Method

This is a quasi-experimental study (^{Fife-Schaw, 2010}) that uses quantitative and qualitative data analysis strategies.

Results

Analysis of the justification for positions regarding police violence against the black suspect

The analysis of the corpus through lexical analysis, after reducing the words to their roots, showed a total of occurrences of 1,122 different words with 428 distinct forms, corresponding to 70 % of the corpus. A number of texts equal to 51 (ICU) was obtained, divided into 270 text segment units (ECU). Descending hierarchical classification (DHC) showed that the analysis corpus was divided into four distinct classes (table 1).

Table 1: CHD of the position in relation to the violence against the black suspect 

The semantic analysis of class 1, called Violence as justice, encompasses positions that conceive of police violence as an appropriate action given the crime committed by the suspect. The repertoires of this class consider police violence a form of punishment given a violation of the law and the ineffectiveness of justice. The word that represents the significant absence for this class is violence, which indicates agreement with the actions of the police. Examples from class 1 repertoires present police violence as justice, being an alternative to impunity, such as: “I agree with the way the police behaved, because it is a way of carrying out justice, of complying with the law” and “I agree because given what is happening in justice, the action of the police is a means of punishing the suspect”.

The content of class 2, called Legal sanction, is less present in the corpus (15.24 %) and refers to police violence as unnecessary, but defends the need for a sanction for the crime committed by the suspect. The word that represents the significant absence for this class is policeman. This absence may lead to the exemption from responsibility of this law enforcement officer regarding the actions carried out against the suspect. In this sense, arrest is indicated as an appropriate action in the police engagement. Examples of the repertoires found here are: “I disagree with aggression in situations like this, but it is necessary to take another type of measure such as socio-educational measures and imprisonment” and “The aggression was unnecessary, but an adequate measure such as prison is necessary”.

Class 3, called Justified violence, is the most representative in the corpus (28.57 %) and presents repertoires that disagree with aggressive actions as part of the police officer's duty, however, they understand aggression as a consequence of the crime committed by the suspect. The word that represents the significant absence in this class is aggression, indicating that the violence practiced is perceived as necessary. In this sense, the arguments tend to justify police violence. For example: “It is not the duty of the police to attack the man, however he must suffer the consequences of his actions for the way he committed the crime” and “The police should not attack, but the individual is paying for committing a crime”.

The semantic analysis of class 4, called Unjustified violence, refers to the lack of necessity for the use violence in the police action. This class did not show any significant absence. The justifications of this class emphasize the inappropriate conduct of the police in the unjustified use of violence, since the suspect had already been arrested. Examples of arguments about unjustified police violence are: “There was no need for violence, as the suspect had already been arrested” and “Since the suspect was already handcuffed, the police action is not justified, there was no need for violence”.

Organizing principles of positions regarding police violence against the black suspect

When analyzing the factorial plane (figure 1), it is observed that the vertical axis organized the participants' positions according to the acceptability of police violence. In the lower part of the factorial plane, the arguments presented in classes 2 (Legal sanction) and 4 (Unjustified violence) reject police violence insofar as this action was considered unnecessary and unjustifiable. In the upper part of the factorial plane, the arguments presented in classes 1 (Violence as justice) and 3 (Justified violence) defend the idea that violence is acceptable, being understood as a form of justice, more precisely as a response to criminality, given the ineffectiveness of the judicial system.

Figure 1: AFC of the position in relation to the violence against the black suspect 

The horizontal axis organizes the positions according to the arbitrariness of the police action and places the arguments referring to the abuse of authority against the arguments referring to the instrumentality of violence. On the left side of the factorial plane, it is observed that class 4 (Unjustified violence) is distinguished from the others, covering arguments that understand the police action as abusive, given that the suspect was already handcuffed. In contrast, on the right side of the factorial plane are classes 2 (Legal sanction) and 1 (Violence as justice) which present arguments associated with the instrumentality of violence. In class 2, there is disagreement regarding the use of police violence, although the need for measures necessary to combat crime is highlighted. Whereas, in class 1, police violence is seen as the instrument through which criminals are punished.

