Print version ISSN 0102-311X
Cad. Saúde Pública vol.18 n.6 Rio de Janeiro Nov./Dec. 2002
|Ana Rosária Sant'Anna 1, 2 |
Marta Julia Marques Lopes 1
Homicides among teenagers in the city of Porto Alegre, Rio Grande do Sul State, Brazil: vulnerability, susceptibility, and gender cultures
Homicídios entre adolescentes na cidade de Porto Alegre, Rio Grande do Sul, Brasil: vulnerabilidade e culturas de gênero
1 Programa de Pós-Graduação em Enfermagem, Escola de Enfermagem, Universidade Federal do Rio Grande do Sul. Rua São Manoel 963, Porto Alegre, RS 90620-110, Brasil.
Abstract The authors present a quantitative and qualitative study on homicides among teenagers in Porto Alegre, Rio Grande do Sul State, Brazil, based on a historical series during the 1990s and the life and death histories in this group, with a special focus on 1997. In that year there were 68 homicides in which the victims were from 10 to 19 years old. Of the 68, 62 were males and only 6 females, or a ratio of 10:1, showing that young males are more vulnerable and susceptible to being murdered. The data indicate that cause of death is influenced by gender culture and that homicides are based on power and status symbols characterizing a kind of virility. This expression of virility in the shaping of violence also appears in the domination of the female body observed in homicides with young women as the victims. The life and death histories of these teenagers highlight the pertinence of the gender-based analysis as a theoretical-analytical category, in addition to analyses considering socioeconomic aspects and social inequity.
Resumo Este é um estudo quanti/qualitativo sobre homicídios entre adolescentes em Porto Alegre, tendo como base uma seriação histórica para a década de 90 e as histórias de vida e morte nesse grupo, particularizando o ano de 1997. Nesse ano houve 68 homicídios na faixa etária de 10 a 19 anos, dos quais 62 do sexo masculino e 6 do sexo feminino, numa razão de 10:1, mostrando que os homens estão mais vulneráveis e suscetíveis a morrer por essa causa de morte. Admite-se, a partir disso, que as formas de morrer sofrem influência das culturas de gênero e os homicídios se sustentam em símbolos e posições de poder que caracterizam um tipo de masculinidade. Essa expressão de masculinidade na forma de violência mostra sua face também no domínio sobre o corpo feminino, observado nas situações de homicídio envolvendo as mulheres adolescentes. Esses achados apontam para a conotação relacional e para a assimetria social, expressão das culturas de gênero, como atributos explicativos. As histórias de vida e morte desses adolescentes mostram a pertinência das análises baseadas no gênero como categoria teórico-analítica, além das análises que contemplam os aspectos sócio-econômicos e as iniqüidades sociais.
This article is the result of a Master's thesis presented in the Graduate Studies Program in Nursing at the Federal University in Rio Grande do Sul (UFRGS), February 2000 (Sant'Anna, 2000). The article adopts an analytical methodology allowing a gender approach to the case series of homicides among adolescents.
Homicides among adolescents are linked to the overall issue of violence and require an analytical effort to jointly interpret the individual data.
Most current explanatory models of violence in Brazil are based on statistical studies which rarely grasp the issue's complexity. The need to rethink this issue and reconstitute a series of factors involved in homicides led us to focus on primary data to correlate education, socialization (gender cultures), delinquency, interpersonal relations, and social conditions. In our opinion, other factors (biological and intra-psychological ones, among others) merit attention, but in this study we adopted a gender culture perspective and the analytical possibilities of the gender concept as defined by the Anglo-Saxon literature.
Mortality from external causes is defined as the sum of homicides, suicides, and fatal accidents. Deaths from external causes are not distributed equally between males and females. According to the statistics consulted in this study, violent deaths occur predominantly among males (some 80% of the total in Porto Alegre from 1990 to 1997).
Mortality distribution by gender shows males with external causes as the third most common causal category every year except 1998, when respiratory tract diseases occupied third place. Among females, external causes were the fourth most common category only in 1990 and were fifth in 1980, 1985, and 1995. These deaths also varied as to the basic cause, with traffic accidents as the main cause among females, unlike males, in whom homicides were the main cause (Secretaria Municipal de Saúde de Porto Alegre, 1999).
