Abstract
Photovoice offers creative participatory action methods for conveying community strengths and challenges with the goal of addressing health inequities. Accelerated by COVID-19 restrictions, photovoice has increasingly become virtual, and this shift has given rise to new considerations including navigating online recruitment and data collection, e-participatory action trends and working with multi-site large qualitative data sets. Within these contexts, the current article discusses the case for and against virtual photovoice, drawing from a large study comprising 110 men’s experiences of, and perspectives about, equitable and sustainable intimate partner relationships. The findings are shared across three themes. The first theme, e-Efficiencies and concessions contrasts increased recruitment reach and data collection cost-savings with vulnerabilities to phishing and challenges for working with participants’ wide-ranging internet literacies and practices. Theme two, Participatory action changed, chronicles the participants’ varied relationships to photography including sourcing third-party and archived photographs. Revealed also were privacy concerns whereby some participants opted for audio only interviews and/or restricted the use of their photographs. The third theme, Reckoning breadth and depth in a large dataset, discusses emergent study design considerations including analytics for interpreting and contextually representing large multi-site projects that are made possible through virtual photovoice. While technological advances and COVID-19 have forged photovoice virtually, the case for and against this trend reveals complex considerations that will likely manifest a continuum of approaches ranging virtual, hybrid and in-person models. In summary, we suggest that integral to weighing the case for and against virtual photovoice researchers will need to thoughtfully adapt to changing technologies, as well as potential post COVID-19 tilts for returning to in-person.
Keywords
Introduction
The early photovoice work from Wang and Burris (1997) yielded significant advances and unprecedented legacy effects for visual methods in qualitative health research. Beginning with projects that powerfully detailed women living in marginalizing conditions, participants’ socioeconomic challenges were highlighted in purposefully lobbying policy change to address health inequities (Wang et al., 1998). This early photovoice work also modeled and seeded participatory action methods characterized by the researcher[s] being physically present and collaboratively interacting with the participants and environments that they studied (Wang, 2003). That said, diversity soon grew for how and where photovoice health research was done (Catalani & Minkler, 2010). For example, a range of qualitative methodologies were integrated including narrative analyses (Mizock et al., 2014), phenomenology (Tanhan & Strack, 2020) and grounded theory (Hatala et al., 2020), while the rise of digital technologies influenced photovoice processes and products, and the dissemination of those study findings (Oliffe and Bottorff, 2022). The COVID-19 restrictions, especially physical distancing, also pressed many photovoice traditions, making obligatory virtual study designs. The current article offers methodological insights regarding the case for and against virtual photovoice based on our experiences of completing a large photovoice study focused on men’s experiences of, and perspectives about, equitable and sustainable intimate partner relationships.
Photovoice Then, now and Going Forward
Wang and Burris (1997) defined and detailed photovoice to suggest that it was driven by three primary goals. First and foremost were the participatory action endeavours in which photovoice purposefully provided opportunities for participants to chronicle their communities’ strengths and challenges. The social determinants of health were central to this goal whereby participant produced photographs and narratives visually illuminated the connections between agency and structures to raise awareness of inequities. Second, participant photographs and narratives were collectively shared to facilitate group conversations, and distil collective views regarding community health challenges and strengths. The third goal was to target policymakers with participants’ consensus views as a means to lobby change. In pursuit of the aforementioned goals, photovoice traditionally relied on researchers completing in-person fieldwork and interviews, and physically being there in the participatory action and co-construction of the data, analyses and targeted dissemination of the study findings (Wang, 1999). Photography was also focal, and acquiring photography skills was idealized early on as a rich by-product of photovoice participation (Wang and Burris, 1997). While some argue that “photography died the moment a camera lens found it’s way onto a cell phone” others distinguish between recreation, hobby and professional photography (Quora, n.d.). Notwithstanding such debates and differentiations, it was ever clear that the normative use of smartphones for taking, editing, storing and sharing pictures forever changed photography – and herein, photovoice forwent some of its early gains (and ideals) for participants building traditional photography skills. Inversely, the predominance for point-and-shoot digital photographs and surge of the selfie, has undoubtedly increased the accessibility, inclusivity and biopic reach of photovoice.
