores

Abstract:

Various factors may affect the likelihood of individuals who commit offenses during adolescence continuing to offend into adulthood. This study aimed to: (1) Describe and compare recidivism rates among 350 adult men who had gone through the juvenile system in the countryside of São Paulo; (2) Evaluate the prediction of recidivism according to psychosocial profiles; (3) Assess the influence of race on recidivism. Official recidivism data from the sample were collected and analyzed using logistic regression analysis, revealing that a more markedly problematic psychosocial profile was associated with greater chances of recidivism, while being Black was linked to higher chances of criminal prosecution. This study highlights the importance of identifying which psychosocial profiles are associated with a higher likelihood of persistent offending to target more effective interventions. It also reveals the presence of racial bias in the Brazilian criminal justice system, indicating structural racism.

Keywords:
juvenile delinquency; teenager in conflict with the law; psychosocial development; racism; criminology

Resumo:

Diferentes fatores podem influenciar as chances de um indivíduo que comete ato infracional na adolescência continuar a cometer delitos na adultez. Esse estudo teve como objetivo: (1) Descrever e comparar as taxas de reincidência de 350 homens adultos com passagem pelo sistema socioeducativo no interior de São Paulo; (2) Avaliar a predição da reincidência em função de perfis psicossociais; (3) Averiguar a influência da raça na reincidência. Foram coletados e analisados dados de reincidência oficial da amostra e, utilizando análises de regressão logística, observou-se que o perfil psicossocial mais marcadamente problemático se relacionou com maiores chances de reincidência e a raça preta, com maiores chances de processamento penal. Esse estudo mostra a importância de se conhecer quais perfis psicossociais se associam com maiores chances de persistência na conduta infracional, para focalização de intervenções mais efetivas, e que há um viés racial presente no sistema criminal brasileiro, denotando racismo estrutural.

Palavras-chave:
delinquência juvenil; adolescente em conflito com a lei; desenvolvimento psicossocial; racismo; criminologia

Resumen:

Diferentes factores pueden influir en la probabilidad de que una persona que delinque en la adolescencia siga delinquiendo en la edad adulta. Este estudio tuvo por objetivo: (1) Describir y comparar las tasas de reincidencia de 350 hombres adultos que pasaron por el sistema socioeducativo en el interior de São Paulo; (2) Evaluar la predicción de reincidencia según los perfiles psicosociales; (3) Verificar la influencia de la raza en la reincidencia. Se recopilaron y analizaron datos de reincidencia oficial de la muestra y, mediante análisis de regresión logística, se observó que el perfil psicosocial más marcadamente problemático se asoció a mayor probabilidad de reincidencia y la raza negra a mayores probabilidades de enjuiciamiento penal. Este estudio demuestra la importancia de conocer los perfiles psicosociales asociados a mayores probabilidades de persistencia en la conducta delictiva, para poner el foco en intervenciones más eficaces, y marca la existencia de sesgo racial en la justicia penal brasileña, lo que denota racismo estructural.

Palabras clave:
delincuencia juvenil; adolescente en conflicto con la ley; desarrollo psicosocial; racismo; criminología

Adolescents who have committed delinquent acts constitute a heterogeneous group in terms of their exposure to risk and protective factors associated with the persistence of delinquent behavior. Identifying these factors, both individually and collectively, is crucial as it provides valuable insights to guide effective intervention and prevention programs in the field of juvenile delinquency (Basto-Pereira & Farrington, 2022 ).

Currently, there is extensive evidence regarding factors that significantly increase the likelihood of adolescents exhibiting a persistent pattern of delinquent behavior (frequent and stable, tending to continue into adulthood) (Campbell et al., 2019 ; Cox et al., 2018 ; Johnson & Moyer, 2021 ; Moffitt, 2018 ; Sapori et al., 2020 ), including national empirical studies (Galinari & Bazon, 2021 ). In a meta-analysis, Basto-Pereira and Farrington ( 2022 ) highlight five main dimensions among the risk factors for persistent delinquency: (1) Problems related to education/employment; (2) Association with delinquent peers; (3) Family problems; (4) Alcohol/substance abuse; (5) Specific mental health problems (e.g. anxiety, stress). The authors argue that the cumulative effect of these factors significantly increases the likelihood that individuals who have engaged in delinquent behavior during adolescence will continue to offend into adulthood.

On the other hand, protective factors are variables that can diminish or nullify the negative effects of risk factors, such that the manifestation of delinquent behavior can be discontinued, even for individuals with a significant presence of risk factors. Examples include participation in pro-social activities, interest in performing well at work or school, self-control, and a strong bond with a pro-social adult. Komatsu et al. ( 2019 ) used protective factors to classify 4 groups of adolescents in conflict with the law (from most protected to most vulnerable). By analyzing re-entry data from 858 adolescents, they observed that the group with the fewest protective factors (the most vulnerable) had the highest re-entry rates into the adult system, while the group with the most protective factors (the most protected) had significantly lower re-entry rates (Komatsu et al., 2019 ).