The first phase of this study was conducted using secondary statistical data from the State Secretariat for Health and the Environment (SSMA) and the Porto Alegre Municipal Health Secretariat (SMS) to construct a historical series of homicides among adolescents in the 1990s, allowing for an assessment of the magnitude of this cause of death in the Porto Alegre population (Secretaria Municipal de Saúde de Porto Alegre, 1999). The second stage of the study focused on the year 1997; we conducted an in-depth study of the 68 teenage homicides that occurred that year, with a reconstitution of the life history of these adolescents. Note that the complete statistics are not presented in this article due to space limitations, but they are available to interested readers.
According to Anglo-Saxon authors, and Joan Scott (1990) in particular, gender is a constitutive element in social relations based on perceived differences between the sexes, and is the first form ascribing meaning to power relations. The author emphasizes that changes in the organization of social relations always correspond to changes in representations of power, and that such changes do not necessarily take a single direction, but involve four different situations: culturally constructed symbols, represented by modalities and their contexts; normative concepts, expressed in religious, educational, scientific, political, and legal doctrines that determine the meanings of male and female as the product of a social consensus rather than a conflict; construction of gender not exclusively through the kinship situation but also through economic and political organization; and finally, subjective gender identity constructed from historically situated social organizations and representations. Scott further highlights that these four elements operate interdependently, but not simultaneously, as reflections of each other, thus requiring historical research into the relationships between them.
Another author who has studied gender and social inequities with an emphasis on epidemiology is Breilh (1996), who identifies asymmetrical gender relations as factors producing inequity, related to cultural and historical issues in the sexual and biological differences between men and women. The health/disease process in the male/female universe is constructed in the midst of social contradictions and specific conditions in social reproduction. There are patterns between males and females that differentially expose individuals to the protective and destructive processes influencing their health.
The homicide-related situations described in this study are in a sense "compatible" with what is socially legitimated as the standard for virility, since attributes related to physical strength, courage, and tenacity, among others, constitute a positive form of virile social identity (Lopes, 1996). Such sex-related "attributes" constitute what we refer to as "genders" and partially explain the prevalence of homicides among males and the violence that is intimately related to relations of power and domination, making men victims of their own actions.
Breilh (1996) adds that violence is directly related to inequity, which can be influenced by a combination of three factors: economic oppression, ethnic discrimination (including racial discrimination), and gender violence.
Within this context, we adopt other concepts capable of helping us "deal with" the issues raised. We thus use the concept of "vulnerability" as defined by Lert (1996) as the variants constituting the individual and collective dimensions defining a risk situation. Thus, in this study, vulnerability is constituted by the variants contributing to the context of violent deaths among adolescents. As for individual vulnerability, we high light the need to always contextualize it in reference to a situation, or to a given moment within a given society.
Based on this conceptual approach, what we refer to as a "risk situation" is the set of factors or vulnerabilities which - from the point of view of public health - constitute the social facts and relations (especially gender relations) that predispose to violent deaths. Thus, this article adopts the definition of vulnerability according to the meaning of social risk of delinquency and death.
Due to the complex, multidimensional nature of the proposed analysis, we adopted a pluralistic methodology capable of harboring perspectives that would allow for a multifaceted understanding of the theme. This choice led us to propose an approach to what could be called a combinative or hybrid terrain. From this same perspective, the methodological/analytical proposal used statistics for the analysis of quantifiable data and content analysis to categorize and/ or quantify the qualitative or discursive data.
Several authors discuss the use of a hybrid methodology. Cortes (1998) refers to the use of qualitative analysis as an adjunct or illustrative method in research that mainly uses quantitative analytical techniques or the combination of the two methodologies, without a clear predominance of one over the other.
With regard to the joint utilization of quantitative and qualitative methods, Minayo (1992) argues that the quality of the facts and social relations are their inherent properties and that quantity and quality are inseparable and interdependent, hence requiring the dissolution of quantitative/qualitative dichotomies.
In this study, the statistical analyses are frequently punctual, cannot be generalized, and comprise an interpretative analytical framework. They frequently highlight the broader scenario of the specific facts under investigation.
The subjects of this study were adolescent homicide victims from the year 1997 and their families. Adolescence was defined according to the Brazilian Ministry of Health as the age bracket from 10 to 19 years (CNPD, 1997).
The year 1997 was chosen during the project's preparatory phase, in 1998, when the database of the National Mortality Information System (SIM) underwent a quality control process which allowed to work with these data.
