Project Overview
The study looked at women’s experience with household water access, its possible link with violence against women, and the policy/services responding to these issues in Peru and Indonesia. The research questions include participant perceptions about water scarcity in the area, the possible link between water issues and women’ health, development and implementation of policies and programs to address water insecurity, the involvement of women in water-related program in the area, and institutional collaboration or inter-sector responses to solve problem regarding water insecurity in the area. To help answer these questions, the researchers collected data from relevant stakeholders in the target areas using semi-structured interviews with community members, health and water management practitioners, and other stakeholders in Indonesia (n = 24) and Peru (n = 51). Focus groups were also held with men and women, separately, in both locations (approximately n = 10 participants in each focus group) and a multistakeholder meeting (approximately n = 25 participants) was used to present results to community members and receive feedback from local and regional stakeholders.
Data Description and Collection Overview
The data in Indonesia were collected from October to November 2020 (Phase 1) and October 2021 (Phase 2); the data in Peru were collected in December 2021 (Phase 1) and March 2022 (Phase 2). Interviewed were women who report problems with water security and household tensions; healthcare workers who have knowledge of gender-based violence in the area; local community leaders; local non-governmental organizations working in water/women protection issues; and local government representatives responsible for water management. The interviews were audio-recorded. Audio-recorded interviews were examined, and selected portions were transcribed and translated into English from local languages.
Qualitative data were also collected from four workshops in both Indonesia and Peru. These workshops were with local women, men, healthcare workers, and water management leaders. The workshops were used to share findings and collect data on action planning. Selected parts of the workshops were audio-recorded. Audio-recorded segments were examined, and select portions were transcribed and translated into English from local languages. Recordings from both the interviews and workshops were saved as free lossless audio coding (.flac). Qualitative textual data was saved in Word.
An additional survey was administered to approximately 700 women in both Indonesia and Peru. Adult women from randomly selected households were asked if they would like to participate in the study. Data was collected from these women and was entered into a field-friendly tablet offline. The data was pseudo-anonymized immediately, meaning participants were assigned a code at the beginning of the survey and their name was only recorded on a master key that was destroyed once data collection concluded. This master key linked participant codes to names. The quantitative data was then entered into Excel. Once the data was cleaned, it was exported into STATA for analysis.
During data cleaning and analysis, all data were anonymized to the fullest extent using the method of generalization hierarchy. This method is considered appropriate for cross-sectional surveys with a moderate risk threshold. The generalization hierarchy was applied for indirect identifiers including respondents' socio-demographic information. The representation of age was changed into 5 year intervals; marital status was presented as married and never married; number of children was changed into 2 member intervals; number of household members was changed into 3 member intervals; location was changed from village to sub-district, and then district; educational level was changed into 3 different levels (below or equal to primary education, secondary education, and tertiary education); household income was changed into 3 groups (low, middle, high) for all data that is being released publicly. Women were also able to place themselves on a socioeconomic ladder.
Data collection instruments were created in collaboration with international and local non-governmental and governmental partners to ensure they meet community standards for conducting research on sensitive topics. Data collection instruments were piloted with local focus groups to ensure they were culturally appropriate before used in the full-scale research project. All data processing was conducted by team research staff and did not include third parties.
At the end of each day, researchers and assistants downloaded the quantitative data and audio-recordings, and saved typed qualitative data onto a password-protected computer that was securely stored by the principal investigator and co-investigators. Only the research staff had passwords to access the data. Any textual documents were kept in a locked safe. Every 2-3 days, the PI or Co-PIs accessed a secure internet connection to digitally back up coded files to their university-provided OneDrive for Business, which required a secure authentication process to access. The qualitative and quantitative data within OneDrive were fully anonymized by redacting identifying information such as locations.
Selection and Organization of Shared Data
The documentation files consist of participant consent forms for both Peru and Indonesia, an interview questionnaire, and a QDR-created document containing the Spanish and English definitions/codes used in the interviews. Additionally, an Excel document was created to organize key quotes that pertained to themes of interest. In order to select these quotes, two researchers read through the original transcripts and used the qualitative data management program, Dedoose, to tag relevant lines. These excerpts were then inserted into an Excel file, allowing the qualitative data across the two countries to be compared.
