Trans Women, Travestis and Violence in Brazil

Including trans women in violence statistics means questioning institutional power over categories and the naming of harm. The Brazilian Estudo Preliminar de Violência contra as Mulheres Trans e Travestis makes both problems visible. Produced by DataSenado in partnership with the Observatório da Mulher contra a Violência, the report combines accounts of harm with an unusually revealing description of classification and correction. Inclusion is a process involving questions, interviewers and institutions, rather than the simple addition of an identity label.

The report published on March 7, 2024 concerns 21 Brazilian trans women and travestis aged 16 or over, contacted by telephone between August 21 and September 25, 2023. It explicitly describes the study as exploratory. Random telephone-number contact does not by itself establish representative inference: the researchers lacked stable population information needed to calibrate estimates. These interviews therefore cannot supply a national prevalence rate or a reliable comparison of rates with cis women.

Who reaches the questionnaire?

The question routing on page 4 distinguishes a person’s present identification from registration at birth. That distinction matters, but its operation is not self-evident. The report describes six cis women initially classified as trans because their birth registrations had contained clerical errors. It also identifies interviewer coding errors and a case in which sexual orientation was confused with gender identity. Neither administrative records nor an interviewer’s interpretation can substitute transparently for a person’s account.

Auditing corrected classifications, rather than simply treating a database field as a fact. The report states that all interviews marked as involving trans women or travestis were reviewed; this does not establish that everyone wrongly excluded elsewhere was found. Its note about ordinary telephone practice is particularly instructive: interviewers may infer sex from voice or from gendered marital-status words. A feminist methodological question follows from that description: whose participation can be lost before a supposedly neutral questionnaire begins?

The actual routing also has boundaries. Participation required identification as female and an age of at least 16; the resulting study cannot stand for every nonbinary person, younger person or gender-diverse experience. Travestis appears as a social and political self-identification in the Brazilian context, not a synonym for wearing particular clothes. Respecting that specificity is different from claiming that one sequence of questions fully captures everyone’s identity.

Naming violence is not the same as listing an act

Nine participants reported domestic or family violence at some point in their lives; five of those reported such violence in the preceding twelve months. The study then asked all participants about thirteen concrete situations during the preceding year, including humiliation and threats by someone in an intimate or family relationship. These are different questions and time windows. The number of listed situations is not a number of victims, and lifetime declarations cannot simply be subtracted from recent behavioral responses to calculate an underreporting rate.

One exchange makes the distinction tangible. A participant describes public humiliation as normal, then recounts being called a man despite identifying as a woman. The report also discusses damage to her hair extensions as property violence connected to gender identity. Ordinary objects and apparently familiar words can carry harm that a broad question fails to elicit. This does not authorize a researcher to diagnose a person or dismiss her self-description; it asks whether the available vocabulary makes an injury easier or harder to recognize.

Knowing a protection service exists is another matter from using it or receiving support there. The section on services collects knowledge and perceptions, not a demonstration that institutions reliably welcomed the participants. References to law belong to the report’s 2024 context, rather than a current legal guide. The distinction prevents a formal route of protection from being treated as proof of practical access, just as recognizing an act does not establish that reporting it will be safe.

Statistical absence is an institutional question

The report’s closing discussion connects missing population information with the difficulty of making statistically valid claims. It also calls attention to information collected through another household member, whose response may hide an identity they refuse to acknowledge. These are the report’s arguments about its research setting, not proof that Brazil’s entire statistical infrastructure has remained unchanged since 2024. The useful question concerns how an institution can recognize people in one setting while making them difficult to count in another.

More collection alone does not resolve that contradiction. Privacy, the ability to refuse and control over disclosure remain important, especially when a household or interviewer may be unsafe. Nor should the limitations of 21 interviews make their accounts disposable. They can identify problems in categories, interaction and recognition without carrying the burden of representing a whole population. Treating situated testimony seriously and refusing unsupported national estimates are compatible responsibilities.

Data Feminism offers a framework for asking who defines categories and performs the labor of correction; this report supplies a concrete institutional encounter in which to use those questions. Japan’s survey of violence experiences offers a separate comparison of behavioral wording, self-recognition and help-seeking. Their samples and populations should not be combined. Reading them together makes inclusion a sustained methodological responsibility: who can answer, what can be described, and what consequences follow after an experience enters the record?