‘Meta’ analysis: considerations when using Facebook within research and evaluation studies: a research note

ABSTRACT

The use of social media platforms as a means of data collection for researchers is increasing. With the number of active users of Facebook across the globe exceeding two billion, the platform offers a means to reach and hear from diverse populations and gather information-rich data. Despite an increase of researchers utilising Facebook and publishing their findings, there has been less focus on describing the issues researchers need to think about to undertake their work in line with ethical principles, which is particularly important when engaging people in research. This research note provides a worked example of the practical steps adopted when establishing a Facebook Page and Group to recruit and gather data from individuals receiving care and support. The paper offers guidance for researchers on the issues to consider and provides insights into how using Facebook can be aligned with good research governance practice and ethics approval processes.

KEYWORDS:

• Facebook research
• social media
• online methods
• research governance and ethics
• Facebook groups

Introduction

Founded in 2004, the Meta Incorporation’s social network platform ‘Facebook’ is the largest and most used globally (Dixon, 2022; Hall, 2022). By September 2017, Facebook exceeded more than two billion active users, and by December 2022, had almost three billion (2.96) active users (Dixon, 2022).

Facebook Groups are spaces for individuals to communicate about shared interests with each other. Facebook members can create Groups based on their interests and invite others to join. Security settings are adjustable between ‘public’ (anyone on or off Facebook can view posts and membership), ‘private’ (only members can view other members and posts), and ‘secret’ (unsearchable, only accessible via invitation from another group member) (Lijadi & van Schalkwyk, 2015).

Using Facebook for data collection

Prior to the COVID-19 pandemic, Facebook Groups were already a popular means for users to connect with others, and their use has increased since then with 1.8 billion people actively using Facebook Groups each month worldwide (Hutchinson, 2020; Martin, 2022). Given the global reach and active use of Facebook and Facebook Groups, the platform offers a means of reaching diverse populations and gathering rich data from users (Franz et al., 2019).

Three principal ways have been described within which research can be undertaken using social media (Franz et al., 2019). Firstly, active analysis is the active involvement and communication of researchers with Facebook participants. Secondly, passive analysis involves researchers observing the information patterns or interactions between Facebook Group members. There can be challenges associated with accepting researchers as outsiders within communities of interest – also referred to as ‘lurking’, researchers do not interact or involve themselves, but observe online communication discreetly without being seen (Salmons, 2015). Thirdly, self-identification uses Facebook as a research recruitment tool for data collection, and participants are self-selecting. Social media may also reduce attrition and participant engagement via regular social media posts and updates (McGinley et al., 2015). An increasing number of studies, spanning various subjects have made use of Facebook to generate recruitment and data collection (e.g. Kent et al., 2016; Lijadi & van Schalkwyk, 2015; Medley-Rath, 2019; Penderson et al., 2015). However, previous studies have largely focussed on the data and findings gathered by this approach, rather than detailing, in-depth, the practical processes and ethical considerations to be contemplated before using Facebook for data collection.

This research note focuses predominantly on the approach we adopted for a Facebook Group, specifically on recruitment, planning for and mitigating for risk, Group membership, and data collection and analysis.

Aim

In this research note, we describe our approach to establishing and organising an online Facebook presence for individuals in receipt of care and support as assessed by local authority social services departments (whether service users or unpaid carers) in Wales, UK. Our intention was to utilise Facebook to recruit, engage, and hear from those service users and unpaid carers about their experiences under the Social Services and Well-being (Wales) Act (SSWBA).¹ This work formed part of a wider study to independently evaluate the implementation of the Act (the IMPACT study)² (see, among others Llewellyn et al., 2022, 2023).

The IMPACT study

The national evaluation of the SSWBA ran from 2018–2022 and examined the implementation and outcomes of the Act for individuals in need of care and support, unpaid carers and family members, the paid workforce supporting them, and the organisations who have responsibilities to discharge the Act’s duties (Llewellyn et al., 2023). IMPACT was supported by a study expert reference group, with three citizen co-chairs.