Using cluster analysis, Campbell et al. ( 2019 ) investigated risk factors for persistent delinquency by using latent profile analysis in a sample of 1,263 adolescents prosecuted in court. The authors identified subgroups with distinct profiles regarding exposure to risk factors and, consequently, differing needs for intervention. They classified the groups as follows: Minimal Intervention Needs (Profile 1); Social Behavior and Social Bonding Needs (Profile 2); Maximum Treatment Needs (Profile 3). Although Profiles 2 and 3 showed similar levels of exposure to the assessed risk factors (total scores), they differed in the domains that characterized them, indicating the need for differentiated psychosocial interventions. Profile 3 (Maximum Treatment Needs) had the highest rates of recidivism/reoffending, while Profile 1 (Minimal Intervention Needs) had the lowest rates. These authors, along with others (Cuervo et al., 2020 ; Mulder et al., 2019 ), argue that the association between recidivism/reoffending rates and risk scores is not perfectly linear. Results may be more accurate when methods that group adolescents based on levels of exposure to risk factors and similar intervention needs are used.

In the Brazilian context, Galinari and Bazon ( 2021 ) conducted a study with a similar methodology (latent class analysis) to identify psychosocial profiles in a sample of 400 adjudicated adolescents. The authors identified four different profiles in terms of delinquent behavior patterns and psychosocial characteristics. Profile 1 was very similar to the general population in terms of average scores on the assessment tools used, while Profile 4 had scores that were most different from the general population, characterized by greater exposure to risk factors for persistent delinquent behavior beyond adolescence.

Hypothesizing that the profiles (1 to 4) represent a progressive increase in exposure to risk factors related to persistent delinquent behavior, with Profile 1 being the least exposed and Profile 4 being the most exposed, the study proposed here is aimed at verifying the rate of recidivism/reoffending in the context of Criminal Justice (considering time progression) for the entire sample, as well as for the individual profiles. It was hypothesized that adolescents whose data contributed to describing the most problematic profiles in terms of behavior patterns and psychosocial characteristics — Profiles 3 and 4 — would have a higher likelihood of recidivism due to their increased exposure to risk factors associated with persistent delinquent behavior. In Brazilian literature, no studies were identified focusing on the analysis of profiles in terms of exposure to risk factors and their relation to the likelihood of recidivism among adolescents in conflict with the law. The few studies found use linear methods (rather than clustering methods) to identify and analyze variables associated with recidivism (Sapori et al., 2020 ).

It is important to emphasize that recidivism data generally refers to official records related to new offenses reported to control agencies (such as police or justice). Thus, official data as a measure of recidivism/reoffending requires critical consideration, as the operations of control agencies are not “neutral” in their methods of operation. One should consider the possibility of selectivity governed by variables that are not necessarily the same as those governing the criminal phenomenon itself (Sapori et al., 2020 ; Soares & Ribeiro, 2018 ). In Brazil, for example, “race” has been highlighted as a characteristic that can bias the operations of control agencies. Literature argues that non-white individuals (especially Black and Brown - multiracial individuals) tend to be more monitored, apprehended, prosecuted, and convicted for crimes compared to white individuals, due to racial discrimination and labeling processes that make this segment of the population a target for control agencies (Sinhoretto & Lima, 2015 ; Soares & Ribeiro, 2018 ). As a result, official data may end up reflecting the selective actions of the system itself (Soares & Ribeiro, 2018 ).

According to Labeling Theory, individuals are labeled as criminals through (pre)conceived notions that define the biosocial characteristics of those more likely to commit offenses. In Brazil, Black and poor men are more intensely monitored, accused, and sanctioned because they are socially marked as inferior (Soares & Ribeiro, 2018 ). Following this perspective, criminal justice system institutions produce crime and criminals in a selective manner (Soares & Ribeiro, 2018 , p. 91), as they exercise differentiated surveillance over individuals considered potential criminals.

In the Brazilian sociocultural context, although the debate about race as a criterion for penal selectivity is intense (Soares & Ribeiro, 2018 ), few empirical studies address this topic and show whether this variable constitutes indeed a bias, especially in the transition between “juvenile delinquency” and “adult criminality” (Sinhoretto & Lima, 2015 ). Therefore, this study also focuses on the social variable “race,” seeking to observe “if” and “to what extent” it impacts recidivism/reoffending outcomes in the studied sample. In this study, the term “race” is used based on the concept of “social race,” present in other Brazilian studies (Guimarães, 2011 ; Soares & Ribeiro, 2018 ), understood as a nominalist concept, that is, one that expresses something that does not actually exist in the physical world but has effects in social reality (Guimarães, 2011 ). This study aimed to: (1) Describe and compare recidivism rates among 350 adult men who had gone through the juvenile system in the countryside of São Paulo; (2) Evaluate the prediction of recidivism according to psychosocial profiles; (3) Assess the influence of race on recidivism.

Method

Participants

As previously mentioned, the participants were the same as those in the study conducted by Galinari and Bazon ( 2021 ), which collected data from 400 adjudicated adolescents in a town in São Paulo state for committing an infraction. However, due to the lack of data for 50 participants, they could not be included in the analyses of this study, resulting in a final sample of 350 individuals.