In the 25 years since Wang and Burris’ (1997) seminal work, a plethora of cross-sectional photovoice studies have illuminated wide-ranging health issues including grief (Creighton et al., 2013), social connectedness (Bonell et al., 2022), suicide bereavement (Oliffe et al., 2020), fathering (Creighton et al., 2014), smoking (Oliffe et al., 2010), physical exercise (Thandi et al., 2018), prostate cancer (Oliffe, 2009) and depression and suicidality (Ferlatte et al., 2019). These diverse health-related photovoice projects have shared thick descriptions of specific sub-groups and phenomena predominately within small sample-size and single-site studies (Jarldorn, 2019). Though some photovoice studies had moved online prior to the pandemic (Tanhan and Strack, 2020), COVID-19 restrictions hastened moves for photovoice research to be done virtually. Specific to COVID-19, Rania et al. (2021) and Ferlatte et al. (2022) detailed virtual photovoice challenges for managing online meeting technologies, and establishing trust and connectedness between researchers and participants. Similarly, Chen (In Press) reported challenges for converting potential to actual participants in virtual recruitment wherein the lack of responsiveness to email correspondence was a contributing factor. Acknowledged also was the time-consuming nature of some technologies and e-communications, for which participants’ uneven skillsets were implicated (Ferlatte et al., 2022; Rania et al., 2021). Relatedly, COVID-19 normed the use of virtual qualitative interviews with concessions including the need for researchers to adapt their preparation and pacing, and accommodate unstable internet connections contrasting data collection labour cost-saving benefits (Oliffe et al., 2021). The current article builds on these beginning insights (Chen, In Press; Ferlatte et al., 2022; Rania et al., 2021; Tanhan & Strack, 2020) by offering project-informed acumens to discuss the case for and against virtual photovoice.
Locating the Current Photovoice Study
The interconnections between men’s mental health and their intimate partner relationships informed the current photovoice study’s focus on better understanding what constitutes equitable and sustainable partnerships (Oliffe et al., 2022). Specifically, distressed and disrupted intimate partner relationships significantly increase men’s mental illness and suicide risk (Grundström et al., 2021) while married men report health benefits, and have a higher life expectancy in comparison to single men (Jia and Lubetkin, 2020). Taking a strengths-based approach we were interested to understand men’s experiences of, and perspectives about, equitable and sustainable intimate partner relationships, as a means to informing tailored programs for men to build better partnerships. The study focused on men 19 to 44 years-old who spoke and understood English, and a $100 CAD honorarium was offered to participants. Within this criterion we maximized diversity by recruiting participants worldwide from any ancestry, locale and sexual identity. The study was purposely designed to be 100% online, and one primary objective was to test the feasibility and acceptability of conducting photovoice research virtually.
Recruitment was completed via social media platforms Twitter, Facebook, LinkedIn and Instagram. Thirty organizations and individuals that work with men or focus on men’s health around the world were also messaged via email and provided project recruitment materials for their social media posts and newsletters. Five hundred and eighty-nine potential participants contacted the project manager via email, and one of six researchers (2 women and 4 men based in Canada) followed-up with 239 potential participants to schedule a brief (approximately 10 minutes) eligibility/intake meeting via Zoom. Of these 239 contacts, 60 (25%) did not respond to our email reply, and 179 (75%) confirmed their interest in the study by scheduling an eligibility/intake meeting. The purpose of the eligibility/intake meeting was to explain the study and detail participation including the completion of Qualtrics hosted digital consent[s], demographics and survey questionnaires, and uploading of 5 to 10 photographs to illustrate their experiences of, and perspectives about, equitable and sustainable intimate partner relationships. Eligibility/intake meetings were completed with 159 men (20 men did not attend the scheduled meeting), and the study information and hyperlinked Qualtrics hosted enrolment forms were subsequently sent to these potential participants by email. Of those 159 potential participants, 48 (30%) scheduled an individual photovoice Zoom interview at intake while 111 (70%) indicated that they would prefer to submit their photographs before scheduling an interview. Of the 159 potential participants who completed the eligibility/intake meeting, 49 (31%) men did not complete the consent or submit photographs to formally enroll in the study. In summary, follow-up with 239 email contacts who had expressed interest in the study resulted in an overall conversion rate to participation of 46%, and we completed individual Zoom interviews with 110 men July through December, 2022 (see Figure 1: Recruitment Flow Diagram). Recruitment flow diagram.
Participant Demographics (n = 110).