When it came to engaging with service users and unpaid carers, our original intention was to focus only on ‘traditional’ forms of data capture, including face-to-face interviews and focus groups, and online surveys. However, the onset of the COVID-19 pandemic required a shift towards remote forms of data collection. The creation of an online presence via a Facebook Page and Group not only offered a means to manage this challenge but, as suggested by one of our citizen co-chairs, it would enable us to reach out to our target group where they were, rather than expecting service users and carers to find and approach us.

Therefore, a public Facebook Page offered a means to increase awareness of our study and its outputs, whilst a private Facebook Group offered our target population an alternative means to participate, providing an additional data source to complement our existing methods.

Using Facebook in academic research: considerations and permissions

Ethical approval was secured from the authors’ institution [Ref: 210304 HR]. Our target group (service users and unpaid carers) were considered ‘vulnerable’ in our committee’s definition, requiring a high-risk ethics application to be submitted. Our intention to use Facebook as a recruitment and data collection approach was a first for our Faculty. Below we discuss the key aspects highlighted from the research governance review that was undertaken. Our intention is that this level of detail will be helpful for others making similar applications elsewhere.

Recruitment

Recruitment was congruent with active analysis, and research self-identification; moderators posted information to the Facebook Page, and actively participated in generating Group content. However, awareness of our Page, and recruitment to our Group was facilitated by trusted gatekeeper organisations.

Gatekeeper organisations, defined as those controlling access to groups, institutions, or organisations (Singh & Wassenaar, 2016), were invited to share the IMPACT Facebook Page and Group link with their members and contacts. The use of gatekeeper organisations for recruitment formed part of the overall IMPACT study. Trusted gatekeeper organisations came from extant relationships the study team held with community groups and organisations particularly within the voluntary sector, carers’ organisations, and with other networks of groups and organisations.

Prospective participants were reminded that the Group was one means to contribute to the IMPACT study, and that they could instead choose to take part in an individual interview, a focus group, or survey if preferred.

Planning for and mitigating risks: the importance of Facebook Group rules

The Facebook Group rules were set by the moderators/administrators (Facebook/Meta, n.d.-a) and outlined the importance of creating and maintaining a safe and respectful environment. Members were advised that moderators would monitor all activities and public comments, and that rules were established and would be updated if/when required. Examples included barring taking screenshots and sharing posts, spamming, adverts, offensive language, racism, and hate speech. Collecting data via Facebook simplified the management of such issues; if members were offensive, moderators were able to mute them, give a warning, or remove them from the Group.

Members were advised that they should be cautious when sharing personal information with others, encouraged to keep discussions within the Group, and asked to seek permission before direct messaging other members – noting that such messages were private and would not be moderated. Following ethical review, an additional rule was implemented which asked members to consider whether the sharing of any personal information might be potentially distressing to themselves, or other members before posting/sharing.

Additional information outlining how we would mitigate risk included:

Privacy guidelines

We made a clear distinction between the IMPACT Facebook Page and the IMPACT Facebook Group. The Facebook Page was public and accessible to anyone; only moderators could post content which focussed on raising awareness of the IMPACT study, and recruitment opportunities. Facebook users ‘liking’ our page received notifications of IMPACT, and its outputs. The private [closed] Facebook Group created a more secure environment for members to respond and participate (Franz et al., 2019). When obtaining data from members to be used outside a Group, moderators must ensure members are given notice, made aware of how their data will be used and stored, to obtain informed consent, and be clear that Meta/Facebook are not collecting the data (Facebook/Meta, n.d.-b).

Managing potential conflict within the group

A plan (with actions) to manage conflict amongst members, and mitigate escalation, was developed to resolve issues as soon as they were brought to the moderators’ attention. For example, in the event of such circumstances, it was agreed that in discussion with individual members (privately via direct messenger) moderators would remind members of the Group’s purpose and rules and to seek tolerance for different perspectives. In the event of unsuccessful attempts to resolve disputes, moderators would retain the sanction of removing Group membership if individuals were adversely affected by the ongoing conflict.

The information sheet outlined that if concerned or distressed, members should in the first instance contact their gatekeeper group or organisation and where contact details were not held, to contact the moderators who would provide gatekeeper details, or a suitable alternative and if needed/required make a referral to a suitable service.