Primary data for 130 participants were collected throughout 2015, while data for the remaining 270 were collected throughout 2018. All participants were male, aged 16 years old or more at the time of the primary data collection ( M = 16.98 years old, DP = 0.78). Among the adolescents, 48% were in provisional detention, 35% were serving a socio-educational measure in a closed facility (detention), and 16% were serving a socio-educational measure in an open facility (Supervised Freedom/Probation). They came from families with the following socioeconomic characteristics, according to the Brazil Criteria (Associação Brasileira de Empresas de Pesquisa [ABEP], 2016 ): 1%-A; 1%-B1; 11%-B2; 37%C1; 33%-C2; 18%-D-E. Based on this classification, the estimated average household incomes were: A – R$20,888; B1 - R$9,254; B2 – R$4,852; C1 – R$2,705; C2 – R$1,625; D-E – R$768 (ABEP, 2016 ).

Instruments

In this study, the data came from this sample that were already analyzed using Latent Class Analysis based on descriptive variables of delinquent behavior patterns and psychosocial characteristics, organized into four subgroups, from which the profiles were derived (Galinari & Bazon, 2021 ).

Three variables were emphasized: (1) Psychosocial Profile (independent variable); (2) Race (independent variable); e (3) Recidivism/Reiteration (dependent variable). The psychosocial profiles described by Galinari and Bazon ( 2021 ) are characterized as follows: Profile 1 (18% of the sample): Represents adolescents with a common delinquency pattern, meaning occasional delinquency, with scores on psychosocial variables also within normative ranges. Profile 2 (16% of the sample): Refers to adolescents with a distinctive delinquency pattern (as opposed to common), but still characterized by low frequency and absence of violence, marked by above-norm scores (indicating difficulties) in personal/psychological regulation variables. Profile 3 (24.5% of the sample): Describes adolescents with a distinctive delinquency pattern, characterized by frequency and diversity of offenses considerably above the normative average, involving violent crimes, and indicative of problems/needs in social dimensions (such as family violence and association with antisocial peers), but with normative scores in personal/psychological variables. Profile 4 (41.5% of the sample): Characterizes adolescents with a distinctive delinquency pattern, marked by frequency and diversity of offenses considerably above the average, involving violent crimes (similar to Profile 3), and above-norm scores in various social and personal/psychological variables, indicating issues across multiple domains. Regarding race, the data, also stored in the database, were collected at the time of the initial data collection (Galinari & Bazon, 2021 ) through hetero-identification using the following categories: White, Brown (multiracial individuals), and Black (no individuals were classified with other races/ethnicities). However, race data were only available for 315 individuals ( n = 315). The recidivism variable, in turn, refers to information collected for this study, related to first-instance cases from the São Paulo Court of Justice.

Procedures

Data collection. The search for the existence of criminal prosecution was conducted using the full names of the individuals through consultation in the Justice Automation System (e-SAJ) , which is publicly accessible. This data was collected between December 2020 and May 2021 (between 1.7 and 5.9 years after the adolescents had turned 18; M = 3,679; sd = 1,426). Only criminal actions and executions involving individuals from the initial sample (400 individuals) were selected and counted for analysis. It is noteworthy that data for 50 individuals could not be used as the search results in e-SAJ for these individuals were inconsistent, making it impossible to differentiate which Criminal Cases referred to individuals in the sample due to name similarities. Thus, the results presented in this study pertain to a total of 350 participants (final sample). Information about criminal actions and executions was collected for each Criminal Cases, with a maximum of five processes being collected per individual, per category (Criminal prosecution or criminal execution).

The collected data were organized into the following categories: (a) Regarding criminal prosecution: Type of offense; Presence of violence or serious threat (yes or no); Existence of a sentence (yes or no), Existence of a conviction (yes or no), Existence of an acquittal (yes or no); Defendant’s imprisonment (yes or no); Substitution by a non-custodial sentence (yes or no); (b) Regarding criminal executions: The case relates to a sanction (yes or no); Type of sanction; Regime progression (yes or no); Progression to which regime (when there was progression); e (c) Other information: Defendant’s imprisonment in any case (yes or no); Possibility of identifying the individual in the e-SAJ system (yes or no). These details for each individual were stored in digital spreadsheets, in addition to the data previously stored from the initial collection (Galinari & Bazon, 2021 ).

Based on this information, three measures of recidivism/reoffending were used: (1) Existence of a criminal prosecution against the individual; (2) Application of imprisonment as a sanction; (3) Existence of a criminal prosecution for crimes involving violence or serious threat. These measures were accounted for as follows: (1) having any criminal prosecution against oneself; (2) receiving a custodial sentence in any regime (i.e. closed, semi-open, or open), determined by a first-instance conviction, when the sentence was not replaced by right-restricting penalty (e.g. Community Service); (3) being prosecuted for crimes involving violence or serious threats (robbery, domestic violence, kidnapping and unlawful imprisonment, rape, aggravated murder).