The epidemiological demarcation for the study of external causes is the historical series from the 1990s and the case series for the year 1997, since the study focused only on families who had lost an adolescent child to homicide during this historical cross-section. Fletcher et al. (1989) define a case series as a prevalence study in a group of individuals with a given disease or injury, held at a single moment in time. This demarcation allowed for an exploratory study of the life and death histories of these adolescents. The seriation was longitudinal for the 1990s and allowed for some comparisons.
Relating to the 68 adolescent homicide victims in 1997, 64 families were located. Eight families refused to participate, and 57 interviews were actually conducted. Three addresses were not located, so families not found and those which refused totaled 11 interviews not conducted, or 16.2% of the total sample. Semi-directed structured interviews were held with the families with the aid of a data collection instrument. The instrument consisted of closed and open questions. The closed questions provided data for the socio-demographic and epidemiological analytical quantification, based on the selected variables. The open questions were structured to supplement the socio-demographic and epidemiological information, and based on the thematic categorizations. They allowed us to unveil, learn about, and understand the dynamics of the individual, family, and social factors that had exposed the adolescents to homicide. In short, they composed the vulnerability to homicide risk situations in the study group.
Considering the notions of vulnerability and risk situation, the study variables were grouped according to three dimensions: socio-economic processes of the adolescents and their families; the adolescents' family processes; and the adolescents' individual processes.
To study the quantitative variables, a descriptive analysis of the data was performed, based on simple frequency tables, central tendency and dispersion measures, and a proportional comparison test. The study also used tables cross-analyzing the socioeconomic variables, gender, and race with variables on previous history and information on the death.
Deaths were also located spatially on the city map using the Georeferenced Information System (GIS), where the spatial unit was the neighborhood, allowing for a comparison of the city's various neighborhoods based on social indicators from the census conducted by the Brazilian Institute for Geography and Statistics, or National Census Bureau (IBGE, 1999). This comparison indicates a concentration of these events in neighborhoods characterized by adverse social indicators.
Since homicides have legal implications, in order to comply with ethical research principles we kept the identity of the research subjects secret and guaranteed the information's confidentiality. Families were informed verbally about the research objectives and received an Informed Consent form, requesting the interviewee's authorization and signature attesting to his/her knowledge of the document's content.
Gender, external causes, and homicides - some data
Table 1 shows that males were the main victims of homicides during the seven year period, although there was a slight increase in homicides among females in all age brackets, especially from 10 to 19 years.
Homicide rates among males showed a certain regularity during the period studied, with a slight decrease from 1994 onward. From 1990 to 1997 the rates varied by 3.7%. Male homicides in the 10 to 19 year bracket varied by 7.6%. For females, the rate varied 3.80% overall and 7.70% in the 10 to 19 year bracket.
In Porto Alegre, all deaths from external causes also occur predominantly in males, with some 80% of the total during the study period. Comparing Tables 1 and 2, the trend in the proportion of deaths from external causes in Porto Alegre is similar to that in the State of Rio Grande do Sul as a whole. However, in Porto Alegre male and female deaths have increased in absolute numbers since 1994, except for the years 1996 and 1997, when deaths from external causes in women decreased by approximately 10%. As shown in Table 2, 88% of homicides occurred in males in Porto Alegre, while inversely, female victims were less than 12%. Specifically in 1993, less than 5% of all homicide victims were has females.
In the 10 to 19 year age bracket there was no variation in the proportion of homicides in males and females until 1992. In 1993 there were no female homicide victims, so the proportion of males was greater than in 1994, but in absolute numbers there was an increase of 33% among male victims that year. The absence of female homicide victims in 1993 may be related to the inclusion of "events of indeterminate intent" among females, which increased by 100% in 1993.
Despite the higher proportion of homicides among males as compared to females, from 1995 to 1996 there was a drop of 25% in deaths among males and an increase of 80% in females. In 1997, male homicide victims increased by approximately 83%, while in females there was a decrease of 3%.
These statistical disparities attest to difficulties in the registry services. This problem hampered the analyses, requiring an effort at constructing a reliable explanatory approach. However, focusing on 1997, the year selected for the in-depth study, produces the following portrait.
Table 3 provides data on adolescents from the 57 families added to the information for the 11 losses (families not located plus refusals), whose data were taken from the death certificates. Among these 68 deaths, males predominated, with 91.2%, or a 10:1 male/female ratio.
