Sample characteristics

Estudo de Socialização Legal de São Paulo¹ consists of a longitudinal survey carried out over four years (2016-2019) with an initial sample of 800 participants aged 11 and over and divided between 400 girls and 400 boys. Due to the size, diversity, and socioeconomic inequality of the city of São Paulo, the research took some important methodological precautions to cover the economic, educational, and developmental differences that could affect the students’ responses from different social backgrounds.²

The initial sample was obtained from the population of students enrolled in the sixth year of elementary school in the city of São Paulo. The 2014 National School Census served as a resource at the time of building the sample. This source indicated that 59% of elementary school students came from public schools and 41% from private schools in São Paulo (Instituto Nacional de Estudos e Pesquisas Educacionais Anísio Teixeira [Inep], ²⁰¹⁴). In order to meet the proportion of students and the geographic distribution of schools in different city regions, we used the Probability Proportional to Size (PPS) method, which generated a list of public and private schools eligible to participate in the study.

It is important to emphasize that all procedures involving minors followed the ethical standards of Resolução n. 466/²⁰¹² [Resolution No. 466/²⁰¹²] of Conselho Nacional de Saúde [National Health Council]. The study was officially approved on September 24, 2015, by the Comitê de Ética em Pesquisa em Seres Humanos [Ethics Committee for Research on Human Beings] of Escola de Artes, Ciências e Humanidades [School of Arts, Sciences and Humanities] (EACH) of Universidade de São Paulo (USP). Only students with a Consent Form signed by their parents were able to participate in the research. In addition, participants were fully informed about their role and anonymity rights and had the option to withdraw from the research at any time.

The company Inteligência em Pesquisa e Consultoria (Ipec) (formerly Ibope) was responsible for data collection during the four waves under the direction of NEV-USP. To conduct each interview, the researcher read aloud the questions and all the response options to the participants. Surveyors then entered the results into the software Survey-To-Go,³ used to minimize errors and omissions, in addition to facilitating the systematization of data.

The study’s first wave (hereafter T1) interviewed 800 individuals between May and September 2016 (50% girls and 50% boys; mean age 11 years). The second wave (T2) took place between August and November 2017 and obtained 742 interviews (loss of only 7.12%; 50.13% of girls; mean age of 12 years). The third wave (T3) took place between August and November 2018 with 723 completed interviews (9.5% loss; 49.5% female; mean age of 13 years). The fourth wave (T4) took place between August and December 2019 with 701 subjects. The final attrition rate compared to the first wave was 12%; 356 boys, 346 girls; mean age 14 years.

The study demographic categories correspond to those officially used by Instituto Brasileiro de Geografia e Estatística [Brazilian Institute of Geography and Statistics] (IBGE) in the Censo Demográfico [Demographic Census] - white, Black, pardo [mixed race], yellow (Asian), and indigenous. According to the last census carried out in 2010, the Brazilian population was composed of 47.73% white, 43.13% pardo, 7.61% Black, 1.09% Asian and 0.43% of indigenous people. In the city of São Paulo, 60.6% declared themselves white, 30.5% pardo, 6.5% Black, 2.2% Asian and 0.1% indigenous (Instituto Brasileiro de Geografia e Estatística [IBGE], ²⁰¹¹). The racial composition of the sample of 800 research participants had the following distribution: 47% white, 27% Pardo, 11% Black, and 15% other or none specifically (Asians and indigenous people were grouped in the category “others” due to the low number of each group). Therefore, the final study sample consists of 47% whites and 53% non-whites.

Analysis variables

The study’s questionnaire is a comprehensive instrument that aims to explore the experiences of children and adolescents in different spheres (family, school, neighborhood) as well as their interactions with authorities (legal and non-legal). In general, the research investigates the formation of attitudes and beliefs towards norms and laws, along with the legitimacy and trust attributed to authorities. The study also measured the exposure to violence, which can affect the perceptions of justice and the democratic system.

For analytical purpose, four variables were selected involving “rules-violating behavior,” “perception of laws,” “legal cynicism,” and “belief in justice” (the phrasing of these questions are in the following tables). The first three variables are based on research carried out in the United States by the Legal Socialization Laboratory of the University of New Hampshire (^{Cohn et al., 2010}; ^{Rebellon et al., 2012}; ^{Trinkner et al., 2012}; ^{Van Gundy et al., 2011}). The variable about belief in justice was based on the works of ^{Dalbert (1999)}, ^{Pimentel et al. (2010)}, and ^{Thomas (2018)}. Finally, we considered the sex (male and female) as the main control variable in the analysis.

