The Ethics of Representing Girls in Digital Policy Spaces

In this article I apply policy frame and visual analysis to explore UNICEF’s advocacy for girls’ education on Instagram. I consider a purposefully selected sample of photos and captions instagrammed from UNICEF’s official account so as to describe the policy framing of girls’ education policy, and population targeting. A parallel goal of this article is to interrogate the ethics of using image-intensive new media data in education policy research. My findings expose the ways in which girls’ images and experiences are used to promote UNICEF’s agenda and advocacy for girls’ education. I show the need for adapting protocols for working ethically with publicly available social media data in education policy research.


Introduction
Alessi and Alessi rightly describe the ethics of new media as an "uncharted area." They explain that "the production and subsequent use of new forms of media results in new dimensions of experience and effects in general that are impossible to understand through an analytical framework based on data gathered from old forms of media " (2008: 68). In contrast to old media-including print communication-new media enables users to be active participants in what they see, hear, and share with others in digital spaces. International development and humanitarian aid organizations, like UNICEF, have historically used old media tools to advance their agendas. Policy actors have expanded their engagement with new media tools to promote advocacy, brand their agendas online, and engage their stakeholder networks (Anderson 2016). I refer to the dynamic construction of policy frames-the interpretive schema used to orient policy goals and direct policy targets (Rosen 2009;Verloo and Lombardo 2007)-and the engagement of organizational-users with individual-users in relation to policy issues through new media as digital policy spaces. Moreno et al. (2013) identify three primary ethical concerns for researchers using new media data: privacy; consent; and confidentiality. Scholars have begun to grapple with how research protections extend to the text and network data sourced from individual users, but not how organizations engage with new media (boyd and Crawford 2012; Moreno et al. 2013;Warrell and Jacobsen 2014). UNICEF's position as a global advocate for children's rights and its organizational new media use creates a compelling case for the necessity of analyzing how girls' education policy discourses are constructed and diffused using image-intensive new media, and the ethical challenges that are presenting themselves to us in using this data in policy research on girls' education.

New Media Ethics in Digital Policy Spaces: A Case Study of UNICEF and Instagram
In this article I describe how UNICEF uses Instagram to frame the policy contexts of girls' education and I identify the ethical complexities of imaging girls in digital policy spaces. Instagram is an "image-intensive social media genre" (McNely 2012: 1) with an estimated 300 million users worldwide. It allows users to upload, manage, modify, and share images with accompanying texts. Following Rosen (2009) and Verloo and Lombardo (2007) in using policy frame and visual analysis, I describe how UNICEF uses Instagram as a policy tool in girls' education policy and international development and I point to the need for expanded protocols on research that involves human subjects for use in new media in education policy analysis.
My selection of UNICEF as a case for analysis was informed by Kirk and Magno's (2009) and Kirk's (2008, 2010) foundational work on the imaging of girls as target populations by international organizations and humanitarian aid agencies working in girls' education, and Nguyen and Mitchell's (2012) research on visual methodologies in education policy research. Magno and Kirk's collaborative work identified the importance of girls' voices and agency in creating and curating policy narratives. Magno and Kirk, and Nguyen and Mitchell discuss the importance of researchers reading images as a method to identify policy frames. Magno and Kirk (2008) concluded that "images are powerful tools used in the construction of gender relations, ethnic/racial difference and power relations as well as conceptions of development, underdevelopment and overdevelopment" (360). I extend their findings to interrogate how girls and education policy discourses about girls are framed by UNICEF through Instagram, and I confront the ethical challenges of using image-intensive new media data in education policy research.