Data analysis. Initially, descriptive statistics were used to characterize the overall sample and by profile concerning the variables of interest (Race, Criminal Prosecution, Imprisonment, Crime with Violence or Serious Threat). Comparative analyses were then conducted using the chi-square test to assess whether there were distinct distributions between profiles for the studied variables. Following this, logistic regression analyses were performed, reporting Odds Ratios, to determine if belonging to a specific profile during adolescence and/or “race” influenced the likelihood of the participant being involved in the criminal justice system in adulthood. Pseudo r² was calculated for the adjusted models to compare them. Unlike linear regressions, pseudo r² does not represent the proportion of variance explained by the model but rather the improvement of the model compared to a null model (Hemmert et al., 2018 ). The analyses were conducted using STATA version 13.

Ethical Considerations

The study was approved by the Comitê de Ética da Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, belonging to Universidade de São Paulo, under CAAE No. 77903617.5.0000.5407.

Results

Table 1 presents the results for individuals whose data could be checked in the e-SAJ system ( n = 350). At least one criminal case was identified for 155 individuals. Of these, 102 were sentenced to imprisonment (meaning they were incarcerated). The overall rate of “criminal cases” was 44%, the imprisonment rate was 29%, and the rate of cases involving crimes with violence or serious threat was 15%.

Regarding the first criminal case, it is noteworthy that out of the 155 individuals prosecuted, 104 were convicted at first instance (found guilty); two were acquitted; for the remaining individuals, no judgment had been rendered by the time of data collection (a judicial decision had not been made on the case), which represents a conviction rate of 98.1%. In terms of the number of criminal cases: 56 individuals had a second criminal case. 15 had a third case. Eight had a fourth case. Three had a fifth criminal case. Table 1 also presents the general distribution (percentage in the total sample; % total) and specific distribution for each subgroup/profile of the existence of criminal cases, cases with imprisonment, crimes with violence or serious threat, and race (for race: n = 315, as there were no race data for all participants).

Table 1
Characterization of the Total Sample and by Group/Profile of the Distribution of Race and Indicators of Involvement with the Criminal Justice System

Variable	Total	% Total	Profile 1	Profile 2	Profile 3	Profile 4	ꭓ²	p
Criminal case/records
	155	44%	37%	33%	45%	50%	6.723	0.10
Criminal case/records involving imprisonment
	102	29%	20%	20%	30%	35%	6.870	0.07
Crime involving violence or serious threat
	52	15%	12%	7%	14%	19%	5.0993	0.16
Race *
White	92	29%	31%	26%	36%	26%	16.673	0.054
Brown	177	56%	55%	65%	46%	59%
Black	46	15%	14%	9%	59%	15%

• ^*
  Note.^*n = 315.

The chi-square test (ꭓ²) revealed no significant differences between the profiles regarding criminal records, imprisonment, crimes involving violence or serious threat, and race ( p > 0.05). Table 2 presents the results of the logistic regression analyses, reporting odds ratios for each of the subgroups/profiles and race concerning different types of involvement with the adult Criminal Justice system, using Profile 1 (closest to the normative population) and the white race as references. This table includes two models: the unadjusted model, where the variables “profile” and “race” were analyzed separately; and the adjusted model, which considers both variables to avoid confounding.

In the unadjusted model, Profile 4 was associated with a significantly higher chance of criminal recidivism compared to Profile 1, with 1.53 times more likelihood of being criminally prosecuted and 1.65 times more likelihood of being sentenced to imprisonment. Race was also predictive of prosecution in the criminal justice system, with Black individuals having 2.04 times higher odds of being criminally prosecuted compared to white individuals. Both profile and race did not show significant results for higher or lower chances of being prosecuted for crimes involving violence or serious threat. In the adjusted model, however, the variables “profile” (Profile 4 compared to Profile 1) and “race” (Black compared to White) showed decreased odds ratios, with confidence intervals falling below 0.99 and p > 0.05, indicating a lack of evidence for the predictive capability of these variables for involvement with the adult criminal justice system.

Table 2
Results of the Logistic Regression Analysis

	Unadjusted Model				Adjusted Model
	OR	95% IC		p	OR	95% IC		p
Criminal Prosecution
Profile 1	Reference				Reference
Profile 2	0.58	0.31	1.06	0.08
Profile 3	1.05	0.64	1.72	0.81
Profile 4	1.53	1.00	2.35	0.04	1.52	0.99	2.35	0.054
White	Reference				Reference
Brown	1.08	0.68	1.72	0.72	1.02	0.64	1.64	0.905
Black	2.04	1.03	4.06	0.04	1.97	0.99	3.93	0.053
Criminal prosecution involving imprisonment
Profile 1	Reference				Reference
Profile 2	0.57	0.28	1.16	0.12
Profile 3	1.07	0.62	1.82	0.79
Profile 4	1.65	1.04	2.63	0.03	1.59	0.99	2.54	0.051
White	Reference				Reference
Brown	1.49	0.89	2.51	0.12	1.41	0.83	2.38	0.19
Black	1.72	0.82	3.58	0.14	1.65	0.79	3.45	0.182
Crime involving violence or serious threat
Profile 1	Reference				Reference
Profile 2	0.36	0.10	1.22	0.10
Profile 3	1.02	0.49	2.13	0.94
Profile 4	1.69	0.89	3.19	0.10	1.66	0.87	3.15	0.118
White	Reference				Reference
Brown	1.20	0.59	2.44	0.60	1.13	0.55	2.30	0.737
Black	1.44	0.54	3.85	0.45	1.37	0.51	3.68	0.524

• Note. Pseudo r² values for the adjusted models, respectively: 0.0172, 0.0165, 0.0115.

