Photovoice Analysis with ATLAS.ti

February 20, 2014

Author:  Ricardo B.  Contreras

In this short summary paper, I will discuss how I used ATLAS.ti in the analysis of photovoice data within the context of an ethnographic study examining temporary migration and the value of labor.  The study is funded by a grant from the National Science Foundation (Cultural Anthropology program) and it involves data collection in communities of the departments of Chimaltenango and Santa Rosa, Guatemala, and in rural or semi-rural communities of the states of Michoacán and Sinaloa, Mexico.  In the first phase of the photovoice component of the study, cameras were given to six households in each of the two communities in Guatemala.  They were asked to take photographs of work-related activities in which they engage in daily life. These could include activities at home, such as taking care of children, cleaning, cooking, doing carpentry, or activities outside of home, such as working in the fields, carrying wood, and shopping.  Each participant (i.e., household) took between 25 and 29 photographs. A total of 320 photographs were analyzed.

First, a few words about the photovoice method.  Photovoice is defined in the literature as “…a qualitative research method in which participants use cameras to generate data (…) [thus] directly involving participants in the research process.  The photographs generated by participants become central artifact for discussion in an in-depth interview and or/focus group” (Novak 2010:292). In other words, photovoice allows for the elicitation of the point of view of participants represented in the form of still images of cultural scenes that they deem meaningful.  This is complemented by an in-depth interview in which the participant explains the cultural scene represented in the photograph and places it within the larger context of daily life and/or community experience.

Following, I will describe how ATLAS.ti was used in the analysis of these photographs and their accompanying in-depth interviews.   The following steps will be described: adding the photographs and accompanying interviews into the project; grouping them in primary document families according to participants; setting global filters by primary document families; using the multi-document view to place photographs and interview side-by-side; segmenting, coding, and hyperlinking quotations from photographs and corresponding interview; hyperlinking quotations from primary and secondary photographs; visualizing domains of analysis in network views; and writing memos.  These steps are represented in the following figure:

Steps

Figure 1. Steps in the photovoice analysis process

Setting up The Project

I added the photographs and corresponding interviews, in audio format, to the project library, and grouped them in primary document families according to participants, as follows:

Primary document family A=”Photovoice-Site-Identifier of the participant”.

Each primary document family contained all of the photographs and the single interview with the person who took those photos.  Inside of the Primary Document Manager, I set the program to show each primary document (i.e., photographs and interviews in audio format) as tiles, which facilitated the identification of photographs.  Additionally, I set the preview size to Extra Large. (Inside the Primary Document Manager go to View/Tiles and View/Preview Size/Extra Large).  The following screenshot shows this primary document family organization with the document view set to tiles:

PD family

Figure 2. Photovoice primary documents in Primary Document Manager

As I stated above, the interviews were not transcribed but rather added to the project as audio documents.  I did so because I realized that several of the photographs that a given participant took were mere repetitions of other photographs and their description did not contribute much to the understanding of the whole.  Thus, I preferred to transcribe them selectively, that is, only transcribe the descriptions that the participant made of the principal photographs, those that represented the core of a cultural scene.  Also, I decided to write down that description as a Comment of the photographic quotation, rather than as a primary document on its own.  The screenshot below shows a Comment on a photographic quotation showing the participant’s description of the cultural scene.

Figure 3. A 'comment' of a graphic quotation

Figure 4. A ‘comment’ of a graphic quotation

Segmenting, coding, and hyperlinking quotations. 

Global Filters and the Multi-Document View

Before loading a photograph and its corresponding interview, I proceeded to apply a global filter for the primary document family that contained that photograph and that interview.  That allowed me to focus on the photographs and interview for each participant at a time.  Using the multi-document view feature of ATLAS.ti, each photograph was loaded side-by-side its corresponding interview audio primary document.  Below, you will see a screenshot showing the applied global filter. Notice that the background of the space where the primary documents are shown turns to a yellowish color.  That is the ‘universal’ indicator that a filter has been applied. (A global filter is applied by right-clicking on the primer document family inside of the Primary Document Manager, and selecting ‘set global filter’).

Global filter

Figure 4. The Primary Document Manager once a global filter has been applied

Once the global filter has been applied, I loaded on the left side a photograph and on the right side the audio interview document, in which the participant described the content of that and all of the other photographs she/he took.  See below a screenshot showing this.

A Note on Coding of Photographs

I decided to code only the photographs that represented the core of a particular cultural scene.  I refer to those photographs as ‘primary photographs’.  I found that normally participants took several photographs describing a particular work activity, but only one or two of them were central and were described in more detail than the others.  For instance, a person could have a photograph showing a woman fetching water.  The person described that photograph and as she did so, she made reference to other photographs that played a complementary role and that helped explain the fetching water activity.  I refer to these as ‘secondary photogrpahs’. So, I decided to code only the primary photograph (or photographs), and to use hyperlinks to connect the quotations from that photograph to quotations from the corresponding secondary photographs.  This is illustrated by the simple figure below:

Central and secondary quotes

Figure 5. Primary photo quotations hyperlinked to secondary photo quotations

Coding Photographic Quotations and Hypelinking Them to Audio Quotations

Following, I selected each of the primary photographs and proceeded to code them following a reference conceptual framework that the research team had created and which was represented in the form of a code-to-code semantic network.  This network is shown in the figure below.

