Photo Elicitation: Interviewing With Participant-Generated Images

Photo elicitation is a qualitative interview technique in which photographs are used to prompt and structure discussion, rather than relying on a participant’s unaided verbal recall. In its participant-generated (or “auto-driven”) form, the participant takes the photographs themselves — typically over a set period, guided by an open prompt tied to the research question — and then sits down with the researcher for a one-on-one interview in which each photograph is examined and its meaning narrated. The photographs are a route into the interview, not the end product; the primary data is the participant’s talk about the images.

Photo elicitation vs. photovoice: who takes the photos, and why it matters

Photo elicitation and photovoice are frequently confused because both put a camera in a participant’s hands, but they are methodologically distinct:

• Photo elicitation is an interview-enhancement technique. The images (participant-taken, researcher-supplied, or archival) exist to stimulate and deepen a one-on-one interview. The unit of analysis is usually the individual participant’s interview transcript.
• Photovoice is a Community-Based Participatory Research method built around Wang and Burris’s 1997 framework. It is structurally group-based, oriented toward critical dialogue and social action (the SHOWeD prompt sequence, policy-maker presentations), and treats participants as co-researchers throughout, not just interview subjects.

A useful way to hold the distinction: every photovoice study uses a form of photo elicitation to interpret the images, but not every photo-elicitation study is photovoice. This guide covers the interview protocol directly — for the participatory, group-based, advocacy-oriented method, see the photovoice guide.

Origins: Collier (1957) and Harper (2002)

The technique traces to John Collier Jr.’s 1957 paper “Photography in Anthropology: A Report on Two Experiments,” published in American Anthropologist (vol. 59, issue 5), which compared interviews conducted with and without photographs as prompts and found the photo-supported interviews longer, more detailed, and less prone to interviewer fatigue effects. The method’s most-cited methodological statement is Douglas Harper’s “Talking about pictures: a case for photo elicitation,” published in Visual Studies (vol. 17, no. 1, 2002). Harper’s central argument is that images engage a different mode of cognitive processing than words alone — he points to the relatively older, more basic neurological pathways that process visual information compared with language — and that this difference is precisely why photo-elicitation interviews tend to surface material that a purely verbal interview does not reach: participants report the photographs jog memories, associations, and emotional responses that a direct question doesn’t.

A practical protocol for participant-generated photo elicitation

1. Frame the research question and the photo brief

Write the prompt participants will photograph to, and pilot it before fielding. A prompt that is too narrow (“photograph your medication routine”) constrains the interview to what you already expected to hear; a prompt that is too broad (“photograph your life”) produces images with no shared frame for comparison across participants. Most published protocols sit in between — e.g., “photograph the things, places, and moments that matter to your experience of [the phenomenon under study] over the next two weeks.”

2. Recruit, consent, and equip participants

Informed consent for a photo-elicitation study has to cover two separate acts: participating in the interview, and taking photographs that may include other people, private spaces, or identifiable locations. Cover both explicitly, and address upfront what happens if a participant photographs an identifiable third party who hasn’t consented (see the ethics section below). Smartphone cameras have made provisioning a study-owned camera optional in most contexts, but decide deliberately — a provided device standardizes image format and resolution and avoids excluding participants without a capable phone.

3. The shooting period

Give participants a defined window (commonly one to a few weeks in published protocols) and a rough image-count guide rather than a hard cap, so participants aren’t rationing shots against an arbitrary limit. The researcher is deliberately absent during this stage in the participant-generated variant — that’s what separates it from researcher-driven photo elicitation, where the researcher selects or takes the images in advance.

4. The photo-elicitation interview itself

The interview is typically semi-structured (see designing semi-structured interviews): the participant selects images to discuss (their own selection is itself data — what they chose to include or leave out), narrates what each one shows and why they took it, and the researcher probes with open follow-ups (“what was happening just before this?”, “what do you notice looking at it now?”) rather than imposing an interpretation. Photographs function as a third point of shared attention between researcher and participant, which softens the direct interviewer-interviewee dyad of a standard interview and is frequently reported to ease rapport, particularly on sensitive topics or with participants who find open-ended verbal recall difficult.

5. Third-party image consent and data handling

Any identifiable person appearing in a participant’s photograph who is not the participant needs their own consent pathway before that image is used in any output, or the image needs to be excluded, cropped, or the face obscured. Build this into the protocol before fieldwork starts, not as a post hoc fix: decide whether participants will seek on-the-spot consent from anyone they photograph, whether images with unconsented identifiable third parties are usable for the interview but not for publication, and how images are stored and retained separately from (but linked to) the interview transcript.

