Photovoice is a participatory qualitative research method in which participants take their own photographs of their lived experience, environment, or community in response to a research question, and then discuss those photographs in a structured group process to generate data and identify themes for action. Unlike photo-elicitation, where a researcher supplies or selects images for participants to react to, photovoice puts the camera in participants’ hands: they decide what to photograph, and their captions and group commentary become the primary data, alongside the images themselves. It was developed as a Community-Based Participatory Research (CBPR) technique and is used across public health, community development, education, and social work research wherever a study wants participant-controlled, visual evidence of everyday conditions rather than researcher-mediated description.
Origins: Wang and Burris (1997)
Photovoice was formally introduced by Caroline Wang and Mary Ann Burris in their 1997 article “Photovoice: Concept, Methodology, and Use for Participatory Needs Assessment,” published in Health Education & Behavior (24(3), 369-387), building on earlier fieldwork with rural women in China’s Yunnan Province. Wang and Burris drew the method’s theoretical basis from three strands: Paulo Freire’s critical consciousness-raising pedagogy, in which dialogue about visual and textual material helps people identify and act on the structural conditions of their lives; feminist theory’s attention to whose standpoint produces knowledge and whose images get taken and circulated; and documentary photography’s tradition of using images as social evidence, redirected here so participants are the photographers rather than the photographed.
Wang and Burris described three goals for the method: to enable people to record and reflect their community’s strengths and concerns, to promote critical dialogue and knowledge about personal and community issues through group discussion of photographs, and to reach policymakers. That third goal is what distinguishes photovoice from a purely descriptive visual method — it is designed from the outset as advocacy-oriented, participatory research, typically closing with a public exhibition or presentation to an audience with some capacity to act on the findings.
The Photovoice Process, Step by Step
Implementations vary, but most follow a recognizable sequence derived from Wang and Burris’s original protocol:
- Recruit and train participants. Participants are recruited from the community or population the study concerns. Training covers camera or phone operation, basic composition, and — critically — the ethics of photographing other people (see Consent and Image Ethics below), usually formalized with a written agreement participants sign before fieldwork begins.
- Introduce the prompt or theme. The researcher poses an open question tied to the study’s aim (for example, “What helps you stay healthy in this neighborhood?” or “What makes it hard to get to class on time?”) rather than a checklist of subjects to photograph, so participants retain interpretive control over what counts as a relevant image.
- Photograph over a defined period. Participants take photographs, typically over one to several weeks, documenting whatever they judge relevant to the prompt in their daily environment.
- Select and caption. Each participant selects a small number of their own photographs — often those they consider most significant — and writes or dictates a short caption or story explaining what the image shows and why they chose it.
- Discuss in groups using SHOWeD. Participants present their selected photographs to the wider group, and the group works through the images using a structured discussion method (SHOWeD, detailed below) that moves from description toward analysis and action.
- Codify. Wang and Burris describe three overlapping ways researchers and participants codify what emerges from the discussions: identifying recurring issues or themes, identifying representative stories that illustrate those themes in participants’ own words, and identifying theories — patterns that explain why a condition exists, which point toward what could change it.
- Exhibit and disseminate. The project typically closes with a public exhibition, report, or presentation of selected images and captions to community members, service providers, or policymakers, which is also where photovoice’s advocacy goal is most visible relative to other qualitative methods.
The SHOWeD Discussion Method
SHOWeD is the most widely used discussion structure for working through photovoice images in groups. It is a mnemonic for a sequence of five questions asked about each photograph, moving participants from literal description toward critical reflection and action:
| Letter | Question | Purpose |
|---|---|---|
| S | What do you See here? | Establishes a shared, literal description of the image’s content before interpretation begins. |
| H | What is really Happening here? | Moves from surface description to the participant’s account of the situation behind the image. |
| O | How does this relate to Our lives? | Connects the individual photograph to shared experience across the group, not just the photographer. |
| W | Why does this situation, concern, or strength exist? | Pushes discussion toward underlying causes rather than stopping at description. |
| D | What can we Do about it? | Orients the group toward the method’s action/advocacy goal, generating the material an exhibition or report will present. |
Facilitators generally work through all five questions for each photograph rather than treating them as optional prompts, since the sequence itself — description, then context, then relevance, then cause, then action — is what produces analyzable data rather than a simple photo caption.
Consent and Image Ethics
Photovoice raises a layered informed consent problem that most interview- or survey-based qualitative methods do not: photographs frequently capture identifiable third parties — neighbors, family members, strangers in public space, children — who never consented to be part of the study at all.
Most protocols distinguish three separate consent obligations:
- Participant consent to take part. Standard research consent covering the study’s purpose, the training and photography task, how captions and discussion will be recorded and used, and how images may be exhibited or published, including whether the participant’s own name or only their photographs will be attributed.
- Third-party or bystander consent. Because photovoice sends participants into everyday settings rather than restricting them to a lab or interview room, ethics review boards typically require participant training on when a photograph subject’s consent is needed before the photograph is taken or before it can be used in reporting — commonly enforced with a photo-release form participants carry and ask identifiable subjects to sign, and a study rule that images of identifiable people who have not consented are either not taken, not retained, or are altered (blurred, cropped) before analysis or exhibition.
