Female community health agents’ structure of social thinking about domestic violence against women

This study aimed at describing community health agents’ representational structure on domestic violence against women. A quan-titative-qualitative research study based on the Theory of Social Representations in its structural approach and carried out in a municipality from inland Bahia, Brazil. The community health agents participated through free evocation and centrality techniques: choix-par-bloc, constitution of word pairs and mise-en-cause, from May to August 2019. Data analysis was carried out by means of the EVOC software (Ensemble of Programs Permettant I’analyse des Evocations), similarity analysis and mise-en-cause analysis. These professionals’ representational structure is organized from the central elements of disrespect and sadness, which attribute negative meanings to the representation regarding their positions and repercussions; the other elements integrate specific information to the structure of the representations, justifying them. It is concluded that the understanding regarding organization of the community agents’ social thinking about the phenomenon allows its problematization, as well as the elaboration of prevention and coping strategies for women in situations of violence, the ag-gressors and the community.

introduction Domestic violence against women (DVAW) has aroused both public and scientific interest for representing a problem of significant magnitude that affects families and society in a multifactorial way.It is a phenomenon with historical-cultural roots permeated by beliefs, traditions and values that ground its interpretation and coping, even in the health area 1 .
The social construction marked by inequalities between men and women naturalized by men's power, strength and figure and by domination over women 2 .The asymmetric power relation becomes a driver of violence because it reinforces the dominant masculinity model that minimizes women to the idea of passivity and victimization 3 .
In this scenario, DVAW represents a type of violence directed to women and mostly practiced by intimate partners 3 .The World Health Organization estimates that, at the global level, one out of three women has already been a victim of physical and/or sexual violence at moment of her life 4 .
A research study carried out by the DataSenado Institute on DVAW in Brazil pointed out that, in 2019, the percentage of women who suffered some type of violence corresponded to 27% of the interviewees, and those responsible for the reported aggressions were partners and former partners in 37% of the cases 5 .
Health services are part of the first contact of women in situations of violence and must present an ethical, safe and respectful approach with implementation of actions that seek to minimize the consequences of the problem 6 .Considering that the Family Health Strategy (FHS) has professionals who experience the needs of the community and maintain the bond, it is believed to be a great ally in identifying, preventing and coping with DVAW 7 .
Among the professionals who are part of the family health team, Community Health Agents (CHAs) have the possibility of knowing the family dynamics and creating bonds, being key actors in coping with DVAW situations 8 .A study carried out in the municipality of São Paulo showed that there is certain invisibility of DVAW in Primary Health Care (PHC) and pointed out that the CHAs are the professionals who first identify cases of DVAW in the territory and that, from this, they share the situation with the team 9 .
In this context, health professionals' representations about DVAW can exert an influence on targeting of their actions 10 .Social representations (SRs) constitute a set of concepts, statements and explanations of everyday interactions, and should be considered as a common sense theory, as they are not created in isolation, they are shared and, once created, they acquire a life of their own and transform reality 11 .
It is believed that describing the structure of CHAs' social thinking about DVAW improves knowledge by making it possible to understand the bases of their representations, ideas, images, values and experiences that relate to each other and influence the planning of actions and the way of proposing social practices for coping with the problem.
Consequently, the objective of the current study was to describe the representational structure of community health agents about domestic violence against women.

Methods
This is a descriptive and exploratory study of a quantitative and qualitative nature with theoretical and methodological support from the Theory of Social Representations with a focus on its structural approach or the Central Nucleus Theory (CNT) 12 .The CNT proposes that SRs are a double system made up by the central nucleus that defines, organizes and gives meaning to the elements of the representation, and by the peripheral system that comprises representations of a conditional nature, which are more flexible, practical and constitute an essential content of the representation 12,13 .
The study loci were 18 Family Health Units (FHUs) from an urban area in the municipality of Jequié, Bahia, Brazil.The inclusion criteria corresponded to urban area units, with double or single teams and with full teams according to the National Primary Care Policy 14 during the data collection period.