UNICEF and Imaging Girls as Target Populations
The gendered dimensions of policy making mean that women and children are uniquely attractive to policy makers and international organizations, especially where education is incorporated as a development strategy (Burman 2005;Chant 2006). UNICEF's primary population of interest is children and its policy agenda is broadly conceived of as the promotion of their well-being. It positions itself as an influencer of global decision-making on humanitarian issues involving children and youth, and as unique among international organizations working with these populations (UNICEF 2015).
UNICEF sets stringent standards for the use of children's images. Its policy for photographing children says, The accurate and respectful visual representation of children everywhere is part of defending children's rights, including their rights to expression, privacy and protection. Recognizing that children are frequently at risk of abuse, discrimination, stigma or other exploitation if their name or visual identity is known, UNICEF promotes international photography norms that protect children's identities-in both the making and the use of images of children-as needed (UNICEF 2015).
UNICEF further requires that all personally identifiable information be withheld or obscured unless the child, parent, or adult guardian provides permission for use of the image (MediaWise and UNICEF 2005).
The images and texts shared through UNICEF's Instagram account include a unique identifier and a photographer credit. The distribution of credited, official photos signals that UNICEF's Instagram activities are best represented as placemaking processes. McNely defines placemaking as posts that "involve an organization 'placing' their identity within specific material locations as a way of reinforcing the organizations' core image" (2012: 3064). Meriläinen and Vos's (2011) analysis of international nongovernmental organizations' online engagement indicates that new media is used to share information, generate activism, and leverage their expertise and networks to inspire policy change. Instagram provides an opportunity for organizations, like UNICEF, to use image and text to placemake their work.

Research Questions
I conceptualize girls' participation as their active engagement in creating or curating the images that UNICEF instagrams. In their visual analyses of the pictures that UNICEF uses and of the policy framing of the United Nations Girls' Education Initiative (UNGEI), Kirk and Magno (2009) and Magno and Kirk (2008) highlight the importance of imaging girls' educational process and outcomes and including their participation in image production and message dissemination. Following their work, I formulate two questions: What processes and outcomes of girls' education does UNICEF construct through its use of Instagram? How is girls' participation reflected in image and message production and diffusion?
A primary concern for researchers using social media-sourced data is whether the images and texts used in their analyses were collected using appropriate ethical procedures. These procedures include protection of the subject's privacy, consent, and confidentiality (Alessi and Alessi 2008). Most of the photos included in this analysis disclose potentially identifiable information about the girls shown in the photographs. In order to problematize both ethical concerns-how girls' images and lived experiences are used to frame policy goals through new media and how researchers engage with image-intensive new media-I ask: What are the ethical challenges in using image-intensive new media data in the analysis of girls' education policy?
The ways in which organizational-users select and disseminate images of protected and vulnerable populations-like girls-further complicates the ways in which research protections extend to digital policy spaces and the ways in which researchers use this data.

Methodology
My use of policy frame and visual analysis methods in this work focuses, exclusively, on UNICEF's organizational use of Instagram as a policy tool. UNICEF's primary Instagram account is publicly accessible and, as of January 2016, reported over 800,000 followers. Similar to the approach that Preston et al. used in their study of photographs in annual reports, in this article I am most interested in "reading the images" instagrammed by UNICEF as a "way of seeing" (1996: 116) how girls' are constructed as target populations and how the policy issues that affect them are framed. Verloo defines a policy frame as an "organizing principle that transforms fragmentary or incidental information into a structured and meaningful problem, in which a solution is implicitly or explicitly included" (2005: 20). The "various dimensions" (Verloo and Lombardo 2007: 33) in which policy is communicated contextualizes how populations and issues are framed. Because policy frame analysis considers the content, the context/s, and the conditions within which policy messages are communicated, it provides an interpretive framework for both the visual and textual data sourced from Instagram.

Sampling
I constructed a purposeful sample of images instagrammed by UNICEF's official account over a six-week period in 2015. The target populations of interest in this article are girls and young women of school-going age-5 to 25 (UNESCO n.d.: n.p.). I selected images featuring a girl or girls as the primary subject/s of the photo for analysis. I collected these images manually at one time to reflect a non-probabilistic sample of photographs from UNICEF's official Instagram account. I archived a total of 38 photos at this stage. I removed four photographs and their accompanying captions from the sample because the sex of the primary subject of the photo was unclear or male, or because the subject was an adult, and to remove duplicates. I further refined the sample to focus on only the images and accompanying texts that aligned with two of UNICEF's priority areas-Child Survival and Development, and Social Protection and Inclusion. I chose to focus on the images that aligned with these two areas because they were over-represented in the sample. The final sample consists of 28 photos and their accompanying captions.