The insights shared in the current article were derived through post-interview researcher debriefs and weekly team meetings where the study procedures, data collection processes and data quality were discussed and documented as reflexive memos to guide our process evaluation of conducting virtual photovoice. The six interviewers focused their discussions on: 1) operational facets of recruitment and the interviews, 2) the e-participatory aspects of the study and, 3) approaches to managing, analyzing and representing the data collected. To decide and build content for each of these discussion points, notes were labelled and elaborated on. As discrete discernments were derived and ordered, some notes were relabeled, reassigned and/or subsumed. These analytic processes continued in researcher meetings and the writing-up of the current article whereby content categories were progressed to themes with descriptive labels advanced to pre-empt three specific (and intricately connected) findings: 1) e-Efficiencies and concessions, 2) Participatory action changed, and 3) Reckoning depth and breadth in large datasets. Some illustrative quotes accompanied by researcher chosen pseudonyms are provided to contextualize key points.
Findings
e-Efficiencies and Concessions
Technologies are much debated, especially in terms of their reach and usability (Carayon & Hoonakker, 2019). In doing photovoice research virtually there were some salient learnings pertaining to e-Efficiencies and concessions, chiefly related to the operational aspects of participant recruitment and enrollment, and conducting individual Zoom interviews. Extending insights shared by Chen, (In Press); Ferlatte et al., (2022); Rania et al., (2021), the recruitment and enrolment e-efficiencies were most evident in the worldwide reach of our social media campaigns, with diverse and large numbers of men making contact via email to express their interest in participating. Virtual snowball sampling was made possible through community champions, enabling us to include men from 15 countries with wide-ranging ancestries and sexual identities. For example, amid thanking us for the opportunity to be part of the research during the eligibility/intake meeting, Tyre, a 22-year-old bisexual man from the US, offered to promote the study through his sexual minority men’s social media chat group. Affirming “we don’t talk enough about gender and relationships,” Tyre’s endorsement of the study prompted 5 of his network peers to contact us about taking part in the research. Rhetoric regarding men’s reticence to take part in qualitative research, particularly those experiencing marginalizing conditions, was disrupted through the many and varied virtual communities lobbied by participants who had taken part in the study.
Because the virtual reach garnered interest far beyond the 50 participants proposed as the study sample, we were able to purposefully recruit men as data collection progressed. Specifically, we included men from as many countries and cultures as possible, and the feasibility for including more participants was aided by data collection cost-savings. For example, completing Zoom interviews with 14 men who lived in 9 states and 13 cities across the US cost $1,400 CAD (based on researcher salaries) compared to $30,000 CAD (researcher salaries plus travel, accommodation and per diems) if those interviews had been conducted in-person.
One unresolved recruitment and enrollment e-concession was that by relying on social media, potential participants who did not have internet access, as well as those who did not subscribe to the specific platforms that we used (i.e., Twitter, Facebook, LinkedIn and Instagram), were excluded. Social media recruitment also has labour (and budget) implications, and the proliferation of new platforms has amplified challenges for strategically deciding which one[s] to use. We primarily recruited via Twitter with an established account (>3000 followers) based on our understandings that men 19–44 years-old are amongst the highest Twitter consumer groups (Wojcik and Hughes, 2019). Recognizing the fit of photovoice with Instagram, a new account was created to aid recruitment; however, minimal engagement with the recruitment posts yielded no return on that investment. Hence, a concession for e-recruitment was that while Twitter was effective in securing 83 (75%) participants, many men, some of whom may have vested interests in photography and visuals (i.e., Instagram) were not directly reached.
Additional e-concessions for relying on social media recruitment included contact from bots and phishing emails. Our vulnerabilities in this regard were likely increased by the offer of a $100 CAD e-gift card honorarium for study participants. Specifically, we received an unprecedented number of contacts in the 5th week (n = 140; 24% including 80 on August 10, 2022 alone) after two study recruitment tweets were retweeted by accounts with a large number of followers. We suspected 280 (48%) of the 589 contacts were bot initiated phishing emails, detectable by the lone phrase, “I’m interested” in the body of the email, with an empty subject line. We tested some of these emails asking open-ended questions, and immediately received automated bot replies with content irrelevant to the eligibility/intake scheduling information solicited. Needing to vet and decide legitimate contacts amid consulting university information technology services regarding security concerns stalled recruitment, and we experienced high administrative workloads and labour costs – significant e-concessions for study recruitment.