The research participants were the CHAs from the aforementioned FHUs.Selection of the participants was carried out for convenience through telephone contacts with the supervising nurse, when a meeting with all the CHAs from each of the FHU teams was requested so that the researcher could invite them to participate in the study.
The first stage took place in May 2018 in a room made available by each FHU.The participants were 107 CHAs who agreed to sign the free and informed consent form; the data were col-lected by the researcher in a room made available at the FHU where the CHA works, individually.Thus, the participants answered the form containing sociodemographic aspects (age, gender, marital status and family income, among others) and the questionnaire of the free evocation technique.This is a projective/associative test and seeks to access the internal organization and structure of a representation 15 .The participants were asked to evoke five words or expressions that came to their mind after hearing the "domestic violence against women" inducing term.
As an inclusion criterion, the CHA should be in functional activity and having been working for more than six months at the FHU, establishing as exclusion criterion CHAs who were on vacation, bonus leave or health treatment.The data were collected from May to August 2019, in two stages.
For the analysis of the evocations, the EVOC (Ensemble de Programmes Permettant I' analyse des Evocations) software, 2005 version, was used, which makes it possible to perform a prototypical analysis based on the organization of the evoked terms according to the evocation frequency and order; the intersection of these two criteria produces the four-quadrant chart 13,16 .
The chart corresponds to four quadrants that organize the terms as follows: in the upper left quadrant is the probable central nucleus of the SR; they are the most significant terms for the subjects because they present high frequency and the lowest rank (mean evocation order).The words located in the upper right quadrant constitute the first periphery (they have high frequency and also high rank), while the lower right quadrant represents the second periphery, with words that have low frequency and high rank.On the lower left quadrant, we see the contrasting elements that have low frequency and low rank 13,16 .
The second data collection stage took place from June to August 2019: the same group of CHAs that contributed in the first stage was asked to participate in other data collection techniques, with 60 CHAs collaborating at this second moment.
According to the CNT, there are methods to identify the elements effectively belonging to the central nucleus; they should lead to a more consistent final apprehension of the structure of the representation, and it is as necessary as apprehension of the SR content 13 .In this stage of the research, tests for confirming centrality of the elements and their interconnections were used: choix-par-bloc, constitution of word pairs, and mise-en-cause.These tests were planned based on the previous result, that is, the words were organized from the prototypic analysis, so that the candidates to the central nucleus were tested.
Choix-par-bloc, or successive choice by blocks, allows for a quantitative approach to the representation's items.It is through it that the similarity relationships are evidenced within the representation, as well as those of antagonism or exclusion, allowing to compare the relative importance of certain elements 13 .Therefore, a list was prepared with all 15 elements of the four-quadrant chart for the "domestic violence against women" inducing term and the participants were asked to choose the five words that most characterized the object and the five that least characterized it, as well as to write down the remaining five words.
Constituting word pairs consists in identifying the quantitative property of the central element by means of its connectivity.The method required asking the subject to create, based on a corpus (in this case the evocations on DVAW), a set of word pairs that seem to "go together" 13 .
Mise-en-cause, or questioning technique, is a qualitative method to verify centrality of the central nucleus through the salience of the representation object's constituent elements 13 .In this way, it is assumed that the central elements of a representation are non-negotiable, that is, through a negative question, with the "yes", "no" and "perhaps" answer options; every time that, without the presence of one of the cognems, the subjects consider the object unrecognizable, this cognem is central 17 .Thus, the following question was presented: Can there be DVAW without physical aggression?; and so on with the following words: disrespect, cursing, lack of love, sadness, sexual abuse, cowardice, mistreatment and anger.
