Developing a Facebook page as a research tool

I am preparing to set up a Facebook page for my dissertation research. Below I’ll briefly summarise 2 key pieces of literature that will help me to design the page, consider the challenges, ethical considerations and different uses of the platform.

The first text is an article titled: Facebook as a Research Tool for the Social Sciences : Opportunities, Challenges, Ethical Considerations, and Practical Guidelines Kosinski, Michal ; Matz, Sandra C. ; Gosling, Samuel D. ; Popov, Vesselin ; Stillwell, David (2015)

According to the authors, these are some of the benefits of using Facebook for research:

• Facebook constitutes a large and diverse pool of participants, who can be selectively recruited for both online and offline studies.
• Facebook can be used as a powerful data-recording tool because it stores detailed demographic profiles and records of an enormous amount of actual behaviour expressed in a natural environment.
• There is an option to post an advert on Facebook targeting a specific sub population or replacing an online or offline demographic survey with a button that enables participants to donate their behavioural and demographic data to the researchers.
• Participants can receive immediate feedback on their contribution.

And some of the challenges in using Facebook as a research tool:

• Researchers need to rethink research design and acquire new skills, including more advanced statistical techniques.
• Greater care must be taken to meet participants needs and to optimise their experience in the time available.

The authors also identify a couple of different approaches or methods of carrying it out:

Using a snowball sampling approach – This involves convincing Facebook users to invite their friends to join a study. The chances of a study going viral can be increased by making “inviting friends” an integral part of the experience. Although snowballing can introduce biases as the first participants [who are termed seeds in this study] are likely to disproportionately affect the composition of the sample. But the size and diversity of the Facebook population help to minimise the disadvantages of snowball sampling. Given enough participants, the representativeness of the population can be improved by weighting.

Targeting: Recruitment Through Facebook Advertising – This is an alternative to snowball sampling, although there is likely to be associated costs. Facebook ads have been proven to outperform traditional methods such as postal surveys, and they are more cost efficient than Google advertising, online newsletters, and e-mails.

Collecting Facebook Profile Data – Facebook profile information includes self-reported information (e.g., schools attended, current workplace, age or gender), traces of behaviour (e.g., status updates or likes), and data contributed by the others (e.g., photo tags or comments on a user’s wall). At the time of writing [2015] the following categories of information could be recorded with users’ permission: -demographic profile -user generated content -social network structure (list of friends) -User preferences and activities (group memberships, attended events, installed apps) -information about users friends -Private messages. Data collected from Facebook can be further enriched using external models and services such as Pennebaker’s Linguistic Inquiry and Word Count, which is used to tag status updates with their emotional valence. The pitfalls of using Facebook profile data include drawing conclusions that are acceptably free from influences by overt biases, and to assess the impact of potential hidden biases. Some parts of the Facebook profile are self-reported, and others can be selectively removed by the profile owner. Additionally, users can use privacy settings to limit access to parts of their profiles. The platform wide privacy rules are constantly evolving. I will aim to look into the most up to date privacy and data information.

The authors advise for researchers utilising Facebook to include a discussion of ethical considerations related to the design of a study in their publications. The study should discuss the ethical implications of the findings which would support the evolution of standards and norms in the quickly changing technological environment. I will recognise the ethical considerations and other issues of using this method in the introduction to my dissertation, when justifying the methodological approach.

I am unsure at this stage whether I will need to refer to demographic profile information in support of my research. I will also need to decide whether to protect a participants anonymity; potentially, personal information and demographic data will be important to the stories people tell or may be necessary for me in contextualising stories. Likewise, perhaps all I need is the autoethnographic account itself, to analyse thoroughly. It seems likely that I will need to use.= a snowball sampling approach, by asking people to share in their networks or recommend to friends.

The next text is called: Remembering in Public: A case Study of museum-user communication on Facebook, by Emily Oswald (2020). It is a qualitative study of how museum–user communication occurs on the Museum of Oslo’s Facebook page and “toward what ends” the
activity is directed.

On a weekday in September 2016, the Museum of Oslo’s institutional Facebook account posted a black-and-white image from the museum’s photographic collections, showing a curving, snowy street flanked by a four-story brick building. The photograph was captioned “Ser du hvor dette er? [Do you see where this is?] 1966.” In the 36 hours following the publication of the post, 145 users commented on it or replied to the comment of another user; 823 users “liked” the photograph.

Oswald, 2020

Since 2013, the museum has published one historical photograph from the collections every day, captioned each time with the same question, “Do you see where this is?” The museum always comments on its own post to confirm the location shown in the photograph, generally twenty-four hours after the initial post was published.

