Photovoice was developed by Wang and Burris in work with village women in rural China, at first under the name photo novella (Wang and Burris, 1994). Its stated goals are to enable people to record their community's strengths and concerns, to promote critical dialogue through group discussion of photographs, and to reach policymakers (Wang and Burris, 1997). The method draws on the problem-posing education of Freire (1970), on feminist theory and on documentary photography. It shares with photo elicitation the insight that talking about a picture produces a different kind of account from answering a question (Harper, 2002), and differs in placing the camera in participants' hands. A review of photovoice projects in health found wide variation in how fully participants were involved at each stage (Catalani and Minkler, 2010), which is a reminder that the label does not guarantee participation.

Belonging is more than a feeling. Yuval-Davis (2006) separated belonging, as emotional attachment and feeling at home, from the politics of belonging, the boundary work by which some are counted as members and others are not. Antonsich (2010) drew a similar line between a personal sense of being at home in a place and the discourses that grant or refuse inclusion. A method for studying belonging among migrant students should be able to register both: the bench where a friend waits, and the form that asks for a nationality.

Article 12 of the Convention on the Rights of the Child gives children the right to express views on matters affecting them and to have those views given due weight (United Nations, 1989). Hart (1992) warned that much of what is called children's participation is decoration or tokenism. Lundy (2007) argued that voice alone does not satisfy Article 12, and that children also need a safe space in which to form views, an audience that listens, and influence over what follows. These four elements, space, voice, audience and influence, are used below as a test of each stage.

Work begins with the school's leadership, the staff who teach newly arrived students, and where possible a community organisation trusted by migrant families. Three matters are settled in writing before any student is approached: that taking part or declining will have no bearing on grades or standing; that teachers will not be present in discussion sessions; and that the school commits in advance to receive and answer what the students present. Review by a research ethics committee is sought on a protocol that covers the image-specific issues set out below.

Invitations go to all eligible students in the chosen year groups, not to those picked by staff, so that confident speakers are not the only ones heard. Information is given orally and in writing in the languages of the families concerned, with an interpreter present. Consent from a parent or guardian and the young person's own assent are both required, and assent is treated as continuing: it is confirmed at the start of each session. Following Wang and Redwood-Jones (2001), separate permissions cover taking part, being photographed by others, and the release of each image for display.

The first two sessions introduce the idea of the project, basic composition, and the ethics of taking pictures: asking before photographing a person, what to do if someone refuses, and how to convey an idea without showing a face, for instance through objects, places, hands or shadows. Phone cameras are adequate, and devices are lent to anyone without one. The group also agrees its own rules for discussion, including which languages will be used and how translation among members will work.

Each cycle starts from a prompt agreed with the group. Prompts are concrete and open, such as places in school where the student feels at ease, places avoided, and something that would be changed. The word belonging itself may have no close equivalent in a student's first language, and the facilitator works with the group to find phrasings that carry across. Students take photographs over a week and bring a few of their choice to the next session.

Each student presents chosen images and the group discusses them. The questions known by the mnemonic SHOWeD give a structure: what is seen here, what is really happening, how it relates to our lives, why the situation exists, and what can be done about it (Wang, 1999). They move discussion from description towards causes and action, in keeping with the method's origins in Freire (1970). Students write or dictate a caption for each image in the language they prefer. Sessions are audio recorded with permission and facilitated by a researcher who is not a member of the school's staff, working with an interpreter or a bilingual co-facilitator.

In photovoice, participants take part in interpreting the material, through selecting images, placing them in context and grouping them into themes (Wang and Burris, 1997). In practice the group lays out printed photographs with captions, sorts them into piles, names each pile and discusses what is missing. The researcher then carries out a fuller thematic analysis of transcripts and captions using an established approach (Braun and Clarke, 2006) and brings the result back to the group for correction.

The group decides whether, where and to whom its work is shown. Options range from a closed presentation to school leaders to a display open to families. Each image is cleared again by its maker before use. The school's response, given at a later meeting, completes the cycle. In the terms of Lundy (2007), this stage supplies audience and influence, without which the earlier stages amount to an exercise.