Participants’ varied detail-orientation for completing the Qualtrics hosted consent[s], demographic survey and photograph uploads prior to the scheduled interview also added to the administrative workload. In terms of benefits, Qualtrics is a free base utility at our university, its security is approved by the institution’s ethics board, and the platform requires modest bandwidth for participants to input their data. Overall, most participants reported being able to easily access, navigate and upload their photographs to Qualtrics. We checked the Qualtrics data 24-hours prior to the scheduled interview, and 55 (50%) men required email prompts to complete this information, of which 25 (45%) participant interviews needed to be re-scheduled. While the convenience of e-communications is much lauded, efficiencies are oftentimes less evident (Chen, In Press), and the high volume of emails (i.e., an average of 7 for each of the 239 contacts) had labour implications with incomplete enrollments, non-responses, no-shows and re-scheduling proving especially disruptive to the project workflow.
Data collection e-efficiencies included the sequencing and screen-sharing for jointly looking at, and richly discussing participants’ photographs one-by-one. Interviews began with introductory questions inviting the men to share information about their study/work, ancestries and the relationship[s] contexts they would be discussing. By the 15-min mark of the interview we transitioned to viewing the photographs, as men shared their thoughts and narratives for each image. The participant and researcher videos were sidelined and screen-sharing privileged the men’s photographs, focusing the discussion on their images. The centrality of the photographs was powerful, and participants consistently referenced and referred back to specific images in the interview. Equity, as the central study construct, was explicitly asked about. For example, an interviewer asked 41-year-old Canadian participant Agustin, “So what do you think would be equitable with the dishwashing?” prompting him to elaborate on Photograph 1 titled “Chores”: “The thing that we’ve gone through is like, she’s [partner] like, ‘okay I don’t like the way you do them’[dishes], it’s like, ‘okay, well, teach me how to do it, and then I will try my best to do it’, and if it doesn’t work, then you’ve got to do them. But ‘please do them at a time that is, like, sooner rather than later’… like this picture specifically, like gives me anxiety… this is something that provokes a certain emotion for me and so I’m just like, okay, this is something I want to document to say like, ‘I personally don’t like this,’ however, I know that there's a compromise that needs to be made and I think that part is important, especially with chores.” Chores.
Here, and consistently, we noted men’s ease for building rich storylines around their photographs, and this may reflect norms for the live e-sharing and talking to digital images in the everyday. Most participants (n = 100; 91%) interviewed from their home, and their comfortability for talking was also likely aided by being in the locale where many of their photographs and narratives were situated (i.e., the domestic sphere).
The e-concessions regarding data collection included the Zoom auto-transcription option which proved expensive to the extent that we outsourced the last 90 audio interviews to professional transcription services. Specifically, corrections to the first 20 auto-transcribed interviews indicated that the research assistant labour costs were 50% higher than professional transcription. Challenges mostly related to the notepad formatting of the Zoom auto-transcripts, and inaccuracies for capturing researcher and participant accents and linguistic conventions. Some participants also (n = 14; 13%) refused to use the Zoom video option wherein 8 men (7%) indicated they were conserving their phone data, while 6 (5%) cited safety and privacy concerns relating to their sexual identities and/or cultural ancestries. The e-concession here was for us to adapt to audio only interviews, barring visual connections with and/or observations of participants’ affect and locale.
Participatory Action Changed
In photovoice, the participatory action processes and products for making health challenges visible can, in and of themselves, count as interventions. For example, participants’ gains through photography and their narrations of those images can bolster agency, and powerfully draw attention to health inequities (Oliffe and Bottorff, 2022). While photovoice has traditionally relied on in-person fieldwork and interview methods the virtual nature of the current study rendered Participatory action changed.
With regard to potential losses, as Chen (In Press) reports, virtual photovoice physically distanced researchers and participants, changing the tangibility and exchange for what could be seen and learnt through interacting within a shared environment. Foregone in the current study were opportunities for observing and getting to know participants within their material environments, and fostering traditional participatory action researcher-participant connections. For example, Wang et al.’s (2000) study of men and women living in a shelter in Ann Arbor, Michigan, USA, included photography workshops and group walk-throughs ahead of participants taking photographs to illustrate their health, work and life-related struggles and strengths. These processes drew participation through prolonged engagement and content pacing for researchers to build rapport and trust with participants. By contrast, the brevity of our virtual interactions with participants meant that our focus, by necessity, was on communicating specific instructions and lobbying men’s insights through a one-time Zoom interview.