Content, Context, and Subject Position
After refining the sample, following Miles et al. (2014), I created a matrix to log descriptive characteristics of each photo and accompanying text. This matrix, as suggested by Christmann (2008) detailed the foreground and background content which included setting, place, and composition of each image; whether the photograph was a candid shot or a portrait; whether the subject was photographed in pose or in action; along with information about the compelling policy event referenced in the image or text. All posts included in the sample are captioned. UNICEF uses hashtags (for example, #Liberia) to placemake its agenda within each image and accompanying caption.

Agenda Setting, Policy Framing, and Compelling Policy Events
After organizing the images and captions for analysis, I then sorted the data into categories representing two of UNICEF's priority areas as mentioned above, Protection and Social Inclusion, and Child Survival and Development. I selected these two over-arching categories because they best reflect the content and context/s of the images and accompanying captions included in the sample. Education is designated as both a stand-alone and crosscutting thematic area in UNICEF's agenda. All but two of the posts included in the sample reference education in the image or in the caption. I used the captions that accompany each image to identify alignment with UNICEF's priority areas; UNICEF's response to a compelling policy event; how education is framed; and the ways in which girls are constructed as target populations. Compelling policy events refer to humanitarian crises, natural disasters, or social conflicts that provide a context for UNICEF's work. Two compelling policy events are reflected in the data sampled for analysis-the Ebola epidemic of 2014 and UNICEF's humanitarian action in Africa throughout 2015, and Cyclone Pam, which struck Vanuatu in 2015. UNICEF's focus on its work in reference to these two compelling policy events underscores the placemaking function of new media tools by international organizations.

Narrative Voice
I applied Kirk and Magno's (2009) visual analysis of girls as imaged subjects as an analytical framework to interrogate how girls' lived experiences are used to frame UNICEF's work. UNICEF captions photos using two types of narrative voice, organizational voice and subject voice. UNICEF uses organizational voice to craft the policy message that accompanies each image. Subject voice uses a direct quote attributed to the photographed subject. These categories were incorporated so as to reflect Kirk and Magno's recommendation that UNICEF and UNGEI "give greater attention to participation strategies for engaging children and youth and their communities engaged in image and message production" (2009: 28).

Findings
My analysis of the images and accompanying texts sourced from UNICEF's primary Instagram account identifies three policy tools in use: placemaking by imaging compelling policy events; organizational voice; and photo reportage. Education in emergencies is one of UNICEF's policy interests, and it is embedded in both the Protection and Social Inclusion, and Child Survival and Development domains. Each caption references a compelling policy event, an affected region, and UNICEF's policy response. Direct quotations from imaged subjects are included in five of the sampled instagrammed images and their accompanying texts. Of these five, only three include the voices of the girls imaged in the photos. Where both the subject's and UNICEF's organizational voice were used, the subject was quoted directly in her own words and her words were preceded or explained in context through the organizational voice. Subject voice is used sparingly in the sample, most often to contextualize UNICEF's response to a compelling policy event.