What photo elicitation adds over a standard interview

The specific analytic value photo elicitation is used for is access to tacit, embodied, or otherwise hard-to-verbalize experience — the parts of an experience a participant might not think to mention, or might struggle to put into words unprompted, but which the concrete presence of a photograph brings back into reach:

• It bypasses the “tell me about X” recall problem. A standard interview asks a participant to generate relevant material from memory on demand; a photograph re-presents a specific moment, which reliably surfaces sensory, spatial, and incidental detail (what was in the background, what it smelled like, who else was there) that a general question rarely elicits.
• It disrupts the researcher’s assumed frame. Because the participant chooses what to photograph, the resulting image set can surprise the researcher — surfacing categories of experience the interview guide never anticipated, which is exactly the kind of information gain a pre-structured question list forecloses.
• It reduces the direct interrogation dynamic. Talking about a shared object in front of both parties is a different social act than answering a direct personal question, and is widely reported to lower the perceived power differential of the interview, particularly valuable in research with children, marginalized groups, or stigmatized topics.
• The participant’s selection and framing choices are a second data stream. What a participant chose to photograph, what they left out, and how they order or prioritize the images during the interview are analyzable alongside what they say about them — a layer of data a purely verbal interview cannot produce.

Analyzing photo-elicitation data

The primary analytic object is the interview transcript — the “photo-talk” — not the images considered in isolation. Most published studies apply standard qualitative coding to the transcript (see thematic analysis), treating each image-discussion segment as a coded unit the same way any interview segment would be coded. Some studies add a secondary visual-content pass over the images themselves (what is depicted, framing, recurring subjects across the sample) to triangulate against the verbal narrative, but this is a supplement to, not a replacement for, coding what the participant actually said about each photograph — the photographs are prompts for talk, and it is the talk that carries the participant’s meaning-making, which a researcher’s own reading of the image cannot substitute for.

Ethical considerations specific to participant-generated images

• Third-party consent (covered above) is the most common ethics-review sticking point specific to this method, distinct from standard interview consent.
• Private and sensitive locations. A photo brief tied to daily life will often produce images of homes, workplaces, or other spaces the participant would not otherwise document for a stranger — review boards frequently ask for an explicit protocol on what participants should avoid photographing (medical information, other people’s private property, anything illegal) as part of the briefing.
• Publication use. Decide, and state in the consent form, whether images may appear in publications/presentations, whether faces will be obscured by default, and whether a participant can retroactively withdraw a specific image after the interview even if they don’t withdraw from the study.
• Minors. Photo-elicitation studies with children raise the same third-party and location concerns at higher stakes, plus parental/guardian consent for both the interview and the photography itself — treat as a distinct ethics-review pathway, not a variant of the adult protocol.

When to choose photo elicitation vs. photovoice vs. a standard interview

Reach for participant-generated photo elicitation when the goal is individual, in-depth access to a participant’s own meaning-making about a phenomenon they can photograph, and a one-on-one interview format fits your design. Reach for photovoice instead when the project is explicitly participatory, group-based, and aimed at collective critical reflection or advocacy/policy outcomes — photovoice’s group discussion and action orientation are core to the method, not optional add-ons. Reach for a standard interview alone when the phenomenon under study doesn’t lend itself to photographing, or when logistics (a single interview session, no lead time for a shooting period) rule out the extra fieldwork stage a photo brief requires.

Frequently asked questions

What’s the actual difference between photo elicitation and photovoice?

Photo elicitation is any use of images (participant-taken, researcher-supplied, or archival) to stimulate an interview; photovoice is a specific, group-based, participatory-action method built on Wang and Burris’s 1997 framework. Photovoice studies use a form of photo elicitation to interpret their images, but most photo-elicitation studies are not photovoice — they’re individual interviews, not a participatory group process aimed at social action.

Does the researcher or the participant take the photographs?

Either, depending on the variant. This guide covers participant-generated (“auto-driven”) photo elicitation, where the participant photographs independently before the interview. In researcher-driven photo elicitation, the researcher selects or takes the images in advance (including archival or historical photographs) and uses them to prompt discussion.

How many photographs should participants take?

Give a rough guide rather than a hard quota — enough to give the interview real material without turning the shooting period into a burdensome task. What matters more than a specific count is that the prompt is piloted first, so participants aren’t guessing at what “counts” as relevant.

Do participants need a special camera?

No — a personal smartphone is standard in most current protocols. Providing a study camera is a deliberate design choice (for image-format consistency or to avoid excluding participants without a capable phone), not a methodological requirement.

Is photo elicitation a qualitative or quantitative method?

Qualitative. It produces interview transcripts (and a set of images) analyzed for meaning, not a numeric measure — it sits alongside methods like standard interviews and other qualitative approaches in a study’s methods toolkit.

See also: photovoice, designing semi-structured interviews, thematic analysis, and the research methods pillar for the broader qualitative-methods landscape.