To kill or to be killed: the fate of the male gender?
Violence is "a man's thing". This claim is almost a form of praise in certain contexts! The vast majority (91.2%) of the homicide victims were males, showing that men are more vulnerable and susceptible to dying such violent deaths. This does not occur only with homicides, but also with other deaths from external causes, as described previously.
All the homicides with known perpetrators were committed by males, indicating that this type of violence is associated with a type of masculine conduct. In the study population, this is valid for both the victim and the aggressor, thus characterizing a homicide-prone situation predominantly involving men.
Homicides in the study group were related to socially accepted behaviors, linked to power struggles involving macho cultural traits. The majority of the victims were men murdered by other men, and the killers of women were also men.
The lower proportion of female homicide victims (8.8%) also indicates that the motives and situations differ between the genders in this population group. The facts surrounding these murders are different for men and women and are associated with cultural representations of social gender relations. This evidence can be observed in the Tables.
Table 4 shows the distribution of alleged motives for homicides in males. The most frequent explanations provided by families of victims were "senseless" killing, followed by revenge, involvement with drugs, participation in robberies, and "wipeouts" (i.e., victims who were murdered because they "knew too much").
In the words of victims' families, "delinquents" appeared as the main perpetrators of homicides committed against other males, followed by a variety of other aggressors, like strangers, others, friends, relatives, and the police, in that order.
Deaths classified as "senseless" killings involved arguments and fights at parties or in the neighborhoods, triggered by power struggles, differences of opinion between groups, or murders with no explanation whatsoever.
In one situation, a teenage boy had gone to celebrate his birthday at a pizza joint. He was attacked while he was on his way home with a girl cousin. The situation illustrates what the mother defined as a senseless killing. According to her, "My son and niece were waiting at the bus stop and the Passo das Pedras bus was running late, so they took the Leopoldina. And these kids started hitting on my niece in the bus, so my son decided they should get off near an ATM, right? But he could never have imagined that those three guys would come after them. And getting off the bus... those kids came running after them. And I don't know how he reacted, I guess he just stood still and didn't run, just shouted. He just shouted for my niece to run. And he just stood there. So one of those kids pulled a knife out of his backpack and stabbed him. And a policeman who testified caught the kid, said that he saw him when he still just had his hands up and... [weeping]" (mother).
The interview describes a kind of unjustifiable conduct on the aggressor's part in a situation where the victim had played no active part in any crime. The circumstance surrounding his death illustrate an aggressive attitude in a power struggle among males. Competition for "winning behaviors" consecrates such a male identity. Proving one's manliness through socially consecrated conduct based on intimidation and violent acts is one such attitude.
In murder for revenge, the act is supposedly "justified" by the victim's presumed participation in the process. Many of these adolescent victims had displayed violent attitudes, including involvement in fights, physical aggression, and crimes against property and lives, provoking in their aggressors a thirst for justice, configured in an act of vengeance.
The stories of these homicides reveal these situations and illustrate our analyses: "So on the day he died, a Tuesday, he'd been at my house on Monday. He'd been fighting with a guy, at a party, on New Year's Day, and he killed the guy. He killed him. So 21 days after he'd killed this boy who lived nearby... his father killed him, this boy's father killed my kid" (mother).
Another alleged motive for homicides in adolescents was involvement with drugs or drug dealers. Only males were exposed to this risk situation, indicating a kind of conduct involving the conquest of economic power, where "illicit" conduct is associated with challenging the established order, besides highlighting the lack of job opportunities for poor Brazilian youths.
One victim's sister-in-law reported that the drug traffic was this teenager's only alternative for survival, fueling a climate of dispute, fear, and death: "(...) reality? What can I say? He sold drugs... he was involved with drugs. But he didn't kill anybody, wasn't hurting anybody. He was just hurting himself. I don't know what happened... these gang things where they say so-and-so killed somebody, or what's-his-name killed somebody else. But he hadn't done anything. And he didn't do anything. That was all he was doing [dealing drugs], and even cried about it. He used to say to his older brother, my husband, 'I do this to survive. So my old man won't go hungry. I don't like this life of mine. But I got no job, got no education'. And it so happened that he was at a play selling... and then he called out to his brother. He was drunk, high. He called, shouted... I don't know if there was somebody chasing him, he was suspicious... he called out to his brother, but his brother didn't go to the door... just said his brother must be fooling around on the street. So when his brother... his brother just heard the shots. When he heard the shots the whole thing had already happened. When he got to the door the shooting had already happened. There wasn't time. I don't know if it was a hit job... because these things... who sells more... who sells less, you know? In the 'hood... they think they have to... they have to sell more, you know? I think that must of been it, I don't know" (sister-in-law).