Analysis strategy

Data analysis took place in several stages. In order to understand the formation of democratic values, we first analyzed the frequency distributions in each selected variable, which allowed us to carefully examine variation within each question. Using statistical software, Stata (version 14), we calculated the frequency distributions of the selected questions (tables 1, 2, 3, 4, and 5). Next, we identified significant differences between the responses and the demographic characteristics of the students. We created cross-tabulations and significance tests (χ²) to examine differences between boys and girls (Table 6). Some results show that the responses varied significantly between sexes, which suggests differing socialization processes by gender.

A second analysis used correlation tests to examine the relationship between procedural justice and authority legitimacy. The correlations focused on democratic values, belief in justice, rule-violating behavior, perception of laws, and legal cynicism (Table 7).

Analysis of results

The four waves of the research followed the same students from 11 to 14 years of age. Tables 1, 2, 3, 4, and 5 show the univariate frequencies of responses involving rule-violating behavior, perception of the laws, belief in justice, legal cynicism, legitimacy, and procedural fairness of parents and teachers.

Over four years, respondents were consistent in their responses about certain types of rule violations in the early years of adolescence (Table 1). For example, they overwhelmingly reported that they avoided stealing objects from someone or a store without paying for them (T1 “never” = 99% and T4 “never” = 98%). Likewise, vandalizing or destroying other people’s property was an infrequent activity among respondents (T1 “never” = 97% and T4 “never” = 95%).

The percentage of students who reported using drugs increased with age, but the majority reported abstaining from this behavior (T1 “never” = 100% and T4 “never” = 96%). Other rule- -violating behaviors reported by respondents included buying counterfeit products (T1 “never” = 87% versus T4 “never” = 83%) and engaging in academic dishonesty by copying a classmate’s lesson or cheating on the exam (“never” in T1 = 78% versus T4 “once”, “few” and “many times” = 63%).

Table 2 shows frequency distributions of items related to the perception of law. The results show greater variability in responses over time, with changes mainly in relation to the country’s laws. Paradoxically, the belief that laws were beneficial to the country increased with age. At T1, 43% agreed that laws were beneficial, a feeling that increased to 73% four years later (T4), when respondents were 14 years old and came to understand the laws as important values. Student agreement also increased over time regarding fines and/or imprisonment for those who disobey the laws (from 88% of “agree” and “completely agree” in T1 to 94% in T4).

Other measures showed decreases in agreement with laws over time. There was a slight reduction in agreement that laws must be obeyed even when people disagree with them (from 93% of “agree” and “completely agree” in T1 to 88% in T4). Other small decreases in agreement observed over time are included in questions that asked whether some laws could be broken, whether laws would be the same for all and whether some people were above the law. The most considerable disagreement was on the perception that laws protect people (from 97% “agree” and “completely agree” in T1 to 85% in T4) and whether laws could be changed (from 77% in T1 to 61 % in T4).

Respondents were also asked about their beliefs in justice. The results presented in Table 3 indicate that students remained consistent in their views over time. Most respondents agreed (74% at T2 and 75% at T4) that most of what happened in their lives was fair. This consistency and direction were sustained through questions that asked whether they were treated fairly (71% at T2 and 75% at T4 answered “agree” or “completely agree”).

There was more variation in respondents’ global views of justice. Although they reported a positive and fair view of their personal life, they were more skeptical about the world around them. Skepticism also increased as they got older. For example, 71% thought, at least a little, that justice prevailed over injustice when they were 12 years old (T2); perception that dropped to 59% at 14 years old (T4). Interestingly, less than half (45%) agreed that the world was a fair place when they were 12 years old (T2), which dropped to about a quarter (28%) two years later.

A final set of frequencies measured respondents’ legal cynicism. The results are shown in Table 4. Over time, the students remained consistent in relation to some aspects of the legal system. Similar to the questions related to rule-violating behavior, the vast majority disagreed with the possibility of people stealing products in stores without paying for them (99% in T1 and T4), breaking or damaging other people’s properties (99% in T1 and 98% in T4), deliberately beating someone up (98% in T1 and 98% in T4), using drugs (99% in T1 and 97% in T4), breaking the law (93% in T2 and 94% in T4). We observed that respondents were more willing to accept other people buying counterfeit products (93% at T1 and 86% at T4) and cheating at school (93% at T1 and 82% at T4) as they aged.