Though men routinely asked for (and were provided) examples to guide their photographs and conceptualize the photovoice purpose and task, most participants expedited that mission by using third-party web-sourced images and/or submitting their own pre-existing archived (most often smartphone taken and stored) photographs. While participatory action gains for building photography skills through photovoice were eroded prior to COVID-19 by the normative (if not obligatory) use of smartphone cameras, participants’ sourcing of third party and/or their own e-stored photographs underscored their collective estrangement from traditional photography. For example, 8 (7%) participants sourced images from the internet searching key words including communication, conflict and chores. Reasons offered for submitting web-sourced images included disinterest in taking photographs, inability to capture the essence of their story in ways that internet images could, and/or a lack of time for planning and taking one’s own photographs. Sean, a 33-year-old man from the USA, explained that browsing and selecting the genre and specific open-source web images (which he subsequently downloaded and then uploaded to Qualtrics) sufficed for him to reflect on, and talk about, his past relationships in the interview: “All the photos that I found they have that thing where they're like, available to the public… doing that picture project, it was interesting, right. Like, I found myself wanting to reach for photos that were almost like those calendar photos.”
When directly asked about his decision not to take or use his own photographs, and his relationship to images and photography more broadly, Sean explained: “It’s not that we’re getting more impatient [as a generation], it’s that we’re just becoming more effective. So, the truth is, like the idea of going out and taking pictures and stuff, it sounds kind of like a fun project. But when you factor in the time and like the planning and all that stuff, it is so much easier just to go on Google and find the perfect picture. Because when I look at this [referring to a picture on shared screen], it’s like, you know, this is going to describe exactly how I feel about a certain thing.”
Approximately 88 (80%) participants used their own pre-existing stored photographs, similarly highlighting the reflexivity for contextualizing and connecting their images to the study purpose and interview. As Walter a 24-year-old Canadian man said: “I was looking through my old photos, I have over 9000 of them. It kind of brought back memories that I hadn’t thought of in a long time, like dates that I’ve been on, experiences that I’ve had, and it was a bit of a soul searching kind of thing. It made me look a bit more into the past. I thought it was very introspective.”
It is telling that only 14 (13%) men took new photographs (all via their smartphones) specifically for the study. Regarding photograph titles and/or captions 77 (70%) participants completed that in Qualtrics as requested prior to the interview. While men’s distance from traditional photography was evident, their reflexive processes were apparent in the detailed accounts that they provided for their chosen images. Evident here was the shifting nature of participatory action – whereby the men diversely decided their photograph processes and products. Such actions can be argued as embodying photovoice’s contemporary emancipation and participatory action ideals.
While the targeted sharing of participant produced photographs with decision makers and/or the public has also been synonymous with photovoice, some men in the current study did not consent to their photographs being used (or seen) outside of the interview. Specifically, 24 (22%) participants redacted some or all of their photographs based on concerns about their digital footprint, privacy and cybersecurity risks. For example, Justin, a 31-year-old Canadian man, foregrounded that he had purposefully reduced his web presence and visibility when explaining why his photographs could not be shared publicly: “I remember when I got Facebook in 2005 you’d upload like 30, 40 photos to an album. And you’d just tell everybody, and it would be all there… I think as I developed an awareness of surveillance by government, states… even just surveillance by the companies who I am uploading photos to, I got uncomfortable with that and… I do a lot of research, and a lot about history, and it is scary how much stuff is out there on people, and the surveillance out there. And so, I have … been really reducing my internet footprint, removing most of what you can find out about me from the internet. Partially because of getting involved with activism as well and knowing that getting doxed is a real thing and that would endanger myself and family… I just don’t want to open myself up to that possibility, and so having that security is important to me.”
Adding to Creighton et al.’s (2017) discussion about digital photovoice ethics considerations, the current study disrupted taken-for-granted unitary emancipatory gains for participants sharing their experiences and perspectives through photographs.
By contrast, there were instances where virtual photovoice being conducted worldwide was inclusive of under-researched and hard-to-reach groups. For example, we included 20 men from Turkey (with an additional 25 waitlisted), responding to their strong interest in the study. Listing multiple benefits, Cahill, a 22-year-old man from Istanbul, closed our interview by expressing his gratitude for having a voice in a study that was focused on men’s relationships, and conducted worldwide by researchers from a well-known Canadian university: “I was excited about speaking English – because I have been speaking English for a while – and sharing my opinions and voicing my inner thoughts via this experience. The interview with you has just been a blessing. I want to thank you for the joyful moments that I was able to experience with you.”
In addition to honing his English language skills, Cahill, and many men for whom English was a second or third language, valued their inclusion and opportunities to share culturally rich photographs depicting their experiences and perspectives.
