Abortion in Brazil: A Household Survey Using the Ballot Box Technique

Under criminalization and stigma, an unspoken experience cannot simply be treated as nonexistent. Debora Diniz and Marcelo Medeiros’s 2010 article reports initial results from Brazil’s National Abortion Survey, making confidentiality part of measurement itself. It moves abortion from a moralized exception toward an experience requiring public-health knowledge. That move also requires precision about who the study actually covers.

The stratified household survey interviewed 2,002 literate women aged 18–39 in urban Brazil. Participants completed abortion questions themselves and placed the forms in a box; female interviewers collected demographic and social information. Codes paired the questionnaires without identifying individuals. The methods section makes confidentiality a condition for reducing disclosure risk and improving reliable responses, rather than an optional courtesy after an interview.

Counting women’s experience

Fifteen percent reported ever having an abortion, rising to 22% among those aged 35–39. The unit is women reporting experience, not the number of abortions or the incidence within 2010. One person may have had several abortions; older respondents have also had more time to accumulate experience. Turning these percentages into annual numbers, or treating age as causing abortion, would change the question the study answers.

The questionnaire sought induced abortion, but the authors explicitly acknowledge that its short wording may include some spontaneous losses. This is a substantive measurement limit. Confidentiality, comprehensibility and precise classification are not always simultaneously maximized; design must negotiate among them. Anonymity can assist disclosure without making every respondent understand a word identically.

Among the 296 women reporting abortion, 48% reported medication use for the most recent abortion and 55% reported hospitalization because of it. These denominators cannot become all 2,002 respondents, and a most recent experience cannot represent every abortion across a lifetime. The survey collected no medication names, so medication use cannot establish a particular drug’s frequency or effectiveness. The authors develop a public-health argument, not an individual treatment protocol.

The boundaries of anonymity also matter

Self-completion requires literacy, and urban sampling excludes rural women. Calling a survey national does not remove those exclusions; omitted populations may face different resources, risks and care arrangements. The authors also caution against making education a direct cause of the observed association with abortion experience: age composition and other social relationships complicate that inference.

The contribution exceeds a single prevalence figure. It shows how legal and moral environments participate in producing knowledge: safe conditions for answering affect which experiences are counted, while available confidentiality tools affect who can enter a sample. A feminist reading can recognize the work of protecting participants while retaining attention to people missing from the statistics.

Luna and Luker’s review of reproductive justice connects health experience to criminalization and unequal resources. Ross’s article on intersectional reproductive-justice activism asks who can set the agenda. Statistical evidence can support institutional debate without deciding how people should live, or postponing the needs of excluded groups until more data arrive.