The Research Design - Natural Social Spaces: How Do We Research Social Media and Development Trends, Dynamics, and Impact? Learning from a Polio and Routine Immunization and Social Media Research Initiative in Ukraine

The Research Design

To examine how different communities use social media to communicate in Ukraine, within the context of child health and the research questions, the following research design was developed and implemented by Ukrainian researcher Anna Postovoitova with support from The Communication Initiative through USAID and MCSP.

The overall design was based around a qualitative study to examine how different communities in Ukraine use social media to communicate and search for information on polio, immunization, and child health issues. Effort was made to manually search for, monitor, and identify content trends in popular and thematic social media spaces on Facebook and Vkontakte networks. Analysis was made of active discussions and information shared within online pages, with a specific emphasis on the purpose for that engagement and the themes of the vaccination-related content either shared or in the form of questions for others to respond.

From that base, the study examined trending topics on routine immunization during the period October–November 2016. Conclusions were drawn about the possible implications of circulated online pro- and anti-vaccination content for parents' vaccination decision-making. This monitoring timeframe coincided with the above-described renewal of a mass child immunization campaign after a temporary shortage of vaccines, which greatly influenced the dynamic of social media conversations.

Thematic social media spaces are usually created as a public page or in the form of an open or, alternatively, a closed group. During this study, access was requested to closed groups and, if accepted, information was gathered for the study, but without intervention in the group dynamic.

The study's experimental approach of social media mapping involved discovering platforms where communities of parents, health and education professionals, and other civil society actors gather naturally. It involved collecting lists of, first, popular public pages and personal pages with a large number of followers and, second, thematic social media spaces that are organized in the form of a public page or, alternatively, an open or closed group.

Content about polio, vaccination, and routine immunization on each social media space was searched manually and later reviewed and assessed as either supportive of, or resistant to, vaccination efforts.

In addition to describing the character of communities' engagement (preferred social media network, level of activity, topics of discussions) on social media and the sentiment towards immunization, vaccines, and polio within those spaces, the dynamics of social media conversations were analyzed. This analysis included the volume of conversations on a certain topic, frequently asked questions, and time of day (or night) of the most engaging posts.

Monitoring of popular social media spaces illustrates the general interests of Ukrainian people. These pages are major sources for information; the content posted within them can reach a large audience. Results showed that of 300 public pages and personal accounts on Facebook and Vkontakte, 34 included information supportive of routine immunization, 14 had negative sentiment, and 252 were without posts relevant to this topic (Figure 3). Resources with vaccination-related information are limited on the Vkontakte network. Only a few popular social media spaces featured content about vaccines.

In greater detail, four important elements of this overall research design included: where are people gathering online to focus on vaccination and child health issues; how can we gain access to those conversations; how shall we engage as part of the group; and how should we organize and analyze the data?

A. Identifying where people are gathering online

No mapping was available for natural (started and managed by the public, not by government or international development agencies) social spaces that were in place and being extensively used in Ukraine. This was a crucial problem. We did not know where the people we were seeking to engage are gathering in large numbers.

So the first central element of the research design was to map these spaces. The key questions for the mapping were:

• Which are the main platforms and spaces?
• What are they called?
• Who are the administrators?
• How many people are involved?
• Who are these people—what are some of their key characteristics?
• To which locales or issues do the groups relate?
• What are some of the main discussion themes?

In total, 660 social media spaces were identified and selected for further assessment of their relevant routine immunization and vaccination content. These included:

• 100 popular public pages
• 100 popular personal pages on the Facebook network
• 100 popular public pages on the Vkontakte network
• 20 social media spaces of the education community
• 10 social media spaces of the medical community
• 50 health-related social media spaces
• 50 children-related social media spaces
• 173 city groups and public pages
• 34 parents' groups
• 23 social media spaces with significant focus on polio and vaccination

In total, these spaces included 2,427,431 accounts of parents and city residents, 159,076 accounts of health organizations, and 4,735 accounts representing specific polio and vaccination spaces.

As a test of the natural spaces premise behind the research, we compared the scale of these natural spaces with the scale of online spaces developed and managed by development organizations with specific programming goals—for example, by UNICEF Ukraine and WHO Ukraine. The membership ratio was 544 to 1 in favor of the natural spaces.