Discussion

In summary, this study aimed to: (1) Describe and compare recidivism rates among 350 adult men who had gone through the juvenile system in the countryside of São Paulo; (2) Evaluate the prediction of recidivism according to psychosocial profiles; (3) Assess the influence of race on recidivism. The study focused on both the impact of psychosocial factors on the persistence of delinquent/criminal behavior and the potential racial biases in the justice system, aiming to assess the likelihood of recidivism/reoffending among individuals into adulthood.

In the total sample, 44% of individuals had been criminally prosecuted, 29% were sentenced to imprisonment, and 15% were prosecuted for crimes involving violence or serious threats. The rate of criminal prosecution, 44%, is slightly higher than the rate found by Sapori et al. ( 2020 ) in a Brazilian study, which estimated a rate of 30.1% based on data from new police records. Given this study as a benchmark, a lower rate was expected in the current study, as it used only data from the judiciary system, and not every new police record results in an indictment by the Public Prosecutor’s Office. An indictment only occurs if there is sufficient evidence indicating the materiality and authorship of the crime, according to the Code of Criminal Procedure (Decreto-Lei No. 3,689, 1941 ). The observed higher rate in the current study may be due to differences in the systems involved, specific to the management of each state, since the study by Sapori et al. ( 2020 ) was conducted in Minas Gerais and this study in São Paulo. It is worth noting that both studies dealt with data from youths who had been judicialized and had their liberty restrained (provisionally or permanently) during adolescence.

Regarding the prediction analyses of recidivism/reoffending for the psychosocial profiles identified and described by Galinari and Bazon ( 2021 ), the results from the unadjusted model indicated that individuals with Profile 4 characteristics, compared to those with Profile 1, had 1.53 times the chance of being criminal prosecuted and 1.65 times the chance of being sentenced to prison. These results are consistent with one of the study’s hypotheses: that Profile 4, which includes individuals with a distinctive pattern of delinquent/criminal behavior and indicators of greater psychosocial problems/needs, would be more exposed to risk factors associated with the persistence of delinquent behavior beyond adolescence, and therefore have a higher likelihood of recidivism.

The other profiles (2 and 3) did not show significant differences when compared to Profile 1. Higher recidivism rates were expected for Profile 3, in addition to Profile 4, as both present similar patterns of delinquent/criminal behavior, although Profile 3 refers to adolescents with significant social problems but fewer personal problems (compared to Profile 4). Thus, the risk hypothesis of this study, based on Galinari and Bazon ( 2021 ), was partially confirmed, being effective only for Profile 4. The results align with those found in similar studies for profiles equivalent to Profile 4 (Campbell et al., 2019 ; Cox et al., 2018 ), reinforcing that the accumulation of social and personal risk factors is a suitable predictor of continued criminal behavior (Basto-Pereira & Farrington, 2022 ).

Moreover, it is possible that Profile 3 refers to adolescents more prone to committing offenses compared to the normative population, but not necessarily to the persistence of delinquent behavior beyond adolescence. This is consistent with the findings of Basto-Pereira and Farrington ( 2022 ), which showed that problems concentrated in a single risk dimension (in this case, social) were more strongly associated with the initial manifestation of behavior (one set), while problems across multiple dimensions better predicted the persistence of delinquent behavior.

It should also be considered that this result might be due to the variables used by Galinari and Bazon ( 2021 ) in their Latent Class Analysis for identifying the profiles. The authors prioritized variables more directly related to psychosocial intervention needs, rather than those with greater predictive power for future behavior. This is because the initial objective of the clustering was not to quantify levels of risk for persistence in delinquent/criminal behavior, but to identify different clusters in terms of psychosocial needs, considering the importance of socio-educational programs tailored to adolescents (Galinari & Bazon, 2021 ).

The finding that only one profile/subgroup showed higher recidivism rates, while the other subgroups did not show significantly different chances, also aligns with Moffitt’s ( 2018 ) proposal. The author indicates that few adolescents exhibit a life-course persistent pattern of delinquent behavior, highlighting that most show a pattern characterized by “desistance” from criminal activity by the end of adolescence or early adulthood (adolescence-limited). Thus, it is assumed that the sample of adolescents studied encompasses two broad groups: one showing a persistent pattern extending beyond the early years of adulthood (Profile 4), and another showing a non-persistent pattern, limited to adolescence (Profiles 1, 2, and 3).