Girls' Participation and Framing Compelling Policy Events
Girls' voices are only incorporated into the captions in three of the instagrammed photos included in the sample. One of these photo images, Joana is placed in Vanatau following Cyclone Pam. Joana's age is not provided in the caption, but she appears to be in early adolescence. She is pictured alone, seated in front of her home that has been damaged by the Cyclone. The image is a still photograph sourced from a video. A subtitle-style caption appears in the image. Education in emergencies is the dominant frame. Organizational voice is used to contextualize the policy issue as lack of access to schooling and interrupted education. The caption reads: "Cyclone Pam destroyed her house and disrupted classes for tens of thousands of children like her." Here Joana is positioned as a symbolic representative of these "tens of thousands of children" affected by Cyclone Pam. Joana is quoted: "I want to go back to school soon because I miss my friends a lot." Her voice is used to reinforce the framing of access to education-especially in emergenciesas UNICEF's primary education policy goal.
"I like school because I want to be educated. When I grow up, I want to be a doctor." The words referenced above are attributed to Abier, a nine-year-old girl from South Sudan. Her words are included in the caption that accompanies a photo of her seated at a desk in a school classroom that is located in a temporary learning space funded by UNICEF. Abier is identified as an internally displaced person, temporarily relocated to Juba. Organizational voice is used to highlight UNICEF's support for education in South Sudan. The caption begins with Abier's words cited above and closes with UNICEF's: "So far this year, we have provided education to over 58,000 children affected by conflict in the country." Abier's image and her words are powerful symbols of UNICEF's work in South Sudan and provide a face, a name, and a testimonial to legitimate its work.
A girl's aspiration to become a doctor is reflected in another image and caption instagrammed by UNICEF during the collection period. This image features a posed portrait of Lydia, identified as a 15-year-old girl from Nigeria. She is standing in a mixed-age school classroom, holding a blue folder with UNICEF on the front cover. Lydia is quoted: "When the shooting started, we ran in panic into the mountains … I always feel scared when I think of what happened." The caption explains that Lydia is attending a "UNICEF-supported school in a camp for internally displaced people in Yola," after her village was attacked by Boko Haram. Organizational voice is used to narrate Lydia's lived experience fleeing her village, hiding in the mountains, and the trauma she experienced resulting from the murder of her uncle and cousin during the attacks. The caption reads "Lydia is attending school every day in Yola and wants to be a doctor. 'I want to help my community.'" Like Abier, Lydia's access to schooling is framed as a pathway to achieve her future goals.

Placemaking and Education Policy Agenda Setting
A focus on UNICEF's curriculum work emerges in the Protection and Social Inclusion area of its agenda. Two posts highlight an arts-based curriculum implemented in Liberia. The curriculum was developed through a partnership between UNICEF and the Liberian government. It aims to raise awareness of gender-based violence using an illustrated book. In one post, a group of early adolescent children is pictured in front of a chalkboard. The children, two of whom are girls, are holding a copy of this book. The caption explains that they are reading it aloud as an in-class exercise. In the other image, a group of all female students in early adolescence is pictured dancing in a classroom. It is unclear if the UNICEF supported arts-based curriculum is in use in either of the instagrammed photos, though both captions reference the curriculum.
Education is used to frame the experiences, outcomes, and opportunities for girls in all but two of the instagrammed photos. These photos picture girls on their way to school, seated at a desk, or engaged in schoolwork in a classroom. The school is the background in the majority of these photos, further reinforcing the policy focus on girls' access to formal schooling and their right to education as target populations. Securing girls' access to formal schooling is promoted as a dominant policy frame, and UNICEF's work to support teacher training and resources is highlighted in regional context. In all but one post, UNICEF reinforces the protective function of schooling as a safe place for children affected by violent conflict and displacement.
Girls and young women, as target populations (see Schneider and Ingram 1993), are most frequently constructed as daughters and schoolgirls. UNICEF's promotion and provision of education in emergencies is the dominant policy frame in all the photos in which mothers and daughters are imaged. Mothers and daughters are pictured in three of the photos included in the sample; one features a posed portrait of a mother and daughter and two are candid photographs of mothers and daughters going about their daily routines. In a candid-action photo a girl is crying in her mother's arms; the camera captures an emotional moment between them, which is contextualized in the caption. The young girl is identified as six-year-old Rosol and her mother as Neveen. The caption includes, "For Rosol, school is a painful reminder of the violence she experienced during the conflict." One unique characteristic of this photo and its accompanying text is that it complicates the idea that schools are always safe places for children, especially girls and children in conflict and emergency settings. The framing of education within this Instagram photo and caption distinguishes it from the other posts included in the sample. In all others, securing girls' access to formal schooling is the dominant policy frame constructed through both image and text.