The analysis of the data from the centrality tests was performed at three moments.At the first moment, the analysis of the choix-par-bloc or successive choice test by blocks was carried out: for each cognem, a value ranging from +1 to -1 was scored, with +1 for the words of the most characteristic block, -1 for the least characteristic and 0 for the rest.For each group, the mean emphasis of each item was calculated, adding the total values conferred by the sum of the relations between two elements and dividing it by the number of individuals, finding the distance index 13,18 .The operation is performed for each pair of cognems and allows developing the "similarity matrix", with its result arranged in the maximum similarity tree 13 .The data from the constitution of word pairs were analyzed using the similarity analysis technique, with calculation of a similarity index for each word pair (number of co-occurrences divided by the number of subjects); these data were arranged in the similarity matrix and the maximum tree was built to facilitate understanding 13 .
Mise-en-cause is a double-negation technique, which defines the percentage calculation to identify the most frequently chosen answer of each cognem (Yes, No and Perhaps).Confirmation of a cognem as the central nucleus of the representation is based on the negative answers, that is, if the negative answer is concentrated in more than 75% of the answers, it can be stated that this cognem is the central nucleus 19 . This

results
A total of 107 CHAs participated in the free evocations, of which 103 were female, aged from 29 to 67 years old, with predominance (48) pf the age group between 36 and 45 years old, brown skin color (68), married (62) and time of performance as CHAs from 10 to 23 years, with the majority having been working for 20 to 23 years (52).60 CHAs participated in the centrality tests: 57 were female, 37 evangelicals and with family incomes ranging from 1 to 3 minimum wages (46).
The corpus formed by the CHAs' evocations against the "domestic violence against women" inducing term totaled 523 words, of which 99 were different, with a mean evocation order (rank) was 2.60, minimum frequency of 11 and the mean frequency of 21.Data analysis resulted in the four-quadrant chart (Chart 1).
In the upper left quadrant are the most relevant and significant terms, probably constituting the central nucleus of the representation 15 .It is observed that the words physical aggression, disrespect and cursing had a higher frequency and were evoked more readily, justifying their presence in the central nucleus 13 .Physical aggression was the most evoked term and with the lowest mean evocation order, cursing presented a high frequency; however, the disrespect element, despite not presenting a frequency as high as the terms physical aggression and cursing, presented a lower mean evocation order than cursing, which also explains its importance as a likely central nucleus.
In the upper and lower right quadrants are the elements of the first and second periphery.In the first periphery we find the terms lack of love and sadness placed as the most important peripheral elements due to their high frequencies, which may prove to be central 15 .In the lower right quadrant are the most unstable elements to changes, associated with the life context and to the social practices, 15 namely: jealousy, financial dependence, drugs, sexism, fear and death.
In the lower left quadrant or contrast zone we have the following elements: sexual abuse, cowardice, maltreatment and anger, which have low frequencies and were promptly evoked; they are terms evoked by few participants, but which can reinforce the ideas of the first periphery, complement and discuss the central nucleus or even reveal the existence of a subgroup which has a different representation 10 .
In this way, the results of the centrality tests will be presented, aiming to deepen the hypotheses of centrality of the CHAs' representations about DVAW.
In the choix-par-bloc, or successive choice by blocks method, the connections established between the elements presented the conformation that can be seen in Figure 1.
The term disrespect appears only linked to the word physical aggression, which is the strongest link between the cognems.Physical aggression has four links; in addition to disrespect, it was linked to other elements from the second periphery such as male chauvinism, which in turn was linked to cowardice; another element also appears: jealousy; and, on the other side, mistreatment.The cognem mistreatment was linked to three elements; in addition to physical aggression, it has a connection to the term sexual abuse in the contrast zone and presents a strong connection with the element of the probable central nucleus, cursing.Cursing was linked to sadness, an element from the first periphery and was the cognem that made the most connections; in addition to cursing, it is connected to lack of love, financial dependence, anger and fear, and fear was linked to other elements from the second periphery: death and drugs.
In this way, in addition to the term physical aggression having already been pointed out in the analysis of the four-quadrant chart as a probable central element, it can also be considered in the choix-par-bloc analysis and include other elements as probable central terms, such as mistreatment and sadness.