Fights over women are also behind homicides. Keeping power over the turf is based on controlling the women, through the power over the female body, and the women are not allowed to get emotionally involved with someone who does not belong to the group. "He killed my brother because he was seeing a girl from the place where he lived. Cause a lot of people from there where this kid lives don't like people from our street. Cause since my brother lived here, even though he wasn't in any kind of trouble, he lived here on this street, right?" (sister).
Other situations, like gang fights, robberies, fights between friends, and "wipeouts" indicate that vulnerability to homicide is characterized by a kind of "masculine" conduct that involves constructing and circulating in areas susceptible to these risks. Such kinds of conduct are "expected" of males, imposing challenges and risks on their physical integrity.
From this perspective, according to Willis (1991:52), violence is one way of expressing manhood, "regulating a kind of honor". According to the author, it is a way of manifesting virility and maintaining group hierarchy.
This expression of manhood through violence also occurs in the command over the female body. Motives for teenage homicides differed between girls and boys. As shown in Table 4, the main reason for homicides against girls was "jealousy" on the part of the male partner or former partner, the perpetrator.
Such homicides against women also suggest gender cultures as explanatory attributes. The boyfriends of the three teenage girls had characteristics in common: they all had criminal records and had been involved in robberies, homicides, and drug dealing. Their relationship to the girls was one of possession, and the murders all occurred when the females reacted unexpectedly to impositions on their own conduct. As one mother reported, "He was very jealous, I mean sick, because he killed her out of jealousy. She had gone out to dance. He went upstairs with her, she wasn't with him any more, she had separated from him because of this child she had to give away and because he was very jealous of her. He wanted her but he didn't want the children. He stalked her, came to our house, and killed her" (mother).
Saffioti (1994) studied these situations and concluded that public space is eminently masculine and that men are more exposed to violence from other men in these places. Women suffer violence in the private domain, and their aggressors are generally former husbands, partners, or boyfriends.
Based on data from the Brazilian National Census Bureau (IBGE) in 1988, 70% of the homicides perpetrated by men against women are related to the man's refusal to accept a separation (IBGE, 1988). The circumstances involving the three female teenagers in this study who were killed by their [former] partners corroborate these findings.
Circumstances surrounding the murders of the other two teenage girls were as follows: one was with a group that was playing Russian roulette, and since she voiced her opinion against this game, she was killed outright by one of the players; the other was raped and then murdered. The latter case illustrates physical force as an instrument of male power over the female body.
The mother's account of her teenage daughter's death confirms that her daughter had been defenseless: "So he grabbed her and started dragging her away. They said the other girls saw it happening, like he threatened her. They said it was a group [weeping]. They said he used a hunting rifle. And they said that... he did what he had to do, right. That he shot her in the head. She started to shout, to cry out for help. The police said the noise was muffled by a mattress and a pillow so nobody would hear the shots. Then he tried to bury her, so no one would see"(mother).
Risk situations: vulnerability arguments
Teenage deaths need to be related to the risk situations to which these adolescents are exposed. Table 5 illustrates how these situations occur both with males and females and express gender susceptibility through involvement in illicit contexts.
The majority of the teenage girls who were murdered consumed illicit drugs (60%). Among the teenage boys, the proportion of drug users (34.6%) was high, but not the majority, according to the families. This information should be interpreted with caution, since among males there were ten cases in which their drug consumption was ignored by the family members interviewed.
In this population group, only males used weapons, denoting an explicit attack-or-defend attitude. Males, unlike females, adhered more frequently to the private use of all types of weapons, or gun use as a "sport".
No statistically significant difference was observed in the variables among males and females based on whether the deceased had spent time in FEBEM (the State Foundation for the Welfare of Minors, a government reformatory system) or had police records. However, in relation to police records, males had participated in more serious crimes, like robberies, drug traffic, assault, and homicides, while females tended to have been involved in petty theft, for example.