Photo Reportage: "Mercy's Story"
Imaging girls as students reshapes their identity as schoolgirls, and highlights the ways in which childhood is increasingly organized by educational status and opportunity (Bloch et al. 2006). There are four instagrammed images included in the sample that tell the story of a single girl subject. The subject is referred to as Mercy, a nine-year-old girl living in Liberia whose mother passed away during the 2014 Ebola epidemic. I draw from elements of photo reportage and photo story to explore UNICEF's collection and dissemination of these four images. Citing Wolny-Zmorzyński (2007, 2011), Kędra defines photo reportage as consisting of "at least three photographs, ordered sequentially, of which one indicates the place where the event happened, the second specifies the time, and the third one presents the photographer's (visual) commentary to the event" (2013: 28). In contrast to photo reportage, Kędra describes photo stories as consisting of 8 to 12 photographs, taken by a single photographer, on a specific issue or theme. The collection of instagrammed photos that tell Mercy's story falls short of the range noted by Kędra but meets the other two criteria. All four photos are taken by the same photographer, as evidenced in the photo credit at the end of each caption.
These four images and their accompanying captions bring together the three policy tools I observe through UNICEF's use of Instagram: placemaking; imaging compelling policy events; and organizational voice. Taken together, the photo reportage of "Mercy's Story" constructs a policy narrative that placemakes UNICEF's work to advocate for schooling access in emergencies. Table 1 details each photo and accompanying caption included in UNICEF's photo reportage of Mercy's story. Table 1 Caption

Subject Position Setting
Remember Mercy Kennady? When we first met the with foster candid, at home, 9-year-old from #Liberia in November, she'd just lost her mother action getting mother to #Ebola, undergone quarantine and faced dressed stigma. But with the support and love of her foster family, she is starting anew -and she's even attending school for the first time. Click link in profile for more on her story. © UNICEF/NYHQ2015-0355/Grile Nine-year-old Mercy, who is in her first year of school in in school candid, outside Monrovia, #Liberia, lifts a younger girl onto the landing uniform, action school of the school building. Mercy and her 17-year-old lifting brother, Harris, live with their foster mother, Martu younger girl Weefor, in Monrovia's Paynesville suburb. Ms. Weefor, a friend of the family, is caring for the children after they were orphaned by #Ebola last year. © UNICEF/NYHQ2015-0351/Grile Mercy 9, fills a bucket with water at home in the collecting candid, outside Paynesville suburb of Monrovia, #Liberia. Mercy and her water action new 17-year-old brother, Harris are living with their foster home, mother, Martu Weefor. We first met Mercy in November foster shortly after she'd lost her mother to #Ebola and spent 21 family days in quarantine. Now in the care of a loving foster family she's starting anew -click the link in profile for more on her story. © UNICEF/NYHQ2015-0339/Grile Mercy, 9, sits outdoors on a step of her home in the seated on posed outside Paynesville suburb of #Monrovia, #Liberia. Mercy and stairs portrait new her 17-year-old brother Harris are living with their foster home, mother, Martu Weefor, following the death of their foster mother to #Ebola last year. For more on Mercy's story, family click the link in our profile. ©UNICEF/NYHQ2015-0341/Grile #UNICEF Mercy's lived experience is constructed using organizational voice and not her own. She is described in the captions, as detailed above, as a nine-yearold girl whose mother passed away during the Ebola epidemic of 2014. Despite the focus on Mercy as the primary subject, she is never quoted directly. Mercy's photo story is unique among the other photos and captions in the sample because it does not explicitly reference UNICEF's role. In all other photos in which Ebola is the compelling policy event, UNICEF explicitly references its activities to improve school sanitation and hygiene, train teachers, and provide psychosocial support and transitional housing for students. Three of the four photo captions direct the user to UNICEF's website for "more on Mercy's story." Though indirect, I argue that the inclusion of the link-in bio appropriates Mercy's story to promote UNICEF's agenda.