Reckoning Breadth and Depth in a Large Dataset
The proposed 50-participant study sample was based on previous experiences for achieving data saturation, as well as the 24-month project timeline and the funding envelope ($237,000 CAD). The determination to interview 110 men was influenced by participant diversity (e.g., locale, ancestry, sexual identity) yielding wide-ranging experiences and perspectives to the extent that our expectations for achieving data saturation with 50 participants were completely derailed. Herein, Reckoning breadth and depth in a large dataset was an emergent (and somewhat unforeseen) challenge – a consequence of our decision to more than double the number of participants that we had originally anticipated interviewing for the study.
In regard to data saturation, our previous photovoice projects were typical of the field in that they were conducted in-person at single site locales with small samples, and these study designs expedited data analyses and sureties for representing all the participants in the findings. Conducting participant interviews worldwide troubled those logics whereby the men’s diverse backgrounds and agency-structure interactions complicated our usual analytics, including (and perhaps especially) constant comparative approaches, for inductively deriving patterns across the sample. For example, Fisher, a 32-year-old father of two who lived in Hong Kong, weaved locale-specific economic, cultural and class structures when explaining that he had jointly decided with his wife to employ a full-time nanny: “I guess we don’t have other practical options, to be honest… because, we needed two incomes and we needed someone to take care of the kids. My father could help with it, but he is too old to take full responsibility for that… the other factor would be the lack of public childcare services in Hong Kong. We are able to hire a domestic helper and that's quite common in middle-class families. So, our domestic helper is from the Philippines, she lives in our house and she mainly takes care of the kids and the housework.”
Positioning both he and his wife as equally needing to work financially, and for career advancement, their aligned agency for employing domestic help was normed within their specific locale and social contexts. While rationalized as reasonable in responding to the social structures of Hong Kong in which they lived, there might be Western feminist critiques about gender inequities for hiring low-paid female nannies from low-income countries (Barker, 2005; Tronto, 2002). Indeed, the fit of many theories and constructs routinely used to conceptually advance photovoice study findings were drawn into question by the diversity in the sample. Gender equality, equity and reproductive rights for instance - well established and widely accepted constructs in Canada (Government of Canada, 2014) - may not have the same resonance or purchase when interpreting the intimate partner relationship views of men who operate in (and to sustain) cultural systems that (still) limit women’s freedoms and opportunities. For example, Wayne, a 26-year-old Canadian based participant of African ancestry, explained his partner (also of African ancestry) understood and accepted that men do not perform “womanly duties” in the home (i.e., cleaning, dish washing, cooking). Our analyses herewith worked to unpack how gender roles are co-constructed and negotiated by men in intimate partner relationships and map diverse understandings of equity within and across an international sample. Key here was that participants’ experiences of, and perspectives about, equitable intimate partner relationships could reflect distinctly different cultural traditions and gender relations that were normative in their specific locale or community. Indeed, we were hearing diverse agency-structure accounts well beyond the first 50 participants. It was not until we had interviewed more than 80 participants that we had some certainty about the analytics for describing patterns across, and contextualizing categories and variations within, the large study sample.
Crucial to recognizing diverse structures and agencies, and contextually mapping participants’ interactions were our reflexive practices for confronting and building on our own cultural beliefs and values. This included dispensing with ideals about the resonance of Canadian policies and social norms with respect to interpreting participants accounts. Such reflexivity was imperative to conducting a photovoice study that was responsive to equity, diversity and inclusion (EDI) frameworks. Though many in-person photovoice projects include a range of participant socioeconomics and cultures, the current study participants resided in 15 distinctly different countries and this demanded focused attention to contextualizing how that plurality of structures influenced participants’ experiences and perspectives.
Data analyses were further complicated by the sheer volume of 110 interviews comprising more than 2,000 pages of text and 774 photographs. Our previous photovoice studies were characterized by team meetings to discuss 20 to 30 participant interviews, and there was familiarity and detailed recall aided by the researchers having direct knowledge of every participant. The large, diverse data set in the current study forced us to work differently, and we employed a ‘nested’ analytic approach, assigning participant sub-groups to derive and test typologies. For example, in the first empirical article three interviewers reviewed and discussed in-depth 28 participant interviews, inductively deriving, defining and differentiating three typologies for categorizing men’s intimate partner relationship equity. Then, the other three interviewers independently reviewed the 82 remaining interviews to assess the fit, and their classifications were checked and discussed to reach agreement for refining and weighting the typologies.