Class 3 (Justified violence) does not present a clear definition within the spectrum in the horizontal plane (Abuse of authority x Instrumentality of violence), comprising positions that perceive the police violence as a result of the suspect's action of having committed a robbery.

Analysis of the justification for positions regarding police violence against the white suspect

The analysis of the corpus through lexical analysis, after reducing the words to their roots, showed a total of occurrences of 1,114 different words with 303 distinct forms, corresponding to 73 % of the corpus. A number of texts equal to 52 (ICU) was obtained, divided into 164 text segment units (ECU). The descending hierarchical classification (DHC) showed that the analysis corpus was divided into five distinct classes (table 2).

Table 2: CHD of the position in relation to the violence against the white suspect 

The justifications of class 5, called Unnecessary violence, correspond to 22.5 % of the corpus and refer to positions that understand police violence as an improper procedure. For the participants, the police carried out an unnecessary and hasty action due to the lack of confirmation of the suspect's guilt. The word that represents the significant absence for this class is agree, referring to the participants' indignation with the injustice practiced. Class 5 examples emphasize the concern with anticipating the suspect’s culpability: “In this case, I believe that the police action was unnecessary due to uncertainty about the man being the actual suspect” and “Violent action on the part of the police was unnecessary, above all because the man was only a suspect; I believe that they should confirm responsibility for the crime”.

The semantic analysis of class 6 corresponded to 22.5 % of the corpus and was called Illegal violence, referring to the arguments that consider the police officer's action illegal. Violence is the word that represents the significant absence in this class, suggesting that the actions of the police exceeded the limits of their duty. Class 6 examples demonstrate the disagreement between the police action and the law: “I totally disagree with the aggression, the police officer should only act in accordance with the law” and “I disagree with the police action, there was no need for aggression, the suspect should just be sent to prison”.

The semantic analysis of class 7, is the one with the lowest contribution to the corpus (20 %) and was called Arbitrary violence. It refers to the use of violence in an abusive way. The word that represents the significant absence for this class is law, highlighting the arbitrariness of the police action. The examples of this class demonstrate the arbitrariness of the police action and the failure to guarantee the suspect's rights and integrity: "The police officer does not have the right to attack the individual during this action, his duty is to send him to the police station, even if the suspect is guilty” and “Assaulting the individual, even if he is guilty, is unjustifiable”.

Class 8 is the most frequent in the corpus (35 %) and was called Violence as a solution. Here police violence is seen as a way to solve the problem. The word that represents the significant absence for this class is duty, which shows that the compliance of police actions and the duty to respect the suspect's right as a citizen is not a concern for these participants. The justifications of this class admit that, although it is not the most appropriate approach, violence can function as a punishment for the assailant. As examples we have: “I partially agree, because even if violence is not the most appropriate way to handle the situation, the assailant will think twice before committing another crime” and “I agree that violence is not the best way to solve the problem, but we live in a society where laws are no longer effective, so violence works as a punishment for the assault”.

Organizing principles of positions regarding police violence against the white suspect

Figure 2: AFC of the position in relation to the violence against the white suspect 

When observing the factorial plane (figure 2), from the vertical axis, one can see that the participants' positions are organized around the notion of Justice. In this axis, it is possible to observe the contrast between the arguments about procedural justice and the arguments about corrective-punitive justice. In the upper part of the factorial plane, classes 6 (Illegal violence) and 7 (Arbitrary violence) present the idea that the actions of the police are in disagreement with the laws and social duties, demonstrating concern with the criteria and procedures employed in the use of force in an abusive way. In the lower part of the factorial plane, on the other hand, classes 5 (Unnecessary violence) and 8 (Violence as a solution) emphasize corrective-punitive justice in the police action. Class 5's arguments emphasize that the use of police violence, even if seen as a corrective measure, was unnecessary, since it was based on the suspect's presumed guilt. In class 8, in turn, the defense of violence is observed as a response to the impunity for crimes committed. The arguments of this class are organized around the idea that it is up to the police force to punish the suspect using violence.

