Visual Elicitation Thematic Analysis — Image-Driven Qualitative Inquiry

Either approach is legitimate and each has trade-offs. Participant-generated photographs (photo-voice) give participants more agency and tend to produce personally meaningful images, but introduce ethical complexity around privacy and consent for identifiable third parties. Researcher-supplied images are easier to manage ethically and useful when participants cannot easily produce images, but the researcher must justify the image selection. The choice should be driven by the research question and ethical context, and documented in the methods section.

Standard thematic analysis works solely from verbal data — interview transcripts, focus group recordings, or documents. VETA adds a visual stimulus layer: the images actively shape what participants say and become an additional layer of analytic data. Coding explicitly tracks which image prompted which account, enabling the analyst to interrogate the relationship between visual and verbal meaning. This makes VETA particularly suited to embodied, spatial, or experiential topics where verbal prompts alone are insufficient.

The same guidance that applies to thematic analysis generally applies here: 6–30 participants is the typical range, with conceptual saturation (the point at which new interviews add no new themes) as the key criterion. Because VETA interviews tend to be richer and more detailed than standard interviews, smaller samples often reach saturation earlier. Fewer than six participants makes it very difficult to identify patterns that transcend individual accounts.

Yes. Braun and Clarke's thematic analysis framework accommodates both inductive (data-driven) and deductive (theory-driven) approaches, and VETA inherits this flexibility. A deductive VETA study would specify a theoretical framework in advance and apply it to both the visual stimuli selection and the coding process. Researchers should be explicit about which orientation they have taken and why, as the two orientations have different epistemological implications.

Any image that contains identifiable individuals requires explicit consent from those individuals before publication, even if the participant who photographed it has consented. Common strategies include anonymising images (cropping, blurring faces), obtaining additional consent forms for identifiable third parties, or substituting illustrative images that are not the actual data images. Researchers should establish a confidentiality protocol at the study design stage and document it in the ethics application.