At first glance, one might consider that the findings suggest that the socio-educational measures applied were, to some extent, beneficial in helping to interrupt delinquent behavior for those individuals for whom no recidivism data was found by the time of data collection (56%). However, literature provides evidence that most adolescents tend to exhibit adolescence-limited delinquency, even without institutionalization (Campbell et al., 2019 ; Moffitt, 2018 ). In this study, it was specifically hypothesized that individuals in Profiles 1 and 2 would have low chances of recidivism due to their association with a less problematic behavior pattern and fewer psychosocial issues. In this case, individuals who were institutionalized (83% of the sample) might have benefited from less severe measures, as recommended by Section VII of Article 35 of Law No. 12,594, January 18, 2012 (Law establishing the National System of Socio-Educational Assistance). Additionally, evidence suggests that the institutionalization of judicialized adolescents may be associated with less-than-ideal conditions, such as increased emotional stress and symptoms of anxiety and depression (da Silva et al., 2021 ). This hypothesis is even stronger when considering that Profile 1 comprised youths with characteristics very close to those of the normative adolescent population, suggesting a possible natural or spontaneous cessation of delinquent behavior (Moffitt, 2018 ).

Regarding crimes characterized by violence or serious threats, none of the profiles showed significant effects in increasing or decreasing the likelihood of involvement in these types of offenses. It was assumed that individuals whose data allowed for the characterization of Profiles 3 and 4 would be at greater risk of engaging in violent crimes, as these profiles were characterized by the manifestation of such offenses during adolescence. However, this was not observed, indicating the need for more focused studies to analyze specific risk factors related to involvement in violence that were not investigated by Galinari and Bazon ( 2021 ).

Regarding racial issues, it is important to note that the racial distribution in the sample differs from that found in the Brazilian population. According to the Instituto Brasileiro de Geografia e Estatística [IBGE] ( 2019 ), the population is composed of 42.7% self-declared white, 46.8% brown (multiracial individuals), 9.4% black, and 1.1% yellow or indigenous individuals. In the sample, the percentage of individuals identified as white is considerably lower, constituting only 29% of the sample, while the percentages of brown and black individuals are higher than in the Brazilian population, comprising 56% and 15% of the sample, respectively. Thus, there is a higher concentration of non-white individuals in the sample compared to the general population. This finding itself raises concerns about potential selectivity by control agencies regarding juvenile delinquency, due to the overrepresentation of non-white individuals in the socio-educational system (Sinhoretto & Lima, 2015 ). However, it is important to consider that this result may be due to differences in the method of collecting racial data, as IBGE uses self-declaration, whereas the study used heteroidentification data.

Regarding the likelihood of being criminally prosecuted after the age of 18, the odds ratio results for race are alarming. Black individuals, compared to white individuals, had twice the chance (2.04), indicating that race is a significant variable in predicting criminal prosecution. Since race, in itself, has no causal link to criminal behavior, it is necessary to consider the existence of a bias in the selection of individuals who are criminally prosecuted, reinforcing findings from other studies (Campbell et al., 2019 ; Cox et al., 2018 ). This selectivity is likely explained by stereotypes prevalent in the Brazilian justice system (Sinhoretto & Lima, 2015 ; Soares & Ribeiro, 2018 ), leading certain individuals to have a higher chance of being criminally prosecuted based on their race. It is important to emphasize that involvement in the criminal system leads to stigmatization, distancing individuals from social opportunities, such as employment. This creates a vicious cycle that exacerbates social vulnerability and marginalization (Soares & Ribeiro, 2018 ).

Furthermore, the conviction rate (among individuals for whom a sentence had been pronounced at the time of data collection) was 98.1%, with only 2 individuals being acquitted compared to 104 convicted. This data may indicate that once an individual is prosecuted, “exoneration” is very rare, especially for populations in social and/or racial vulnerability, as proposed by Soares and Ribeiro ( 2018 ). In this context, the criminal justice system deepens the existing inequalities among individuals by focusing on penalizing more vulnerable social strata (Sinhoretto & Lima, 2015 ; Soares & Ribeiro, 2018 ), such as the black population, according to the results of this study.

It is certain that in the adjusted model, the evidence regarding the predictive power of the variables “profile” and “race” was not clear. This data may indicate confounding effects between variables or that the sample size is insufficient for fitting adjusted models. Therefore, studies with larger samples should be conducted. Nevertheless, the data from the unadjusted model provide important information about variables affecting criminal prosecution of individuals.

Overall, this study reinforces evidence already established in the scientific literature that juvenile delinquency can be characterized by both continuity and discontinuity, depending on the issues that characterize the adolescent transitioning to adulthood (Moffitt, 2018 ). This result is of academic and practical relevance because it emphasizes the importance of identifying which characteristics/needs of adjudicated youth are associated with higher chances of re-engagement with the Criminal Justice system during the transition to adulthood. With this data, it is possible to optimize the use of resources in the socio-educational system, focusing on effective interventions for subgroups with greater psychosocial needs to prevent the transition from juvenile delinquency to adult criminality.