In the analysis of the horizontal axis, the positions are organized around the illegality of the use of force and present arguments that range from illegal authoritarianism to legal authoritarianism. On the right side of the factorial plane, class 5 anchors its positions in legal authoritarianism. Although classes 6 (Illegal violence) and 7 (Arbitrary violence) are positioned closer to the center of the factorial plane, they present content qualitatively similar to that of class 5. In the discourses of these classes, the police action, in applying excessive force in the detention of the suspect, contradicts the principles of isonomy and law, ignoring practices that comply with the law. In contrast, on the left side of the factorial plane, class 8 (Violence as a solution) anchors its positions in illegal authoritarianism by defending punishment regardless of the rules and laws that guide the professional practice of police officers. In this class, it is observed that violent action by the police is perceived as a disciplinary measure capable of correcting criminal actions and preventing future events of this nature.

Discussion

The present study aimed to analyze whether the arguments used to accept or reject police violence are influenced by the skin color of the victims of this type of procedure. To this end, a scenario was used in which police violence was committed against an individual suspected of having committed the crime of robbery. In one condition, it was mentioned that the suspect was white and, in another, that he was black.

Upon analyzing the positions adopted by the participants regarding the police violence, the results showed that, in general, the participants tend to disagree with the use of police violence. However, the frequency with which positions of agreement or disagreement were presented and the arguments to justify them differed depending on the color of the target of the violence.

With regard to agreement with the police violence (classes 1, 3, and 8), participants presented similar arguments for both suspects (white and black). The justifications used for this agreement are stated through an authoritarian repertoire that aims to defend against an imminent enemy, in addition to fighting criminality, which would justify the use of police violence (^{Jackson et al., 2018}; ^{Lemgruber et al., 2017}). In these judgments, the police action, even if in disagreement with the law, is perceived as effective and is based on its punitive use (^{Farias et al., 2017}). This aspect is important because the perception of effectiveness of police action is a legitimizing factor of police violence (Jackson et al., 2020).

It is worth noting, however, that in the scenario in which the suspect-target of police violence was black, the positions supporting the use of police violence totaled 51.43 % of the corpus (class 1 - Violence as justice: 22.86 % + class 3 - Justified violence: 28.57 %). Whereas, in the scenario in which the suspect-target of police violence was white, the statements favoring the use of police violence (class 8 - Violence as a solution) corresponded to 35 % of the corpus. In this way, the idea of police violence as a legitimate measure of justice against crime is evoked more prominently when the suspect is black, which converges with previous investigations that qualify support for police violence as a discriminatory process (^{Álvaro et al., 2015}; ^{Bryant-Davis et al., 2017}; ^{da Costa Silva et al., 2018}; ^{Farias et al., 2017}).

Differences were also observed depending on the suspects' skin color when analyzing the arguments that disagreed with the use of police violence. In the context in which the suspect was black, police violence was seen as inappropriate and unjustified (classes 2 and 4). Although these classes disagree with the police action, the emphasis of these repertoires is on the inappropriate conduct of the police officers without referring to the illegal nature of these actions, with control of the criminal being necessary (class 2), even if he is just a suspect. In the context in which the suspect was white, the justifications used to disagree with the violence emphasize illegality as the basis of the police violence (classes 5, 6, and 7). For the participants of these classes, the police action is at odds with their essential functions, which would be to guarantee the safety, rights, and protection of citizens and therefore, the arguments support that police officers should avoid the use of force against the suspect.

It is important to emphasize that, in the set of analyses, the presumption of innocence only appeared in the situation in which the suspect was white, anchored in the lack of confirmation of the suspect as a criminal (class 5). While the attribution of blame was seen in the context in which the suspect was black, supported by the indication of imprisonment as an appropriate sanction for the crime committed (class 2). These positions are supported by the results found by ^{Dukes and Gaither (2017}) who observed that black victims tend to be more blamed for their punishments, compared to white victims.