The Ethics of Imaging Girls as a Target Population
My analysis of the data sourced from UNICEF's Instagram account suggests that girls' education policy needs are not constructed as unique or distinct from children's. Their policy needs are framed as access to schooling, without much attention being paid to the policy contexts that enable (or disable) their educational opportunity and mobility. With the exception of one girl, those imaged in the instagrammed photos are reported to be living in temporary housing for displaced persons because of civil unrest, health risks, or natural disaster. Girls are uniquely vulnerable targets for violence and exploitation during crises, and locating them by name, face, and in place could, potentially, increase their risk.
The use of photo reportage through "Mercy's Story" reflects Kirk and Magno's guidance that images of girls and girls' education should show "policy in action" (2009: 26), and not girls in stand-alone, decontextualized settings. This particular case could be strengthened by increasing Mercy's engagement and participation. She is a passive subject in the photos; her story is told by employing a single narrative using organizational voice to describe her lived experience. UNICEF requires that photographers and partners "avoid categorizations or descriptions that expose a child to negative reprisals-including additional physical or psychological harm, or to lifelong abuse, discrimination or rejection by their local communities" (UNICEF n.d.: n.p.). Here, organizational voice may have reduced Mercy's risk of retraumatization. However, by not including her own words to tell her story, UNICEF's photo reportage has constructed Mercy as an "object of rescue" rather than a "subject of agency" (Giorgis and Farrah 2015: n.p.).

Implications
My analysis of data sourced from UNICEF's official Instagram account was guided by three research questions. In answering Research Questions 1 and 2, I investigated how girls are constructed as target populations of interest to UNICEF, and how girls' education policy discourses are constructed through Instagram. My analyses suggest that UNICEF's Instagram captions follow a consistent pattern to placemake its girls' education agenda across compelling events and national contexts. UNICEF uses organizational voice to frame girls' lived experiences and their educational and social outcomes. These findings highlight the opportunities that Instagram creates for organizations to draw attention to compelling policy events that uniquely affect girls. My analyses indicate that UNICEF relies on organizational voice to tell girls' stories so as to placemake its work on girls' education within its larger agenda. Approaches based on photo voice and photo story provide opportunities for UNICEF to better engage girls as creators and curators of the images that it uses to advocate for their policy needs. I conclude, through these analyses, that UNICEF could more deliberately incorporate girls' voices in its Instagram content.

New Media Ethics in Education Policy Research
Returning to the third research question, which focuses on the ethical challenges of using image-intensive new media data in education policy research, I conclude that Instagram can be leveraged as a powerful policy tool to frame girls' education policy goals in development and humanitarian aid contexts. I suggest two guiding practices for researchers using new media data in education policy research.

Interrogate Images as Agenda Frames
Imaging target populations reinforces normative constructions of gender, childhood, and vulnerability (Schneider and Ingram 1993). This creates a double-bind for policy actors because often the images that are most compelling reproduce developmentalist discourses about girls and women as yoked target populations with the same policy needs (Burman 2008;Anderson 2016). Constructing women and girls as a yoked target population conflates their unique and distinct policy vulnerabilities. Education for All (EFA) and the Millennium Development Goals (MDGs) advanced girls' educational access as a pathway to women's empowerment (Anderson 2016;Monkman and Hoffman 2012). MDGs 2 and 3 reinforced the access-focused education policy targets set through EFA Goals 2 and 5; the first of which focused on achieving gender parity and universal primary education while the second aspired to eliminate all gender barriers at all levels of education by 2015. UNICEF was a principal actor in MDGs 2 and 3, and in EFA (Jones 2006). As Lombardo and Agustín (2012);Unterhalter (2005Unterhalter ( , 2012, Kulawik (2009) andLather (1986) have pointed out elsewhere and in different contexts, this may have unintentionally reinforced the structural and institutional inequalities that inform those binary constructions because they do not attend to the intersectionality of gender, race, ethnicity, and place that influences how education policy problems and solutions are framed Imaging children as vulnerable populations is not a new practice in policy and development sectors. I argue that new media expands and accelerates policy discourses because it has the potential to reach larger and more diverse networks compared to the reach of old media (Alessi and Alessi 2008), and modes of discourse circulation like annual reports, policy briefs, and commissioned white papers, for example. This expanded and accelerated access to image-intensive new media data requires new approaches to positionality and reflexivity in qualitative research.
The images and texts that UNICEF constructs and circulates through Instagram are not agenda-neutral. The ways in which girls are constructed as a target population through UNICEF's Instagram account suggest an integrated and intentional approach to new media policy discourse dissemination. All images include a photographer credit and copyright that explicitly communicate that the images reflect UNICEF's organizational brand and agenda. The inclusion of photographer credit and copyright is important for researchers engaging with new media; it situates UNICEF's Instagram content in the public domain, and it provides researchers with assurance that photographs were taken in accordance with UNICEF's guidelines for reporting on children (MediaWire/UNICEF 2005; UNICEF n.d.).