While the breadth of the data was managed with the aforementioned strategies, we also garnered depth by completing discrete sub-group analyses (e.g., sexual minority men (n = 16), Turkey-based participants (n = 20), 19–24-year-olds (n = 39)) purposefully harvesting from the larger sample. With large photovoice datasets, in-depth sub-group comparisons were also possible. For example, the equity views of 20 participants from Turkey could be compared to a sub-sample of 20 men from Western Canada. This approach can illuminate depth through differentiating locale specific interactions (i.e., parental leave for fathers, economic barriers) as a means to benchmarking a range of relevant health policies. We can also employ mixed method analyses integrating survey questionnaires regarding men’s relationship views (i.e., conflict resolution, masculinity, and gender equity) with qualitative data to address, with parallel study designs, the role of specific constructs. For example, sample demographics (e.g., age) can be regressed to report associations with specific survey constructs levering triangulation opportunities for selective coding of relevant qualitative data to elaborate and augment the quantitative findings. While this approach is commonplace in mixed methods research, it has not been a feature of photovoice studies.
Lastly, there have been suggestions that the images submitted by participants in photovoice studies deserve more focused analyses (Oliffe et al., 2008). To manage the large collection of 774 photographs we initially allocated the visual data and accompanying narratives to 23 descriptive codes ahead of subsuming them to 5 broad category sub-groups: 1) communication, 2) domestic work, 3) finances, 4) intimacy, and 5) values. The process and analytics for doing this increased our familiarity with the photographs whereby we began developing interpretations of that data to inductively derive research questions (many of which resided outside our primary research questions). For example, we undertook a formal analysis of the 50 photographs assigned to intimacy to map connections with equity in men’s intimate partner relationships. Large datasets can also offer opportunities for collaboration with colleagues, and secondary analyses (see Thorne, 2013) for graduate students and postdoctoral fellows who have emergent or established interests in visual qualitative research.
Our experiences indicate that there are affordable opportunities and benefits for completing large virtual photovoice studies. That said, Reckoning breadth and depth in a large dataset was an emergent consideration for us, and fully utilizing the data collected carries budget implications in the time and paid-labour needed for analyzing and writing up all the potential findings.
Discussion and Conclusion
Advancing our recent experiences to insights for doing photovoice virtually affords important and timely methodological considerations to guide the field. Temporally, the COVID-19 long play, inclusive of varied transition[s] out of the pandemic, will prolong and perhaps norm virtual photovoice. That said, as restrictions ease and confidence grows, hybrid (i.e., virtual and in-person) photovoice models may emerge, amid some researchers wholly returning to Wang and Burris’ (1997) in-person work. It is across this continuum of possibilities that we return to the current article title - The case for and against doing virtual photovoice – to build discussion about each of the three thematic findings.
Regarding e-Efficiencies and concessions, extending recruitment reach virtually through social media, though effective, was tempered by challenges for strategically deciding and effectively using specific platforms. Case in point being the differences in recruitment returns for our established Twitter and newly opened Instagram accounts. Highlighted here was the need for budget and expertise to lever content, valence and volume capable of tapping channel-specific social media algorithms for recruitment (Peters et al., 2013). Bot-initiated phishing emails, as an ever-growing and significant social media vulnerability also underscore the need to monitor and adjust these e-recruitment efforts giving particular attention to cybersecurity (Alkhalil et al., 2021). These issues and many e-efficiencies and concessions entwine as cause-effect states, confirming the need for multiple recruitment chains including tailored lobbying (in-person and electronic) rather than solely relying on the diverse but somewhat scattergun reaches of social media.
Conducting photovoice interviews virtually provided efficiencies and cost-effective avenues for reaching geographically diverse participant cultures and values (Keen et al., 2022). The normative use of Zoom worldwide – a COVID-19 legacy effect – was reflected in the familiarity and usability amongst participants. Contrasting those e-efficiencies were well-documented concessions including fatigue (Rania et al., 2021), unpredictable and often-times disruptive environments (e.g., noise, low internet connectivity, cross-talk) (Ferlatte et al., 2022) along with emergent ethical considerations including connecting participants who experience emotional distress with locale-specific tailored supports (Creighton et al., 2017). Though we did not observe participant distress directly in the current study, we suggest preparatory work for managing participants’ safety risks in virtual photovoice studies is a key consideration going forward (see Keen et al., 2022; Roberts et al., 2021). In addition, participant e-privacy and confidentiality concerns underscore the need for responsive university information technology security services to assist researchers to ensure the integrity of the photovoice data that is collected, stored and disseminated virtually through those hosting institutions (Canadian Centre for Cyber Security, 2022).