The results showed that, despite disagreeing with police violence, the arguments indicated a differentiated treatment between racial groups and are also in line with research that shows how police violence is more supported when perpetrated against ethnic/racial minorities (^{Álvaro et al., 2015}; ^{Bryant-Davis et al., 2017}; ^{da Costa Silva et al., 2018}; ^{Farias et al., 2017}). This differentiation becomes more evident when we analyze the positions and observe that their organization presents two distinct lines of logic: when the judgment refers to the context of the white suspect, 65% of the arguments highlight the illegal conduct of the police (class 5 - Unnecessary violence: 22.5 % + class 6 - Illegal violence: 22.5 % + class 7 - Arbitrary violence: 20 %), while in the context of the black suspect, 66.67% of the arguments highlight the suspect's criminal conduct (class 2 - Legal sanction: 15.24 % + class 1 - Violence as justice: 22.86 % + class 3: Justified violence: 28.57 %).

These results come close to blatant expressions of racial prejudice and discrimination (^{Allport, 1954}), in which the distinction of treatment stems only from the difference in color between victims of police violence. In fact, social categories are often used to explain complex events of a negative nature, especially when referring to minority groups (^{Tajfel, 1981}). In this way, these positions can contribute to the processes of legitimation of interracial inequalities in our society, either through the preservation of the image of the white group, or through the demeaning of the black group.

Finally, although the analysis of moral judgment was not our objective in this study, upon analyzing the principles that organize the positions regarding police violence (figures 1 and 2), we were able to observe that the justifications were anchored in the basis of closeness and distance to the principles of justice, illegality, arbitrariness, and acceptance of police violence, leading us to the perception and assessment of the police actions as morally right or wrong. In this way, the anchoring of these arguments can reflect action schemes concerning moral reasoning in this position-taking.

In recent years, developmental psychologists have demonstrated how the process of acquiring morality can provide a means by which racism can be thwarted. Morality consists of prescriptive norms about how people should treat each other, contemplating concepts such as justice, equity, and equality (^{Killen & Rutland, 2011}). The process of acquiring morality depends both on children's cognitive development and on their socialization process, including the messages communicated by adults, which help children to understand what is fair or unfair in behavior towards different groups (^{Rutland & Killen, 2015}). Morality, therefore, serves as a frame of reference for analyzing phenomena of a cognitive and social nature (^{Rique et al., 2013}), such as racism.

One of the first models for understanding morality was presented by ^{Piaget (1932/1994)} in Moral Judgment of the Child, in which he demonstrates how, throughout cognitive development, the child moves from classifying actions as “right” or “wrong”, to depending on commands of authority, to adopting abstract principles of justice, by late childhood. Piaget's studies served as a starting point for the work of ^{Kohlberg (1984}), who developed a more comprehensive theory about the stages of moral development. According to Kohlberg (1984), for a social judgment to be considered moral, the norms underlying that judgment must be both prescriptive (specific criteria that must be adopted by individuals in a relationship) and universal (it must be applicable to all people). For him, the formation of moral judgment depends, simultaneously, on the subject's capacity for self-control and on the moral atmosphere of the group in which this behavior takes place (^{Díaz-Aguado & Medrano, 1994}).

^{Kohlberg (1984}) presents a model, according to which it is possible to classify three levels of moral development, which vary according to the conceptions of justice adopted in different situations. The first level, called Pre-Conventional Morality, involves the principle of “right” and “wrong” formulated from the immediate consequences of the action. The second level, Conventional Morality, encompasses moral principles based on the preservation of law and social order. Finally, the third level, identified as Post-Conventional Morality, involves conceptions of justice based on respect for civil rights and on principles such as balance of social relations through equity (^{Díaz-Aguado & Medrano, 1994}; Kohlberg, 1984; ^{Rique et al., 2013}).