Another centrality confirmation test used was the constitution of word pairs and, as a result of applying the instrument, another maximum tree was produced considering calculation of the similarity index, as shown below in Figure 2.
In this maximum tree, it is observed that the sadness element had the highest number of con-nections, with four links to the terms; followed by male chauvinism, disrespect and lack of love, each with three connections to the other terms.In this result, the term disrespect continues to appear as a central element according to the four-quadrant chart, presenting the strongest connection with the term cursing, which, as is the case with the term physical aggression, appears as a probable central element in the prototypical analysis, although it is not confirmed in this result.
Thus, the likely central elements for this test, that is, those that have more connections are as chart 1. Four-quadrant chart created by the female community health agents' evocations given the "domestic violence against women" inducing term.Jequié.Bahia.Brazil follows: disrespect, male chauvinism, lack of love and sadness.It was found that the terms that comprise the peripheral system appear as possible central elements of the CHAs' SRs about DVAW.Thus, these findings make it possible to understand the need to use various methods in studies on SRs in order to certify centrality of the elements and define the representational structure.
The mise-en-cause or questioning technique contributed to the study through the systematic identification and qualitative property of the central elements 20 .In this way, the elements that stood out the most in the analysis of the four-quadrant chart were tested, such as the terms from the possible central nucleus, first periphery and contrast zone.The numbers of "no", "yes" or "perhaps" answers referring to each element followed by the percentage are shown in Chart 2.
In Chart 2, the most characteristic elements of the DVAW SRs are observed; however, in the CHAs' social thinking, the non-negotiable elements are the following: disrespect, sadness and cowardice, that is, these terms presented more than 75% of negative answers to the questioning technique, showing a positive result as belonging to the central nucleus.The disrespect element present in the central nucleus of the four-quadrant chart was the one that obtained the highest percentage in the confirmation with 88.34%; however, we have physical aggression (5%) and cursing (15%) also present in the quadrant of the probable central nucleus not managing to reach 75%, having centrality refuted in this test.
The sadness element present in the first periphery of the prototypical analysis had its centrality confirmed with 78.34%; in turn, lack of love, also present in this quadrant, presented only 31.67%, a very low percentage to confirm its centrality.The elements of the contrast zone were important in this test, the cognem cowardice presented the second highest percentage (85%), confirming it as non-negotiable in the DVAW representation ; the other terms, sexual abuse (3.34%), mistreatment (71, 67%) and anger (56.67%) were not central.Thus, for the mise-en-cause method, the central elements are as follows: disrespect, sadness and cowardice.
Based on the findings presented, it is possible to identify the central nucleus and describe the structure of the CHA's thinking about DVAW.Thus, as an analysis criterion for defining an element as central, a minimum of three methods were established indicating centrality of the element 21 .
It is verified that the terms physical aggression, disrespect and cursing presented themselves as a probable central nucleus in the prototypical analysis; however, only the disrespect element confirmed centrality for also presenting proof in another two other: constitution of word pairs and mise-en-cause.Lack of love and sadness are elements belonging to the first periphery, but

Discussion
According to the structural approach, SRs tend to verify the influence of social factors on thought processes through the identification and organization of relationship structures 20 .It was observed from the set of analyses that the CHAs' SRs about DVAW are structured through two central elements: disrespect and sadness, which attributes negative meanings to the representation regarding the group's stance on the problem and its repercussions.
The central nucleus of a representation must play an evaluative and pragmatic role; for this to occur, normative and functional elements must be formed 22 .Thus, it is observed that these professionals' representations are organized around the attitudinal dimension through the term disrespect and that the affective dimension is defined by sadness.
Disrespect constitutes a normative element; it refers to the subjects' value system and attributes a social aspect to the central nucleus.It is noticed that the disrespect for the CHAs can be understood as an attitude or behavior of the aggressor in relation to the woman, being the trigger of the domestic violence episodes.The central nucleus elements generate a global meaning of the representation, have strong historical and ideological roots and are consensual in the group 21 .