The 173 city groups number may look a little strange. We attempted to identify the two largest online city groups for each of the two largest cities in each of Ukraine's 24 oblasts (province/state) on Facebook and Vkontakte. As can be seen by the odd number, some only had one.

The identification of these groups is of course in and of itself a major research result. Just being able to identify major social media groups, who they engage at what scale, and their key facilitators and administrators, provides important insight into potential online spaces for further engagement. But we wanted to go further. There was a research interest in the content—what issues are being raised, when they are being raised, how they are being answered, and who is doing the answering. In order to get to those questions, it is necessary to gain access to the groups themselves.

B. Gaining access

The challenge of gaining access to these groups to gather the raw knowledge from which to assess patterns and trends and identify strategic implications for development organizations raises both practical and ethical questions.

At the practical level, these are groups defined by interest (e.g., parents' groups) and location (e.g., the Rivne oblast). But these are often public groups that anyone can join. So, even though the researchers are not necessarily parents and do not live in the oblast of interest (e.g., based in Kyiv, but the city group is for the city of Rivne), it is possible for the researchers to join the group. Once they have joined, they can review the knowledge, information, and ideas shared in the same manner as all members of that group.

That ease of access heightens ethical questions. As researchers, we have a different interest from others in the group. The interventions of people who are genuine members of that group require respect and confidentiality. Openness about the purpose of the researcher's role as a participant is also required, even if it is only in a passive or "listener only" role.

There was significant learning from this research on this theme. We had to explore the best ways forward. The research design implications from that learning can be described as follows:

• Group Administrators: Identifying the group administrator(s), making contact with that person/people, explaining the purpose of the research, and seeking permission to engage is a crucial step.
• The Researcher Profile: Most groups require a profile from the people joining. The researchers need to be accurate and clear in their profiles concerning why they are part of this group.

C. Engaging as part of the group

The possibilities for a Hawthorne effect are large. On sensitive issues, in often difficult local and national contexts, people engaged in online groups could alter their behavior—questions, comments, level of engagement, focus of their interest, etc. This distorts what is being observed—distorts what would have been the natural course of the engagement between people in that group. This is what happens on online platforms that are established, facilitated, and branded to a specific development organization, for example, UNICEF.

From the experience in this research, the following two key research principles emerged:

• Remain silent—do all possible not to interrupt the online process that is taking place. As tempting as it may be to engage, resist.
• Engage only when necessary—if you really have to engage, 1) make it a question, 2) provide a link to your profile, and 3) explain the context and/or background to the question posed.

D. Undertaking content analysis

How did we undertake the content analysis part of this research? Again, this was an exploratory process, as there were few, if any, models to follow. After much exploration and refinement, the Postovoitova research ended up organizing and categorizing the content according to the following categories:

• Themes: What is the main content focus of the engagement and what patterns can be detected? For example, is the person engaging to find where vaccines are available, to promote a view that the vaccines are safe, to ask where the vaccines are manufactured, to critique the performance of medical staff?
• The level of proactivity: To what extent was the shared knowledge a proactive intervention and what patterns can be detected? For example, before a mass delivery of vaccines, were people actively sharing this knowledge and asking core questions about access and safety?
• Timing: When does the engagement take place and what patterns can we see based on that timing? For example, evening and late evening time are recorded to have the highest level of critical examination of the safety and added value of vaccines and vaccinations.
• Tone: What tone describes the nature of the engagement within the group and what patterns can be established, for example, the numbers who are frustrated, concerned, content, happy, questioning, supportive?
• Sentiment: Related to vaccines, what is the sentiment expressed, for example, concerned about safety, worried about ability of health system staff to administer, fully supportive, looking for best place to have the vaccine administered?
• Locality: Where are people connecting from, for example, home, work, friends, health center, school? How can content or expert opinions relate to the environment of an intended audience?

---

¹⁰ Postovoitova presentation, page 9 - http://www.comminit.com/files/annapresentation.pdf

---

Editor's note: Above is an excerpt of a study made possible by the generous support of the American people through USAID under the terms of the Cooperative Agreement AID-OAA-A-14-00028. The contents are the responsibility of the Maternal and Child Survival Program (MCSP) and do not necessarily reflect the views of USAID or the United States Government. The full table of Contents is here.

The previous section in this paper is The Research Questions.
The next section in this paper is Conclusion.