← All insights

What photovoice revealed about climate displacement

The study took place in informal settlements in Chittagong. These had grown considerably over the preceding decade as climate pressure in the coastal chars and river-erosion zones of southern Bangladesh pushed families toward the city. Standard demographic surveys of the settlements existed, and BRAC and UN-Habitat had both done coverage mapping. Nobody had a detailed account of what displaced families experienced once they arrived.

We used photovoice. The participants were women from fifteen households across three settlement clusters. Each was given a camera and a broad brief: document your daily life, and particularly the moments where you encounter difficulty accessing something you need. The photographs were the basis for group discussion sessions, organised around what each image showed and why the photographer had chosen it.

Photovoice is usually justified as a way of giving voice to marginalised communities. Its methodological effect is more specific. It changes who controls the research agenda at the point where data are generated. A survey questionnaire asks about what the researcher has already decided to measure. A photovoice protocol asks participants to decide what is worth documenting, and that question produces different findings.

The Chittagong findings differed from the survey data in two ways. The first concerned documentation. Survey questions about barriers to service access usually draw answers about distance, cost and provider behaviour. Several participants had instead photographed encounters at clinics, schools and water distribution points where they had been turned away because they could not produce address proof or identity documents. Displacement had cut their link to formal records. They held NID cards issued for their home districts, and these were not accepted as address verification in Chittagong. Children could not enrol in school. Mothers could not use maternal health services in the public system without a locally recognised identity document. The administrative system had not been designed for families whose documents were issued somewhere else.

The second concerned exclusion inside the settlements. The photographs recorded informal hierarchies: which households had established relationships with local landlords and which had not, which women could use the tube-well at peak hours and which had to wait, and whose children played in which spaces. None of this appeared in facility-level data, which recorded how often the facilities were used and nothing about the social filters that decided who used them.

The group discussions added what the photographs alone could not. Participants challenged each other's accounts, proposed other explanations for images, and sorted the problems into those they thought collective action could address and those they considered fixed. That sorting mattered for programme design. Some of the barriers were structural (documentation policy, clinic catchment rules) and needed advocacy. Others were relational (exclusion from water access, children's play spaces), and the community was already negotiating them.

Photovoice data is rich and hard to aggregate. The method does not suit the systematic coding that yields publishable frequencies, and researchers who force it into that mode usually end up with weak qualitative analysis and weak quantitative description at once. The appropriate output is thick description: case material, direct participant quotation, and documented patterns that illustrate what participants experienced without representing a wider population. Commissioners who want generalisable findings need to understand this before the study begins.

Bangladesh climate displacement data from Internal Displacement Monitoring Centre (IDMC) and UNHCR operational data. Chittagong settlement coverage data from UN-Habitat's Bangladesh Country Programme. Documentation exclusion patterns corroborated by BRAC research on urban informal settlement access.