Violence is a disrespect instance that occurs within the woman's own home, an environment that would serve as protection and becomes unsafe, as the aggressor is inside the house 23 .Thus, disrespect is anchored in values perpetuated by society and, in this context, the inequality relations between men and women generate attitudes of intolerance and disrespect and act as a foundation for violence 1 .
With regard to the prototypical analysis, it is observed that, together with the elements of physical aggression and cursing, disrespect makes up the probable central nucleus.In addition to indicating the conceptual and imagery dimension of the representation, these terms demonstrate the physical and psychological form of DVAW.Physical violence is the most recognized by health professionals when women seek services, as it leaves physical marks, 1 while cursing is understood as a type of psychological violence that impacts health with the same or greater intensity due to emotional damage and reduced self-esteem; and may appear invisible to battered women as they consider it a normal behavior within the family models 3 .
In this context, there are significant repercussions for the mental and physical health of all family members, not only for women who experience violence, mainly due to the perpetu- ated culture of violence 24 .A study carried out in Bangladesh interviewed 87 women in situations of intimate partner violence, where they presented physical violence as the main form of aggression and repercussions in the lives of other family members 25 .
A research study carried out in the city of Petrolina/PE found that many situations of physical aggression were linked to the occurrence of sexual violence/abuse 26 .In the meantime, sexual abuse is found in the contrast zone in the four-quadrant chart, which reinforces the idea of forms of violence in the CHAs' social imaginary and confers meaning to the mistreatment element.In the choix-par-bloc analysis, it is verified that the mistreatment element is important, with possible centrality due to the connections established with the terms physical aggression, sexual abuse and cursing; however, it was not confirmed as central in other centrality tests.
In this context, for these health professionals, DVAW is disrespect for women objectified in acts that correspond to physical aggression, cursing, sexual abuse and mistreatment, which put women's physical and psychological health at risk.Based on this understanding, the CHAs highlighted the image they create of the aggressors, associating them with the cowardice and male chauvinism elements, which reinforce the value system together with the term disrespect.
The term cowardice was part of the central nucleus of the four-quadrant chart from other studies conducted with health professionals on DVAW 10,22,27,28 .In the current study, the cowardice element was part of the contrast zone of the four-quadrant chart and proved to be a non-negotiable term in the CHAs' SRs about DVAW in the mise-en-cause test, although not confirming its centrality in other methods.On the other hand, the male chauvinism element was important in the organization of the similarity analysis of word pairs, with connections to the terms disrespect, cowardice and jealousy.
The attitudinal dimension brings about gender issues as a reflection, being fundamental in interpreting the phenomenon, when realizing that the historical-cultural conditions continue to contribute to the legitimation of women's submission and to superiority of men, who, exercise power over women and children through hierarchical relations 29 .
As a result, stereotyped roles, understood as something natural, resulted in gender inequality 2 and violence started to be used as a manifestation of this masculine domination, that is, submission of the weakest to the strongest, translating into mistreatment 3 .In this way, the patriarchal mentality, which advocates control of women by men, will always be present in aggressions, mainly due to jealousy, reflecting men's fear of losing their sexual and social object 30 .
In this context, around the CHAs' representational structure about DVAW, sadness is revealed as a central and functional element, in addition to feeling cultivated by the female professionals in the face of acts of violence.This element assumes the affective dimension of this representation and plays a functional role, as it is through it that the representation can be anchored in the reality and social practices of the CHAs.Thus, in this study it is indicated that the practice should not only be understood as physical behavior, but as the discourse around the social object 31 .
A research study carried out with health professionals about DVAW pointed out that sadness is a feeling they show when witnessing a situation of violence, in addition to mentioning that this feeling is also experienced by women who experience violence 28 .Another study carried out with CHAs showed that the feeling of sadness permeates their care practices, as they are professionals who recognize the weaknesses of their actions in the face of domestic violence against women 32 .