Specific to Participatory action changed, the early photovoice work of Wang and Burris (1997) trailblazed pre-millennium with in-person studies that purposefully engaged and sought to collectively empower locale-specific participants. That said, the participatory action methods driving that emancipatory photovoice work were somewhat prescriptive, albeit with venerable intent including (but not limited to) participants building photography skills. A key change, and salient learning from the current study were participants’ agency for defining their participatory actions, especially in deciding photograph processes, products and disseminations. We suggest that these shifts reflect and respond to participants’ sensibilities and drive for autonomy within increasingly diverse but connected virtual cultures. This, of course, includes participants’ preferences for which images can and cannot be shared publicly. Herein, there are prompts for researchers to [re]consider photovoice’s participatory action and emancipatory processes through EDI frameworks. In essence, participants’ agency for defining their photovoice involvement has implications for collaborative flexibility in the design of future studies. Specifically, EDI frames can (and should) engage participants on their own terms – especially underrepresented individuals and sub-groups who are routinely excluded (or aggregated and potentially othered) by racism, xenophobia, and/or ableism (Gilbert et al., 2016). While this alone is inadequate for promoting health equity, photovoice done virtually can forge EDI to advance Wang and Burris’ (1997) emancipatory visions by ensuring participants have control of their processes and storylines. Though photovoice is a reasonably well-known methodology there may be benefits to distinguishing virtual photovoice as decidedly EDI based in its social media age (Sujon, 2022) and post-truth era (Sismondo, 2017) by freestyling participant driven participatory action methods (Evans-Agnew et al., 2022).
It is important to situate the findings pertaining to Reckoning breadth and depth in a large dataset as contextual – wherein these results flowed to and from our decision to increase the sample size while the research was in progress. The caveat here is that collecting more data was, and always should be driven by analytics for rigorously completing qualitative data analyses (e.g., constant comparison and data saturation) (Fusch and Ness, 2015). That said, the feasibility for more than doubling the international sample was only possible by doing the photovoice study virtually. Two important considerations emanate here. First, virtual photovoice studies do not have to be large and multi-site. Indeed, there can (still) be small, single-site virtual photovoice studies dedicated to thickly describing local contexts and lobbying specific policy change. Second, irrespective of the study sample size the EDI frames guiding design and recruitment in virtual photovoice must also be visible in the study findings. To this end, data analysis plans should purposefully ensure context-rich representations reflecting the array of participants who took part in the research, as well as distilling discrete patterns within and across the sample (Dewidar et al., 2022).
In terms of limitations, we acknowledge the cross-sectional study design as shorting longitudinal insights to photovoice changes across (and after) COVID-19 - a period marked by great uncertainty and rapid adaptions to virtual environments. The study topic – men’s experience of, and perspectives about, equitable and sustainable intimate partner relationships – might also be understood as influencing the high interest for participating as a by-product of the pandemic and/or #MeToo era. It is also important to acknowledge the inclusion of some approximates in quantifying the metrics shared in the current study. These shortfalls can be addressed through longitudinal and mixed gender samples inclusive of study specific (and comparative) process evaluations for doing virtual photovoice.
In conclusion, the ubiquity of smartphone cameras, digital photographs and e-images has forever changed traditional photography and expectations about photography, amid COVID-19 accelerating people’s diverse (and normative) virtual connectedness and multiple-residencies. Taken together, it can be argued that photovoice was (and is) well-suited and nimbly adaptive to expand its virtual presence. While photovoice has moved with these (and many other) changes - the lost physical presence and materiality that characterized the early work is undoubtedly grieved (albeit to varying degrees). It may also well be that the much-touted post-pandemic ‘return to normal’ inevitably realises a new chapter for photovoice – a methodology forever changed and changing in the digital era. Irrespective of that continuum, the case for and against doing virtual photovoice will continue to be influenced by changing technologies, as well as people’s post COVID-19 tilts for being there in-person.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was made possible with funding from New Frontiers in Research Fund (Grant # F21-04676). John L. Oliffe is supported by a Tier 1 Canada Research Chair in Men’s Health Promotion. Simon M. Rice is supported by a Dame Kate Campbell Fellowship from the Faculty of Medicine, Dentistry and Health Sciences at The University of Melbourne. Zac E. Seidler is supported by an Investigator Grant Fellowship from the National Health and Medical Research Council of Australia (2008170). Calvin C. Fernandez is supported by PhD funding via John L. Oliffe’s Tier 1 Canada Research Chair in Men’s Health Promotion. Office space provided to John L. Oliffe by the Wyatt Trust, Adelaide, South Australia.