Conceptual framework

Figure 6. The guiding conceptual framework represented as a code-to-code semantic network

At the same time, I selected the section of the audio interview document in which the participant described that photograph (as well as the photographs that complemented that one), and proceeded to create a free quotation out of that segment.  I did not code the audio interview quotation because doing so would duplicate the coding of a cultural scene and would not contribute much to its understanding.  Instead, I hyperlinked that quotation with the quotation from the photographic document using the “Audio Explanation” relation, which I created.  The figure below shows this:

Hyperlined audio and photo quotations

Figure 7. Photo quotations hyperlinked to audio quotations

One of the interesting properties of hyperlinking the audio quotation with the photographic quotation is that clicking on the hyperlink next to the latter activates the former.  Thus, you can hear how the person describes and explains the cultural scene, without having to hear to more than that. 

Hyperlinking Primary Photographic Quotations to Complementary Photographic Quotations

I only coded the quotation from the primary photographic document. Then, I hyperlinked that quotation to those from secondary photographic documents that complemented the description of the cultural scene that is being represented by those images.  I used the hyperlinking relation “Complements”, which I created for this purpose.  (In ATLAS.ti, the system provides a number of standard relations to use in both the code-to-code and the quotation-to-quotation semantic networks.  However, the researcher can create new relations in order to best satisfy the modeling requirements).  The figure below shows this hyperlinking between primary and secondary graphic quotations.

Photo to photo hyperlinks

Figure 8. Primary photo quotation hyperlinked to secondary photo quotation

These hyperlinks can be visualized in a network view, which facilitates the understanding of the photographs, how participants describe them, and the relationships between the different photographs, audio descriptions and explanations, and concepts. One of these network views is shown below:

Hyperlinks in networks

Figure 9. Hyperlinks between primary and secondary photo quotations shown in a network view

Integration of Findings into the Theoretical Framework

The findings from the analysis of photographs are integrated into the whole of the analysis, which includes data collected through other methods, besides photovoice.  This integration takes place through code-to-code semantic networks and through memos.  Normally, the writing of memos and the representation of connections in semantic networks are intertwined activities in the sense that they do not take place in a linear order. However, in this case, I tended to first represent findings in semantic networks, and then to write memos motivated by what those networks were allowing me to visualize.

These networks integrate graphic quotations, the audio quotations in which participants describe and explain, the concepts to which these are related, the connection of these concepts with the guiding conceptual framework, and quotations from interviews and fieldnotes that help to further illustrate these concepts.  All of this provides an integrated view of a single knowledge domain, as well as of the relationship between that domain and the whole.  The figure below shows one of these networks.

Integrated network

Figure 10. Network view of a domain of analysis

On the other hand, memos allowed me to construct a narrative in which the pieces from the photovoice analysis were integrated with findings derived from other methods of data collection as well as from the review of the literature.  I wrote memos in a very dynamic manner, motivated in part by the network representations, the literature review, the analysis of co-occurrences (which included references to data collected from several methods of data collection, not only photovoice), the review of work frequencies across types of participants, and the reading of quotations by concept across types of participants (using the Query Tool).  The figure below shows an extract of one of these memos.

Memo

Figure 11. A memo

Conclusion

Using ATLAS.ti for the analysis of photovoice data was an experimental enterprise for me since I had never before worked with this method in ethnographic research.  Additionally, I had never analyzed photographs in such a systematic manner, linking photos with audio documents, and connecting this with the analysis of informal interviews, in-depth interviews, ethnographic videos, fieldnotes, as well as the review of the literature.  Precisely, I found that a major benefit of using ATLAS.ti in this analysis was the ability to integrate different methods of data collection, which, in my view, is one of the main sources of the richness of ethnographic methodology and fieldwork.

A final word of advice: although I did suggest in this short paper a set of steps to follow, I should emphasize that the process of analysis should be iterative rather than linear (although, of course, this is a methodological decision).  For instance, there is the potential for a rich dialogue between segmentation of photographs and interviews, coding, hyperlinking, representation of relationships in networks, and writing integrative memos.  However, this does not have to be, necessarily, a linear process. You may place each one of these activities in any sequence (e.g., start writing memos before creating a network and creating free quotations and hyperlinking them before coding). In sum, take the approach with which you feel the most comfortable and the one which will allow you to obtain the most from the photovoice data.

Reference Cited

Novak, David R. (2010). Democratizing Qualitative Research: Photovoice and the Study of Human Communication. Communication Methods and Measures, 4(4), 291-310.