Situate Data in the Public Domain
One of the strengths of using new media as a policy tool is that it enables dynamic interaction between and among policy actors. This dynamic engagement between organizations and individuals further complicates sampling protocols and the assent and confidentiality of imaged subjects as well as individual-users. The interaction between organizational and individualusers through new media creates the need for adapted research protocols to avoid potential threats to individual-users' privacy and confidentiality. The essential question is whether individual-users' participation in new media communicates assent to participate in research and policy analysis. I posit that participation does not communicate assent because of the potential threats to privacy and confidentiality for individual-users. Researchers should consider users' expectations about how their comments will be used in new media forums. In other words, just because an individual posts a comment to a photo posted by an organization like UNICEF this does not necessarily mean that she or he has assented to have those comments used in academic or policy research. The absence of protocols for individual-user generated data leaves no mechanism for individuals to revoke consent for the use of their data in research. Individual-users' engagement with organizations' official accounts can be, and are often, de-identified.
The inclusion of new media data sourced from individual-users accounts in research makes it possible to link individuals through comments and diffusion metrics-indicators that proxy reception. For Instagram, these metrics include the number of likes and user comments on each post. This potential for disclosure of individuals' privacy and confidentiality creates specific ethical concerns for researchers engaging with new media. A guiding practice to ensure that images are within the public domain is to sample directly from organizations' official accounts and not individual-users' accounts. The inclusion of comments made by individual-users to posts from organization accounts degrades the boundary between the public and private domains.
The incorporation of diffusion metrics to proxy reception presents additional challenges related to credibility and data integrity. A guiding practice for researchers to address potential threats to credibility and data integrity is to consider how their collection strategy might affect the diffusion metrics for individual posts, as well as their subsequent interpretation. As an example of these guiding practices in use, the analyses presented in this article rely exclusively on organization-user data sourced directly from UNICEF's Instagram account. I used a single-point collection strategy to build a sample from this account. To maintain data integrity, I chose not to include diffusion metrics in my analysis because the single-point collection strategy could, potentially, inflate the number of likes and comments on older posts. Diffusion metrics can be used to indicate the speed and spread of discourse diffusion, but I argue that they are not especially useful to understand individual users' sense-making and negotiation of complex policy discourses.

Conclusion
The visual and textual analyses in this article demonstrate the ways in which girls' education policy discourses are reconstituted as image-intensive new media content. The inclusion of image-intensive new media data generated by individual-users further complicates the ethical protocols typically used in social science research involving human subjects (Moreno et al. 2013;Warrell and Jacobsen 2014). The widespread availability of new media data necessitates adapted protocols to seek informed consent and to protect individuals' privacy and confidentiality. With these adapted protocols, future research can investigate internal and external sense-making processes used in sourcing, editing, and diffusing new media in digital policy spaces. This research would provide important insight into how policy frames are negotiated both off and online.
EMILY ANDERSON is an Assistant Professor of Education at Centenary University. Her scholarship explores the construction, diffusion, and negotiation of girls' education policy discourses and her current research investigates the role/s of new media in girls' education policy, activism, and advocacy. She is curious about the ways in which policy actors engage with new media to frame girls' education discourses and to engage communities in policy processes.