- Minors and vulnerable subjects. Where children, patients, or other subjects with limited capacity to consent may appear in frame — a common issue in school, family, or clinical settings — protocols generally require a parent or guardian’s separate consent, or a rule against identifiable images of that population, decided in advance with the IRB or equivalent ethics committee rather than left to participants’ judgment in the field.
Because photovoice’s endpoint is often public exhibition rather than a confidential dataset, ethics review typically also covers exhibition consent separately from participation consent: a participant may agree to take part in the study and the group discussion while declining to have specific images or their name displayed publicly, and protocols should give them that option at the point of selecting images for exhibition, not only at intake.
Analyzing Photovoice Data
Photovoice produces two linked data streams for each selected photograph — the image itself and the participant’s caption or SHOWeD discussion transcript — and analysis typically treats them together rather than coding the image and the text independently. Common approaches include:
- Thematic analysis of the paired data. Captions and discussion transcripts are coded using standard qualitative coding procedures (see thematic analysis and worked examples of coding qualitative interview data), with the associated photograph retained as context for each coded segment rather than analyzed as a separate visual dataset.
- Wang and Burris’s codify step. As in the process outline above, researchers and participants work collaboratively toward themes, illustrative stories, and explanatory theories, rather than a researcher coding transcripts alone after the fact — the participatory element extends into analysis, not just data collection.
- Participant-led theme sorting. Some protocols have participants themselves group photographs by emergent theme during a workshop session, with the researcher’s role limited to facilitating and recording the categories participants arrive at, which functions as a form of ongoing member checking built into the method rather than a separate validation step added afterward.
- Exhibition as validation. Because the public exhibition presents participants’ own photographs and words back to the community, it also serves informally as a check on whether the analysis represents participants’ intent accurately — a structural echo of member checking that is distinctive to participatory visual methods.
Strengths and Limitations
| Strengths | Limitations | |
|---|---|---|
| Data control | Participants decide what is photographed and how it is captioned, reducing researcher framing of what counts as relevant. | Findings depend heavily on what participants chose (or were willing/able) to photograph; sensitive but important conditions may go undocumented if unsafe or embarrassing to photograph. |
| Access | Surfaces everyday conditions (housing, commute, workplace, environment) that an interview alone may not prompt a participant to describe in detail. | Requires camera/phone access, some technical comfort, and time for a multi-week fieldwork period, which can affect who is able to participate. |
| Ethics burden | Structured consent and training protocols exist and are well established from 1997 onward. | Third-party consent in public/uncontrolled settings is genuinely harder to guarantee than in an interview room; requires more IRB planning than most qualitative methods. |
| Output | Produces a public-facing exhibition or report suited to advocacy and stakeholder engagement, not just an academic write-up. | Analysis of paired image-plus-text data is less standardized than text-only coding, and image rights/consent add a data-management burden after collection ends. |
Photovoice vs. Related Visual and Participatory Methods
Photovoice is sometimes confused with photo-elicitation, but the two differ on exactly the point that matters most for research design: in photo-elicitation, the researcher (or an existing photo archive) supplies the images and participants are interviewed about their reactions to them; in photovoice, participants take the photographs themselves, which is what makes it a participatory method rather than an elicitation technique layered onto a conventional interview. Photovoice also overlaps with broader qualitative research methods such as focus groups (the SHOWeD discussion is essentially a structured focus group organized around images) and standard interviews (captions are frequently collected through a brief interview about each selected photograph). Researchers weighing photovoice against a purely text-based method such as grounded theory coding of interview transcripts should also budget for the additional ethics review and image-management steps described above, which text-only designs do not carry.
Frequently Asked Questions
How many photographs does a photovoice participant typically take or submit?
Protocols vary; participants commonly take considerably more photographs during the fieldwork period than they ultimately select, then narrow down to a small set — often in the range of a handful to a dozen — that they discuss with the group and consider most significant relative to the prompt. There is no fixed number specified by the original Wang and Burris protocol; researchers set the range in their own study design.
Do photovoice participants need to be trained photographers?
No. Photovoice is explicitly designed for participants with no photography background; the training step covers basic camera or phone operation and, more importantly, the ethics of photographing identifiable people, not photographic technique or composition.
Can photovoice be conducted with smartphone cameras instead of dedicated cameras?
Yes. The original 1997 protocol used disposable or point-and-shoot film cameras because that was the available technology; contemporary implementations commonly use participants’ own smartphones, which does not change the underlying method, though researchers should still address device access and data-transfer/privacy handling for images as part of consent and data management planning.
Is photovoice a qualitative or quantitative method?
Photovoice is a qualitative method. Its outputs — photographs, captions, and group discussion transcripts — are analyzed using qualitative techniques such as thematic coding rather than statistical analysis, though some studies count or tabulate theme frequency across participants as a secondary, descriptive step.