From the choix-par-bloc similarity analysis, it is observed that sadness has a higher number of connections contributing other affective burdens to the representation, such as fear, anger and lack of love.
Fear brings about mixed feelings in the DVAW context both for health professionals and for women.Regarding the female professionals, this affective term is linked to professional unpreparedness to act in the face of the problem and to the fear of suffering reprisals from the aggressors, thus limiting their care practice; and for women, to the fear of feeling guilty for the situation, submission to the aggressor and fear of death 10,31 .
Men's power struggle over women entails control and fear 31 .A study carried out in Saudi Arabia pointed out that women are reluctant to reveal the domestic violence suffered due to several factors, with fear of losing their family and shame among them 33 .
This entire context allows the CHAs to expose the feeling of anger, which can be reflected by all the influence of social roles between being a man and being a woman naturalized in society and the absence of decision-making by women in search for changes.A research study carried out with CHAs pointed out that revolt is seen as a feeling in the professionals resulting from the woman's repeated reconciliations with the aggressor, which discourages them from helping and supporting women in situations of domestic violence 7 .
The experiences of CHAs with women in situations of violence envision that health care should enable developing identification and intervention strategies with provision of support 34 .
Lack of love can be represented by the CHAs through their practical experiences, where they tend to observe the interpersonal conflicts experienced by the women, reflecting on lack of self-esteem and on the man-woman relationship.Violence situations reveal a counterpoint: men who, from idealized husbands, become aggressors; and they also present in this confrontation the limit between love and violence 3 .
In the meantime, striving for love and family, women begin to justify their partners' aggressive behavior by using alcohol and/or other drugs or via some personality disorder 3 .In turn, abusive drug use is a common reality among the aggressors, and health professionals tend to evidence it as one of the triggering factors of DVAW 28 .
In this way, the central nucleus comprised by the elements disrespect and sadness defines and distinguishes the CHAs' SRs about DVAW, whereas the other elements integrate the information that is specific to the structure of the representation, justifying them and linking them to the practices.

conclusion
From the prototypical analysis, it was possible to understand the CHAs' social thinking based on shared and cognitively activated elements, presenting the terms "physical aggression", "disrespect" and "cursing" as a possible central nucleus.
Throughout the structural analysis, based on the four-quadrant chart, tests were carried out to define the centrality and role of the elements that make up the representation.In choix-parbloc, it was observed that the elements with the highest number of connections were "physical aggression", "sadness" and "mistreatment".In the maximum similarity tree of word pairs, the most salient elements were disrespect, lack of love, sadness and male chauvinism.In mise-en-cause, as verified in the results, the non-negotiable elements for the CHAs in the DVAW context were disrespect, sadness and cowardice.Thus, centrality in the SR on DVAW occurred for the disrespect and sadness elements, by confirmation in three methods applied.
It is understood that, through these elements, CHAs contribute the attitudinal and affective dimension of the representation of DVAW, assuming a normative and functional role.Disrespect organizes social thinking based on the CHAs' values and judgments regarding the aggressors' acts, and sadness contributes the affective and practical burden of these professionals through their experiences in the DVAW context.
It is believed that the understanding of the organization of CHAs' social thinking about DVAW allows the family health team to problematize the phenomenon, as well as the elaboration of prevention and coping strategies with women in situations of violence, aggressors and the community.
study complies with Resolutions No. 466/2012 and No. 510/2016 of the National Health Council that regulate research involving human beings, and it was submitted to and approved by the Research Ethics Committee of the State University of Southwest Bahia under opinion No. 3,233,780/2019 and CAAE: 07558718.1.0000.0055.
Figure 1.Maximum similarity tree based on choix-par-bloc for the elements of the "domestic violence against women" term.Jequié, BA, Brazil, 2019 (n = 60).

table 1 .
Distribution of the answers to the questionnaire technique (mise-en-cause) for the "domestic violence against women" term.Brazil, 2019 (n = 60).