Consequently, the results presented here show that, on the one hand, the stance towards police violence was anchored in the principles of corrective-punitive justice, in the instrumentality of violence, and in the abuse of authority. The meanings of these arguments are primitive and follow an expiatory justice, which would be in accordance with the pre-conventional morality described by ^{Kohlberg (1984}). In this light, moderating circumstances, such as legal aspects (suspect's rights and police officer's duties), were not considered in the judgment. For these participants, punishment must be immediately applied after the transgression (^{Colby & Kohlberg, 1987}), regardless of the color of the target of the police violence.

On the other hand, police violence was organized around procedural justice and legal authoritarianism. This view reflects a set of principles of justice at the level of conventional morality, and some aspects reflect the principles of post-conventional morality. In the conventional view, the arguments take into consideration the maintenance of the system of rules to resolve conflicts, ensuring impartiality in the protection of society; moreover, in the post-conventional view the arguments imply concern for rights and social welfare in the situation (^{Colby & Kohlberg, 1987}). However, this concern for rights was only revealed in the context in which the suspect was white.

In view of this, it can be said that the position towards police violence seems to vary depending on the color of the target of the violence and presents qualitative changes in its meanings depending on the type of justice reasoning used by the participants. This means that discussing the problem of racial discrimination in a context where the use of physical force is perceived as legitimate implies seeking to know how moral judgment and racism are related to the way police violence is perceived.

Final considerations

Police violence has generated a discourse that reinforces the belief in its necessity, by justifying and legitimizing violent actions in the fight against crime. As a result of this repertoire, defense and guarantee of protection are replaced by the abusive use of force (^{Jackson et al., 2018}; ^{Lemgruber et al., 2017}; ^{Sinhoretto & Morais, 2018}). As a result, the perception of legitimacy of police actions is not always in line with an expression of justice, nor does it presuppose an assessment of the criteria of equality and proportionality. Although justice is considered the synthesis of ethical values, in the context of interracial relations these values are submitted to the interpretation of the symbolic position that different social groups occupy in society.

Based on the results of this study, we were able to observe a tendency regarding the different treatment between Whites and Blacks that is inclined to justify and legitimize social and cultural practices that propagate racial hierarchies. In this way, Blacks continue to be the target of a process of social exclusion, resulting from discriminatory situations in which prejudice develops. Furthermore, we consider that the arguments presented indicated that the way this violence was perceived is associated with different types of moral reasoning.

The results obtained show that it is necessary to analyze what happens in different contexts to understand the different positions of individuals regarding police violence and how their positions can support behaviors that, developed by legitimate institutions, such as the police, would go unnoticed without being seen as discriminatory practices, but as legitimate procedures. It is important to emphasize that the scenario of police violence presented in this study is fictitious and that it is not intended to attribute to police officers, individually, any prejudiced characteristics, but to reflect on real social situations that are conducive to the manifestation of prejudice and discrimination. And here we point out a line of research that can be deepened: how the military police, themselves mostly black, position themselves regarding police violence whose preferred target is the black population.

Another aspect that was not analyzed in the work presented here is the influence of skin color and social class of the participants in taking a position on the type of violence investigated here. It is true that, taken together, the results from ^{da Costa Silva et al. (2018)} show that public school students in a northeastern Brazilian city, mostly black and lower class, are more supportive of police violence when it is committed by a black person than in the situation with a white person. Furthermore, these youths would give greater compensation to a victim of police violence when the victim was white than when the victim was black. Equally important in this work was the fact that young people were also more supportive of lowering the age of criminal responsibility from 18 to 16 years in situations where crimes had been committed by Blacks than by Whites. However, the variable “participant skin color” was not directly investigated. Thus, this is a line of work that also deserves to be investigated in future works.

Finally, although the positions of the participants of this work present characteristics congruent with different forms of moral judgment, the relationship between these constructs was not tested. Future studies should be carried out to overcome these limitations.