Confidentiality and anonymity

The Facebook Page and Group function does not allow for pseudonyms; to join the platform, a user profile is required. Participants were advised via the information sheet that their contribution would be visible to other members and the moderators, therefore they would not be anonymous or confidential. Moderators monitored the Group throughout and ahead of data collection activities commencing, and on those occasions, members were reminded of the need for confidentiality, for example, so that they did not share information they were not comfortable with, and that they were free to choose to take part in some or all activities. Members were informed via the information sheet and consent form, that moderators had a duty to report any disclosures causing concern of a risk of harm to members or others to the appropriate authorities.

Group membership

To join the closed Group, members were required to be 16 years and over. Requests to join the Group were approved subject to agreement with the Group rules, and the completion of three questions to confirm their status as a person receiving care and support services and/or an unpaid carer living in Wales.

Members could represent themselves as individuals, or as part of an (unpaid) community group or organisation, for example, parent carer groups, advocacy groups etc. Only moderators were able to view responses and in the event of a partial completion, made contact privately via ‘Facebook Direct Messenger’ to clarify missing information and review the reply before accepting to the Group.

Group members providing consent vs no consent

Approved members were issued an information sheet containing a link to a consent form (via an online encrypted survey platform approved for use by the authors’ institution). This provided an accessible format in the event of potential formatting issues when trying to type names and initials onto a Word document, where a wet signature is unavailable. Upon receipt of consent, participants were confirmed as an active member whose data could be used in analysis.

To avoid excluding anyone from the Group, we decided not to remove members not providing consent. Our reasoning was to ensure members felt comfortable in the Group; if a member did not wish to provide consent when first approached, we did not want to prevent them from changing their mind, contributing, and providing their consent at a later date. However, acknowledging the need for informed consent to use participant data, the information sheet advised that if consent was not provided that ‘any information you provide through your contribution to the activities in the IMPACT Facebook Group will not be used in the analysis or findings of the study’. The study team reviewed this decision throughout the data collection period and embedded a copy of the information and consent forms (with online links) to the Group, as a reminder and for ease of access.

Data collection and analysis

Data collection occurred during the COVID-19 pandemic across fluctuating periods of lockdown (June-October 2021). Data collection comprised of a timetable of fortnightly bilingual polls (eight cycles), with regular reminders/updates posted to maintain engagement (McGinley et al., 2015).

Poll questions were drawn from IMPACT topic guides and based on the five principles of the SSWBA, well-being, prevention and early intervention, co-production, voice and control, and multi-agency working. Polls were multi-choice and respondents to each poll were sent a follow-up open question privately via Direct Message, which they could choose to reply/not reply to.

Group polls and private message data were extracted following the poll closure and the post and content removed. This meant there was no historical data to read or comment on later and ensured the Group page was current and up-to-date. Ahead of each poll commencing, a public post was issued reminding members that they could withdraw without giving a reason; choose to participate in the poll or not; if preferred, they could respond via direct message to the moderators instead; and that only contributions from consenting members would be included for analysis. In total, n = 63 service users and/or carers joined the Group, of which responses to each poll ranged from n = 6 to n = 15. A small number also choose to take part in an interview instead of contributing via polls. Following extraction, data (poll and free-text responses) were merged with data collected via interviews, focus groups, and surveys. NVivo 12 was used to organise IMPACT study data, and a coding framework developed using thematic analysis (Braun & Clarke, 2006).

Discussion

In this research note, we have provided a worked example of the practical steps that are needed to ensure that research activity using Facebook aligns with good research governance practice and academic ethics approval processes to engage with, and hear from, individuals in receipt of care and support.

Our research note contributes insights into establishing and organising an online Facebook presence as means of recruitment and data collection. This addresses the lack of detailed descriptions of approaches used (Franz et al., 2019), and an absence of guidance for researchers seeking to establish a Facebook presence for these purposes. In addressing this gap, our paper offers a set of practical consideration to support and inform other researchers. It is applicable to researchers in subjects beyond social care who engage with people in research, for example health, psychology, and sociology. It also has potential applicability across the social sciences, particularly in providing opportunities for those who may find more ‘traditional’ approaches daunting, to take-part and contribute.

