Impact of a Privacy-Enhancing ACASI Survey on Menstruation Stigma Response Patterns in Ghana

How to cite

Furlani, Irene. 2026. "Impact of a Privacy-Enhancing ACASI Survey on Menstruation Stigma Response Patterns in Ghana." AEA RCT Registry. AEARCTR-0018075. https://doi.org/10.1257/rct.18075-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Irene Furlani*

* Primary investigator

Registry ID
AEARCTR-0018075
DOI
https://doi.org/10.1257/rct.18075-1.0
First registered
Last updated
Status
In development
Trial dates
2026-03-13 – 2026-04-03
Countries
Private
JEL codes
C83, C88, C93, J16, O12
Registry record
http://www.socialscienceregistry.org/trials/18075

Abstract

Menstruation stigma affects school participation and social inclusion. Evidence on menstruation stigma relies on face-to-face surveys collected with tablets in a Computer-Assisted Personal Interviewing (CAPI) setting. Estimates on a sensitive topic like menstruation stigma can be affected by the presence of the enumerator in a CAPI format. These biased estimates can flaw cost-effectiveness calculations for interventions addressing menstruation stigma. Policy decisions may be based on systematically biased measures.
In this study, we evaluate the effect of a privacy enhancing Audio Computer-Assisted Self-Interviewing (ACASI) survey method on reported levels of menstruation stigma. We created a new plug-in to implement the ACASI format aimed at minimizing the enumerator and third-party presence bias. The plug-in allows respondents to automatically play pre-recorded questions and select color-coded answers. Therefore, we compare the stigma levels around menstruation in surveys conducted in a CAPI format with those in our privacy-enhanced ACASI method.
We conduct a randomized survey experiment assigning senior high school girls in Northern Ghana to a stigma survey module, either in CAPI or ACASI format using SurveyCTO. This design allows us to estimate the differential effect of the privacy-enhancing ACASI format on levels of menstruation stigma, and in particular on school participation and social inclusion.
We hypothesize that our privacy-enhancing ACASI format increases the reported level of menstruation stigma, implying potentially higher cost-effectiveness for interventions addressing menstruation stigma.

Intervention

This is a randomized survey method design. The study analyses responses in a menstruation stigma section comparing the standard enumerator-led Computer-Assisted Personal Interviewing (CAPI) employed in the control group, to a privacy-enhancing Audio Computer-Assisted Self-Interviewing (ACASI) in the treatment group. Respondents in the treatment group can independently listen to pre-recorded questions and select color-coded answers. With the support of a software developer, I created a plug-in for SurveyCTO (the data collection platform used in this study) designed to increase respondent privacy and minimize potential bias introduced by the presence of the enumerator and third parties. The plug-in allows participants to listen to pre-recorded survey questions through headphones, instead of having the questions read aloud by an enumerator. Response options are presented using a simple color-coded interface, enabling participants to select a color rather than read or write text.

Experimental design

The eligible population for this study consists of senior high school girls aged 16 years and older attending selected schools in and around Tamale in Northern Ghana. To be eligible, respondents must have experienced menstruation at least once. The study sample will be drawn from students who participate in menstrual health information campaigns conducted in participating schools prior to the survey. The study is expected to include between 20 and 24 schools located within approximately a 50 km radius of Tamale. The anticipated total sample size ranges from 480 to 576 respondents. Participants will be randomly assigned to one of two survey methods, resulting in a sample of approximately 240 to 288 respondents per treatment arm.

Individuals will be randomly assigned to one of two survey methods for the menstruation stigma section. Randomization is implemented prior to data collection through a pre-generated treatment list that determines which survey module is loaded for each respondent. Enumerators are not informed about treatment assignment in advance, ensuring that the allocation process remains independent of enumerator behavior.

Both survey methods share the same introductory questions and overall structure. The section containing the stigma-related questions is always placed at the end of the questionnaire. Enumerators introduce this section in both conditions and explain the survey procedure to respondents. The color order of the answer options in also randomized.

Design details

Randomization method
Randomization done in office on Stata
Randomization unit
Individual (student)
Clusters
No cluster but we expect between 20 and 24 schools.
Observations
Between 480 and 576
Arms
Approximately 240 to 288 respondents per treatment arm.
Primary outcomes
Experienced stigma index, self-internalized stigma index, perceived societal stigma index
Secondary outcomes
In addition to the stigma indices, the analysis will consider a secondary outcome capturing respondents’ comfort with the survey process, the non-response rate, and each question of the stigma module.

Cited by

No citation data available for this registration yet.

Keywords