Best practices

This section brings together concrete examples of how researchers address ethical, legal, and methodological challenges in visual social research. The practices described here were provided by researchers and professionals in the field and aim to offer food for thought, practical solutions, and concrete insights at various stages of the research lifecycle.

Help us expand the best practices! If you have a best practice to share, use the button below to submit it

Best Practice Submission Form

6. Archiving & Preservation Phase - Alternative Descriptions:
Sensitive images stemming from couples’ and friends’ relationships were replaced by textual descriptions.

Author: Katharina Lobinger | USI Università della Svizzera italiana

Short Description

This is an example of a research project where images could not be shared on an open data repository for future use, but image descriptions were necessary to make interview transcripts and the verbal data understandable. For long-term preservation and open data, the project thus “Visualized relationships” built a dataset in which anonymized interview transcripts and contextual research files were prepared for future scholarly use, while highly sensitive participant visuals had to be replaced by written image descriptions.

Problem Description

The project “Visualized Relationships” investigated visual communication in close friendships and romantic partnerships and therefore works with particularly delicate materials: participant-created and/or participant-selected photographs, additional interviewer-produced images, interview audio files, observation notes, and network drawings. At the preservation stage, the project faced a central open data dilemma: visual materials are highly valuable for understanding visual communication practices, but as the visuals stem from close or intimate relationships, they are too sensitive to be made fully accessible on a data repository.

Implementation Strategy

The project team addressed this issue by preparing different categories of materials for different levels of long-term accessibility. First, the project established that anonymous transcripts would be made accessible on a data repository for secondary uses, accessible only to researchers with permission. Second, these transcripts were not archived as isolated text files. For each interviewed dyad (couple or friendship dyad), the research team created a case-based data package consisting of three interviews, metadata, photo descriptions, network drawing, and observation protocol. This means that future users will receive a richer anonymized analytical unit. Third, the project explicitly states that visual data cannot be made available in the repository for further research use for reasons of data protection and research ethics. Instead, descriptions of the images were produced and made accessible. While they do not fully substitute the visual material, they signal its existence and make the interview transcripts understandable. This procedure was also communicated to participants when obtaining consent.

Data & Metrics

Visuals that couples and friends use in their daily life. These visuals were either provided by participants (e.g. important or typical images used within the relationship) or photographs taken by the researchers with a dedicated camera (photographs of images that were shown during interviews).

What Worked Well

Not sharing visual data on repositories protected participants, some of which were vulnerable or in secret relationships that should not be exposed. Furthermore, in some cases large numbers of visuals were collected per dyad, which provides very personal and intimate insights into the relationship. These were relevant to consider during analysis but too sensitive to share for future uses.

Limitations & Challenges

The approach requires substantial additional documentation work because every excluded visual had to be translated into a short, written description. These descriptions do not present adequate substitutes for the visuals themselves but rather serve to make the interview transcripts understandable as they transcribed texts often refer to visual material. Another limitation is that the future analysis of visual data is not possible.

References & Links:
Lobinger, K., Lucchesi, F., & Tarnutzer, S. (2024). Visualized relationships (ViRe) - Image type overview and short descriptions. Zenodo. https://doi.org/10.5281/zenodo.10557227

Lobinger, K., Lucchesi, F., & Tarnutzer, S. (n.d.). Pair and individual interviews with couples and friends in Switzerland (ViRe) (Version 1.0) [Dataset]. FORS data service. https://doi.org/10.48573/FRMK-2S13

Tarnutzer, S., Lobinger, K., & Lucchesi, F. (2024). Image types revisited. A texto-material approach for creating image types. Studies in Communication Sciences, 24(1), 103–121. https://doi.org/10.24434/j.scoms.2024.01.3894