Recommendations referenced in this text/things that worked well:

• more users commented in response to quizzes or requests for help from the museum, leading the authors to observe that users “adopt explicit museum invitations to participate when participation is deemed relevant” (Gronemann et al., 2015, p. 183).
• Frobenius and Harper (2015) proposed that users treated comments as a “quasi-conversational space”
• To develop a data corpus appropriate for qualitative analysis by a single researcher, thirty-six posts were selected (one post from each month for the three-year period).
• As they published their comments, users were attentive to the technical features of Facebook as a platform, suggesting multiple understandings of how a comment on a post or a reply to a comment might function.

Things to avoid:

• The genre of pseudo questions, in which a museum posed a question it already knew the answer to, likewise resulted in few comments from users.
• The genre “stories” that constituted the majority (41%) of posts “focus[ed] on the museum pushing nicely illustrated content, inviting little or no audience involvement” (Gronemann et al., 2015, p. 183), and elicited few responses.
• It may be difficult for users to manage the large number of potential participants in the conversation-like spaces of comments on a Facebook post.
• Gauging whether and how to appropriately engage with an institution on Facebook involves the added difficulties of knowing how to communicate with an institution or organization and the presence of users from outside one’s own network

Oswald undertook the following data analysis in 3 consecutive procedures: 1) Observation of user comments that were generally about the location in the picture, comments were surprisingly diverse in terms of what users talked about. Users’ comments on a single post also referred to personal memories of the neighborhood pictured, described urban renewal and the built environment, expressed appreciation to the museum for publishing the photograph, and referenced a song by a popular Norwegian folksinger that included the name of the neighborhood. 2) Categorisation of user comments. 3) Types of user activity (identifying a place, sharing and soliciting knowledge, and remembering in public) were selected for further analysis. Comments were arranged into the table below:

This study was incredibly interesting and relevant to my research project. Findings include that commenting on the museum’s post was, for many users, an activity related to displaying an identity on social media, rather than interacting with others, but I will be mindful to ensure that posts and content are strategically phrased to encourage conversation where possible. However, the findings also suggest the importance of investigating how and why posing questions may produce different responses from users and different types of user activity including dialogue. The user response to the Museum of Oslo’s Facebook concept strongly supports Gronemann and others’ (2015) conclusion that Facebook users respond to museum posts that involve a specific invitation to participate or interact. It also provides empirical support for the claim that museums’ use of social media can create opportunities for such communication “to develop around collections and disciplinary knowledge” (Russo, 2012, p. 151).

Further things to research: 1) Data about posts published by the Museum of Oslo between October 1, 2013 and September 30, 2016, were downloaded using the social network analysis
software, Nodel XL. 2) Reeves and Brown’s (2016) conclusion about the usefulness of conversation analysis for studying social media.

I have started to research Facebook groups including local history groups where members can share memories and compare old and new photographs (above), ‘art groups’ that welcome contributions of art from specific time periods and styles alongside a list of criteria, do’s and don’ts (below). I have started observing interactions, comments and conversations, alongside whether an image is posted alongside a leading or an open question and what the nature of comments are. Next, I will decide whether to use a group or a profile page, what title to give it, what to include in the data consent text and produce a plan for the first post and where to distribute it. I will design the first post using notes I took from Autoethnography (Adams et al, 2014). In particular, Linda Barry’s recommendations of the following steps:

• 

1. Beginning with a prompt (a noun— such as “dogs I have known” or “phone numbers” or “cars”—or a verb— “telling,” “missing,” or “finding out”). Use this prompt to make a list of the first 10 images you think of/about. Review this list and pick one of these images.

2. Picture your image by answering who, what, when, why, where, and how questions: Who is in the image with you? What time of day is it? What time of year? When did you arrive in the image? When did others arrive? Why are you in the image? Why are others there? How are you positioned, speaking, and/or moving in the image? How are you and/or others relating in the image?

3. Get inside your image by writing about the scene: What is in front of you? To your left? To your right? Behind you? Above you? Below you?

4. Set a timer and write uninterrupted for 10 minutes about, in, and through your image, beginning with the words “I am . . . .” 

I can adapt these questions and potentially disseminate them throughout different posts to encourage Facebook users / research participants to undertake their own autoethnographies and to suggest a structure for them to follow in their accounts. I will also experiment by integrating a number of the different interview techniques that are listed in the book:

• Emergent interviewing , in which informal interactions lead into more structured question-and-answer sessions or in which interviewers visit with participants in their everyday environments and contexts, conversing and asking about experiences, practices, and perceptions as they happen
• Sensory-based interviewing , in which researchers and participants focus on sensory experiences, surroundings, metaphors, or memories
• Participatory photo (or photo-voice) interviewing , in which participants are asked to provide or take photos of their identities, experiences, cultures, and other aspects in their lives; the meanings and significance of these photos are then discussed collaboratively with researchers
• Interactive interviewing , in which researchers and participants share personal/cultural experiences and tell (and perhaps write) their stories in the context of their relationship.