This means that, over time, students became personally involved in rule-violating behavior and also justifying other’s misbehavior. Interestingly, respondents became more cautious in some areas over time. In T2, 66% of adolescents disagreed that there were no right or wrong ways to make money. The percentage of adolescents who disagreed with the same statement increased to 79% two years later.

Finally, Table 5 presents descriptive statistics regarding the legitimacy and procedural justice of parents and teachers. In general, respondents maintained positive attitudes towards authority figures. Those who perceived their parents as fair also understood them as legitimate authorities. The same occurred in relation to teachers, although at lower levels.

Table 6 presents the results of four variables we analyzed using Chi-square tests (χ²) to understand differences in results between boys and girls. Regarding rule-violating behavior, girls and boys tend to behave similarly. However, two items show a significantly differences. The number of girls who reported (or admitted) having purchased counterfeit products was higher than that of boys (χ² (df = 3) = 9.80; p <0.05; n = 702). In addition, more girls than boys reported copying lessons or cheating on the test (χ² (df = 3) = 8.70; p <0.05; n = 701).

In terms of perception of laws, girls tend to have a more negative (or more critical) view than boys (χ² (df = 3) = 15.60; p <0.001; n = 699). More boys than girls agree that laws serve to protect people (χ² (df = 3) = 11.13; p <0.05; n = 701). More boys understand that laws can be changed (χ² (df = 3) = 9.20; p <0.05; n = 697), and finally, more boys think that laws represent important normative values (χ² (df = 3) = 10.60; p <0.05; n = 696).

Regarding the belief in justice, there were also significant differences relate to gender. Boys tended to have a more positive view of justice than girls. For example, more boys than girls agree that people get what they deserve (χ² (df = 3) = 10.89; p <0.05; n = 700). More boys than girls understand that the world is a fair place (χ² (df = 3) = 11.69; p <0.01; n = 701). Moreover, more boys than girls think that adults are fair in making important decisions (χ² (df = 3) = 9.42; p <0.05; n = 699).

Finally, we measured the level of legal cynicism. As with other categories, girls tended to have a more cynical view of laws. More girls than boys agree that it is okay for people to use drugs (χ² (df = 3) = 8.34; p <0.05; n = 701). More girls than boys understand that it is okay to do anything as long as it does not physically hurt someone (χ² (df = 3) = 17.95; p <0.001; n = 701). In addition, more girls than boys think that people should live in the moment without thinking about the future (χ² (df = 3) = 11.31; p <0.01; n = 698).

The final analysis compared the strength and direction of the relationships between parents’ and teachers’ procedural justice, legitimacy, and perceptions of democratic values. Table 7 presents the correlation coefficients (Pearson’s r) between the highlighted variables. Results show moderate and strong relationships between the variables. Respondents tended to view parents and teachers alike, that is, those who perceive parents as fair also viewed teachers are fair. The correlation between the legitimacy of parents and teachers represents one of the strongest correlations verified in the analysis (r = 0.68; p <0.001).

Looking at parental procedural justice, there are positive associations in the perception of laws (r = 0.18; p <001) and belief in justice (r = 0.34; p <001). These trends follow the same direction between the teachers’ perception of procedural justice, the perceptions about the laws (r = 0.30; p <001) and the belief in justice (r = 0.36; p <001). In other words, respondents who perceive teachers and parents as fair tend to have positive views of laws. Procedural justice measures were negatively associated with attitudes towards rule-violating (r = -.22; p <001) and legal cynicism (r = -.12; p <01). We interpret this as children and adolescents with fair parents and teachers break the rules less often and are less cynical about the legal order.

There were similar results regarding the legitimacy of parents and teachers. Students who perceive parents as legitimate authorities are more likely to form favorable views towards laws (r = 0.44; p <001) and justice (r = 0.42; p <001). On the other hand, parental legitimacy and rule- -violating behavior were negatively correlated (r = -.22; p <001). It means that students are less likely to break the rules by legitimizing parental authority.

Interestingly, respondents who saw their parents as legitimate had a more cynical view of laws, although this association is statistically very weak (r = 0.08; p <05). The results are similar in relation to the teacher’s legitimacy. Generally, students who legitimized teacher authority were less involved in rule-violating behavior and simultaneously had more favorable perceptions of laws, justice, and democratic values in general.