Utilising online social media platforms such as Facebook for participant recruitment and data collection offers increased opportunities to reach a range of different communities. In the UK, 92% of adults were recent internet users in 2020, and the number of disabled users, and older adults (over 75 years) had increased since 2019, rising to 81% and 54% respectively (Office for National Statistics, 2021). These figures, together with the global reach and size of Facebook’s active user population implies that even under-represented groups such as older adults and disabled people are relatively large (Kosinski et al., 2015).

With a surge in active Facebook users during Covid-19, and 2022 figures showing an all-time high, the platform is a means for researchers to reach and engage with a breadth of participants across the globe. However, ensuring that research undertaken via this means is well-thought out, and conducted ethically, particularly when undertaking research with people is essential.

Whilst we have demonstrated the way that Facebook can be used in evaluation and research studies, limitations – like poor internet access, connectivity, and lack of devices – can clearly prevent participation (Penderson et al., 2015). Our study focussed on a specific piece of legislation, and population group in Wales. In this context, we considered there was little or no tangible gain to individuals falsely certifying their status as a service user and/or carer to take part. However, in studies focussing on e.g. politically charged topics with significant public opinion or media interest, researchers should remain mindful of the potential for deliberate misinformation and unauthentic engagement. The minimum aged to join the Group was 16 (the minimum for Facebook is 13 years); whilst we did not implement a rigorous process to confirm age, member profile photos and feeds indicated all were over 16 years. Nonetheless, given the number of children under 13 years with registered social media accounts (e.g. Ofcom, 2022), future research should consider how best to manage and mitigate potential risks when undertaking studies where falsified age-verification may be possible.

The time required to ensure that everything is in place to undertake the study should not be underestimated. During the process of launching and growing our membership within the Group, moderator responsibilities included following up individual requests to join but who had not confirmed their status as a service user and/or carer, or agreement of the Group rules.

However, following the steps described in this paper, we were able to effectively manage the Facebook Group once data collection had commenced. Acknowledging all the above, the Facebook Group was an effective approach through which people – who may otherwise have not taken the opportunity to have their say – could contribute to our study. This research note hopefully provides an insight into the type of issues others will need to consider if they want to do the same.

On reflection, we would have liked to see more people engage with our Facebook Group. Nonetheless, our approach was one part of an integrated multi-method study, and rather than a weakness, we view engagement numbers (n = 63) via this approach as a strength. These were service users and/or carers who may not have otherwise participated. We developed our approach in the absence of the ‘blueprint’ that we have provided in this paper, and hope it simplifies the process and enables other researchers to harvest the potential of Facebook in future research and evaluation studies.

Acknowledgments

We would like to thank Welsh Government for supporting this study. We would also like to thank our colleague Lisa Griffiths for their role in co-managing the Facebook Group.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The work which informs this research note was commissioned by the Welsh Government [C410/2017/2018] as part of the National Evaluation of the Social Services and Well-being Act (Wales), 2014.

Notes on contributors

Sarah Wallace

Dr. Sarah Wallace is a Senior Research Fellow at the Welsh Institute for Health and Social Care (WIHSC) University of South Wales. Sarah is an experienced qualitative (including phenomenology and participatory approaches) and mixed methods researcher. Her subject expertise is within the field of Violence Against Women, Domestic Abuse and Sexual Violence (VAWDASV), and she co-chairs the VAWDASV Research Network Wales. Sarah’s associated research interests span health and social care. Her work has included research with service users, carers and families, the social care workforce, and evaluating social prescribing interventions.

Mark Llewellyn

Mark Llewellyn is Professor of Health and Care Policy and Director of the Welsh Institute for Health and Social Care at the University of South Wales. Mark has undertaken more than 100 specialist research studies and evaluative reviews and gained considerable experience of managing and delivering complex and sensitive studies specifically in health and social care, and across the public sector in the UK. He has developed an expertise in applied research and evaluation methodologies and his work is focused on providing a robust and independent evidence-base in all WIHSC’s work.

Notes

1. https://www.legislation.gov.uk/anaw/2014/4/contents

2. https://wihsc.southwales.ac.uk/evaluation-implementation-social-services-and-well-being-wales-act-gwerthuso-gweithrediad-deddf-gwasanaethau-cymdeithasol-llesiant-cymru/