Another contribution of the study is to shed light on reflections about racism and its possible reflection in the Criminal Justice system. Indications were found, in line with other studies, that race is a factor contributing to increased chances of criminalization of individuals (Campbell et al., 2019 ; Cox et al., 2018 ; Soares & Ribeiro, 2018 ). The empirical demonstration of penal selectivity within the Brazilian sociocultural context is academically relevant as it corroborates existing literature and highlights structural racism in Brazil.

Several limitations of the study can also be noted: A limitation pertains to the fact that the data used by Galinari and Bazon ( 2021 ) for profile identification were collected at two different time points ( n = 130 in 2015; n = 270 in 2018). This implies different time intervals between the data collection for identifying and characterizing the profiles of adjudicated adolescents and the data collection conducted in this study through the e-SAJ system. However, it is important to emphasize that the results from analyzing the data together and separately (considering data from 2015 and 2018 independently) were very similar, providing confidence in presenting the results from the combined analysis.

Another aspect to consider is that the review of Criminal Justice records was conducted only within the context of the state of São Paulo, meaning that data from other states were not accessed. Since the data used by Galinari and Bazon ( 2021 ) for describing the profiles referred to adolescents serving socio-educational measures in a municipality in the interior of São Paulo, and thus living in the state of São Paulo at the time, it is assumed that focusing solely on information from this state would sufficiently encompass the sample data.

It is important to note that this study, by examining delinquency recidivism/reoffending solely through official data, more accurately reflects the functioning of control agencies. Future studies should aim to integrate both official and self-reported data to overcome this methodological limitation, particularly since self-reported data alone may also underestimate the true extent of the phenomenon (Gomes et al., 2022 ).

In summary, this study highlights the importance of understanding which psychosocial profiles are associated with higher chances of persistence in delinquent behavior and identifies a racial bias present in the Brazilian criminal justice system, indicating structural racism. Finally, it is hoped that the findings of this study will contribute to the development of more effective and targeted interventions within socio-educational systems and also help to uncover racial biases within the Brazilian justice system.