Bibliography

Adams, Karen, Burns, Cate, Liebzeit, Anna, Ryschka, Jodie, Thorpe, Sharon, & Browne, Jennifer. (2012). Use of Participatory Research and Photo-Voice to Support Urban Aborigenal Healthy Eating. Health and Social Care in the Community, 20(5), 497-505.

Baker, Tamara A., & Wang, Caroline C. (2006). Photovoice: Use of a Participatory Action Research Method to Explore the Chronic Pain Experience in Older Adults. Qualitative Health Research, 16(1), 1405-1413.

Berbés-Blázquez, Marta. (2012). A Participatory Assessment of Ecosystem Services and Human Wellbeing in Rural Costa Rica Using Photo-Voice. Environmental Management, 49, 862-875.

Caricia, Catalani, & Minkler, Meredith. (2010). Photovoice: A Review of the Literature in Health and Public Health. Health Education Behavior, 37(3), 424-451.

Cooper, Cheryl M., & Yarborough, Susan P. (2010). Tell Me-Show Me: Using Combined Focus Group and Photovoice Methods to Gain Understanding of Health Issues in Rural Guatemala. Qualitative Health Research, 20(5), 644-653.

Fitzpatrick, Annette L., Steinman, Lesley E., Tu, Shing-Pin, Ly, Kiet A., Ton, Thanh, G. N., Yip, Mei-Po, & Sin, Mo-Kyung. (2012). Using Photovoice to Understand Cardiovascular Health Awareness in Asian Elders. Health Promotion Practice, 13(1), 48-54.

Garcia, Carolyn M., Aguilera-Guzmán, Rosa Maria, Lindgren, Sandi, Gutierrez, Rodolfo, Raniolo, Blanca, Genis, Therese, . . . Clausen, Lisa. (2013). Intergenerational Photovoice Projects: Optimizing this Mechanism for Influencing Health Promotion Policies and Strengthening Relationships. Health Promotion Practice, 14(5), 695-705.

Harper, Krista. (2009). Using Photovoice to Investigate Environment and Health in a Hungarian Roman (GYPSY) Community. Practicing Anthropology, 31(4), 10-14.

Harper, Krista. (2012). Visual Interventions and the “Crises of Representation” in Environmental Anthropology: Researching Environmental Justice in a Hungarian Romani Neighborhood. Human Organization, 71(3), 292-305.

Johnson, Ginger A. (2011). A Child’s Right to Participation: Photovoice as Methodology for Documenting Experiences of Children Living in Kenyan Orphanages. Visual Anthropology Review, 47(2), 141-161.

Lai, Shalini. (2012). A Scoping Review of the Photovoice Method: Implications for Occupational Therapy Research. Canadian Journal of Occupational Therapy, 79(3), 181-190.

Mejia, Angie Pamela, Quiroz, Olivia, Morales, Yolanda, Pone, Ruth, Limon Chavez, Graciela, & Olivera y Torre, Elizabeth. (2013). From Madres to Mujeristas: Latinas Making Change with Photovoice. Action Research, 11(4), 301-321.

Mitchell, Claudia, DeLange, Naydene, Moletsane, Relebohile, Staurt, Jean, & Buthelezi, Thabisile. (2008). Giving a face to HIV and AIDS: On the Uses of Photo-Voice by Teachers and Community Health Workers Working with Youth in Rural South Africa. Qualitative Research in Psychology, 2(3), 257-270.

Novak, David R. (2010). Democratizing Qualitative Research: Photovoice and the Study of Human Communication. Communication Methods and Measures, 4(4), 291-310.

Strack, Robert W., Lovelace, Kay A., Jordan, Toshia Davis, & Holmes, Anita P. (2010). Framing Photovoice Using A Social-Ecological Logic Model as a Guide. Health Promotion Practice, 11(5), 629-636.

Wang, Caroline. (2006). Youth Participation in Photovoice as a Strategy for Community Change. Journal of Community Practice, 14(2), 147-161.

Wang, Caroline, & Burris, Mary Ann. (1997). Photovoice: Concept, Methodology, and Use for Participatory Needs Assessment. Health Education Behavior, 24(3), 367-387.

Wang, Caroline, & Pies, Cheri A. (2008). Using Photovoice for Participatory Assessment and Issue Selection. In Meredith Minkler & Nina Wallerstein (Eds.), Community-based Participatory Research for Health (pp. 183-197). San Francisco, CA: John Wiley & Sons, Publishers.

Wiersma, Elaine C. (2011). Using Photovoice with People with Early-Stage Alzheimer’s Disease: A Discussion of Methodology. Dementia, 10(2), 203-216.

Resources on Photovoice

About the Author

Ricardo B. Contreras, PhD

Ricardo B. Contreras, PhD, is an applied cultural anthropologist with a research agenda situated at the intersection of migration, health and community.  His undergraduate degree is from the Universidad de Chile and his graduate degrees from the University of South Florida.  He is the president of Ethnographica Sociocultural Research and has been an instructor of ATLAS.ti since 2005.  He resides in Corvallis, Oregon.