These risk situations partially explain the predominance of homicides among males, since these young men displayed attitudes that trigger violence, thus becoming more exposed to this outcome. Meanwhile, the vulnerability of the five teenage girls who were murdered is related to passive attitudes towards their aggressors, which is culturally expected of women and occurs more explicitly in the more vulnerable social strata. This relationship is confirmed by the difference in plans for the future: 60% of the teenage girls (3/5) had no explicit ideals, unlike the teenage boys, whose main life projects were to study and get ahead professionally. The two teenage girls who had plans for the future were focused on more immediate situations (that is, the near future), like work and consumer goals, life changes, and dedication to family.
The involvement of teenage girls in violent situations has reinforced the idea that their emotional ties to the male perpetrators still accounts for their victimization. This attests to the fact that the cultural basis of male-over-female domination continues to express itself in symbols and life styles. This basis, far from protecting "manhood", makes males susceptible and vulnerable to becoming aggressors and aggressed. Without intending to fuel the debate on the supremacy of one gender over the other, we do wish to contribute to reflection and to highlight comprehensive elements for constructing a reality that will no longer reproduce such inequities.
BREILH, J., 1996. El Genero Entrefuegos: Inequidad y Esperanza. Quito: Centro de Estudios y Asesoria en Salud. [ Links ]
CNPD (Comissão Nacional de População e Desenvolvimento), 1997. Jovens no Brasil: Diagnóstico Nacional. Brasília: CNPD. [ Links ]
CORTES, S. V., 1998. Técnicas de coleta e análise qualitativa dos dados. In:Pesquisa Social Empírica: Métodos e Técnicas (C. B. Neves & M. B. Corrêa, org.), pp. 11-147, Porto Alegre: Instituto de Filosofia e Ciências Humanas, Universidade Federal do Rio Grande do Sul. [ Links ]
FLETCHER, R.; FLETCHER, S. & WAGNER, E., 1989. Epidemiologia Clínica. Porto Alegre: Artes Médicas. [ Links ]
IBGE (Fundação Instituto Brasileiro de Geografia e Estatística), 1988. Manual de Atualização Cartográfica. Rio de Janeiro: IBGE. [ Links ]
IBGE (Fundação Instituto Brasileiro de Geografia e Estatística), 1999. Síntese de Indicadores Sociais 1998. Rio de Janeiro: IBGE. [ Links ]
LERT, F., 1996. L'Approche de la Vulnerabilité Face à la Dépendence dans le Champs de la Santé Publique. Seminaires de Recherche du Crips. Paris: Centre Régional d'Informatiment de Prevention du SIDA. [ Links ]
LOPES, M. J. M., 1996. Divisão do trabalho e relações sociais do sexo: Pensando a realidade das trabalhadoras do cuidado de saúde. In: Gênero e Saúde (M. J. M. Lopes, D. E. Meyer, & V. R. Waldow, org.), pp. 55-62, Porto Alegre: Artes Médicas. [ Links ]
MINAYO, M. C., 1992. O Desafio do Conhecimento: Pesquisa Qualitativa em Saúde. São Paulo: Editora Hucitec/Rio de Janeiro: ABRASCO. [ Links ]
SECRETARIA MUNICIPAL DE SAÚDE DE PORTO ALEGRE, 1999. Porto Alegre em números: Dados populacionais, nascidos vivos e mortalidade, 1998. Prá Saber: Informações de Interesse à Saúde, 4:21. [ Links ]
SAFFIOTTI, H., 1994. Violência de gênero no Brasil atual. Estudos Feministas, número especial:443-461. [ Links ]
SANT'ANNA, A. R., 2000. Vulnerabilidade ao Homicídio: Sócio-História das Mortes Violentas entre os Adolescentes na Cidade de Porto Alegre em 1997. Dissertação de Mestrado, Porto Alegre: Escola de Enfermagem, Universidade Federal do Rio Grande do Sul. [ Links ]
SCOTT, J., 1990. Gênero: Uma categoria útil de análise histórica. Educação e Realidade, 15:5-22. [ Links ]
WILLIS, P., 1991. Aprendendo a ser Trabalhador: Escola Resistência e Reprodução Social. Porto Alegre: Artes Médicas. [ Links ]
Submitted on 5 May 2001
Final version resubmitted on 23 August 2001
Approved on 17 January 2002