References

• Associação Brasileira de Empresas de Pesquisa. (2016). Critério Brasil 2015 e atualização da distribuição de classes para 2016. https://www.abep.org/criterio-brasil
  » https://www.abep.org/criterio-brasil
• Basto-Pereira, M., & Farrington, D. P. (2022). Developmental predictors of offending and persistence in crime: A systematic review of meta-analyses. Aggression and Violent Behavior, 65, 1-11. https://doi.org/10.1016/j.avb.2022.101761
  » https://doi.org/10.1016/j.avb.2022.101761
• Campbell, C. A., Miller, W., Papp, J., Barnes, A. R., Onifade, E., & Anderson, V. R. (2019). Assessing intervention needs of juvenile probationers: An application of latent profile analysis to a risk–need–responsivity assessment model. Criminal Justice and Behavior, 46(1), 82-100. https://doi.org/10.1177/0093854818796869
  » https://doi.org/10.1177/0093854818796869
• Cox, S. M., Kochol, P., & Hedlund, J. (2018). The exploration of risk and protective score differences across juvenile offending career types and their effects on recidivism. Youth Violence and Juvenile Justice, 16(1), 77-96. https://doi.org/10.1177/1541204016678439
  » https://doi.org/10.1177/1541204016678439
• Cuervo, K., Villanueva, L., & Basto-Pereira, M. (2020). Prediction of youth and adult recidivism among Spanish juveniles involved in serious offenses. Criminal Justice and Behavior, 47(4), 399-418. https://doi.org/10.1177/0093854819897282
  » https://doi.org/10.1177/0093854819897282
• da Silva, M. C., Cruz, A. P. M., & Teixeira, M. O. (2021). Depression, anxiety, and drug usage history indicators among institutionalized juvenile offenders of Brasilia. Psicologia: Reflexão e Crítica, 34(1), 17. https://doi.org/10.1186/s41155-021-00184-x
  » https://doi.org/10.1186/s41155-021-00184-x
• Decreto-Lei No. 3.689, de 3 de outubro de 1941 (1941, 13 de outubro). Código de Processo Penal [Criminal Procedure Code]. Diário Oficial da União, seção 1.
• Galinari, L. S., & Bazon, M. R. (2021). Criminal behavior and psychosocial risk factors in Brazilian adolescent offenders: An exploratory latent class analysis. International Journal of Environmental Research and Public Health, 18(19), 10509. https://doi.org/10.3390/ijerph181910509
  » https://doi.org/10.3390/ijerph181910509
• Gomes, H. S., Farrington, D. P., Krohn, M. D., Cunha, A., Jurdi, J., Sousa, B., Morgado, D., Hoft, J., Hartsell, E., Kassem, L., & Maia, A. (2022). The impact of modes of administration on self-reports of offending: Evidence from a methodological experiment with university students. Journal of Experimental Criminology, 20(1), 207-227. https://doi.org/10.1007/s11292-022-09531-z
  » https://doi.org/10.1007/s11292-022-09531-z
• Guimarães, A. S. A. (2011). Raça, cor, cor da pele e etnia [Race, color, skin color and ethnic group]. Cadernos De Campo, 20(20), 265-271. https://doi.org/10.11606/issn.2316-9133.v20i20p265-271
  » https://doi.org/10.11606/issn.2316-9133.v20i20p265-271
• Hemmert, G. A. J., Schons, L. M., Wieseke, J., & Schimmelpfennig, H. (2018). Log-likelihood-based Pseudo-R2 in Logistic Regression: Deriving sample-sensitive benchmarks. Sociological Methods & Research, 47(3), 507-531. https://doi.org/10.1177/0049124116638107
  » https://doi.org/10.1177/0049124116638107
• Instituto Brasileiro de Geografia e Estatística. (2019). Conheça o Brasil – População: Cor ou Raça [Getting to know Brazil – Population: Color or Race]. IBGE Educa. https://educa.ibge.gov.br/jovens/conheca-o-brasil/populacao/18319-cor-ou-raca.html
  » https://educa.ibge.gov.br/jovens/conheca-o-brasil/populacao/18319-cor-ou-raca.html
• Johnson, E., & Moyer, B. (2021). Ethnic differences in the risk assessment of chronic violence review of the literature. Aggression and Violent Behavior, 58, 101553. https://doi.org/10.1016/j.avb.2021.101553
  » https://doi.org/10.1016/j.avb.2021.101553
• Komatsu, A. V., Wenger, L., Costa, R. C. S., & Bazon, M. R. (2019). Factores protectores en adolescentes infractores: Un estudio tipológico [Protective factors among juvenile offenders: A typological study]. International e-Journal of Criminal Sciences, (14), 7. https://ojs.ehu.eus/index.php/inecs/article/view/21304
  » https://ojs.ehu.eus/index.php/inecs/article/view/21304
• Lei No. 12.594, de 18 de janeiro de 2012. (2012, 19 de janeiro). Institui o Sistema Nacional de Atendimento Socioeducativo (Sinase), regulamenta a execução das medidas socioeducativas destinadas a adolescente que pratique ato infracional; e altera as Leis Nos. 8.069, de 13 de julho de 1990 (Estatuto da Criança e do Adolescente) [Institutes the National System of Socio-educational Assistance (Sinase), regulates the execution of socio-educational measures for adolescents who have committed an offence, and amends Law No. 8,069, of July 13, 1990 (Statute of the Child and Adolescent)]. Diário Oficial da União, seção 1.
• Moffitt, T. E. (2018). Male antisocial behaviour in adolescence and beyond. Nature Human Behaviour, 2, 177-186. https://doi.org/10.1038/s41562-018-0309-4
  » https://doi.org/10.1038/s41562-018-0309-4
• Mulder, E., Brand, E., Bullens, R., & van Marle, H. (2019). Toward a classification of juvenile offenders: Subgroups of serious juvenile offenders and severity of recidivism. International Journal of Offender Therapy and Comparative Criminology, 63(6), 819-836. https://doi.org/10.1177/0306624X10387518
  » https://doi.org/10.1177/0306624X10387518
• Sapori, L. F., Caetano, A. J., & Santos, R. F. (2020). A reiteração de atos infracionais no Brasil: O caso de Minas Gerais [Juvenile offender recidivism in Brazil: The case of Minas Gerais state]. Revista Direito GV, 16(3), e1975. https://doi.org/10.1590/2317-6172201975
  » https://doi.org/10.1590/2317-6172201975
• Sinhoretto, J., & Lima, R. S. (2015). Narrativa autoritária e pressões democráticas na segurança pública e no controle do crime [The authoritarian narrative and democratic pressures on public safety and crime control]. Contemporânea: Revista de Sociologia da UFSCar, 5(1), 119-141. https://www.contemporanea.ufscar.br/index.php/contemporanea/article/view/299
  » https://www.contemporanea.ufscar.br/index.php/contemporanea/article/view/299
• Soares, F. C., & Ribeiro, L. M. L. (2018). Rotulação e seletividade policial: Óbices à institucionalização da democracia no Brasil [Labeling and police selectivity: obstacles to the institutionalization of democracy in Brazil]. Estudos Históricos (Rio de Janeiro), 31(63), 89-108. https://doi.org/10.1590/s2178-14942018000100006
  » https://doi.org/10.1590/s2178-14942018000100006

• How to cite this article:

  Georgini, G. A., Galinari, L. S., & Bazon, M. R. (2024). From juvenile justice to criminal justice: Race and psychosocial profiles as predictors. Paidéia (Ribeirão Preto) , 34 , e3425. doi:https://doi.org/10.1590/1982-4327e3425

• Support:

  This study received financial support from the National Council for Scientific and Technological Development (CNPq) through a PIBIC scholarship (Process: 129601/2020-9). 129601/2020-9).

Edited by

Associate editor:

Vanessa Barbosa Romera Leme

Publication Dates

• Publication in this collection
  22 Nov 2024
• Date of issue
  2024

History

• Received
  01 Apr 2024
• Accepted
  25 June 2024
• Reviewed
  24 June 2024