Regular Article

International Journal of Qualitative Methods

Volume 21: 1–12
Using WhatsApp Focus Group Discussions © The Author(s) 2022
DOI: 10.1177/16094069221090355
for Qualitative Data Collection: Exploring journals.sagepub.com/home/ijq

Knowledge, Attitudes, and Perceptions of

COVID-19 in Singapore

Pearlyn HM Neo1 , Jane M Lim1, Rayner KJ Tan1,2, and Suan Ee Ong1,3

Abstract
This qualitative study aimed to explore Singapore residents’ knowledge, attitudes, perceptions, and behaviors around COVID-
19 as shaped by different information sources. Through utilizing WhatsApp as a means of conducting digital focus group
discussions (FGDs), participants were involved in five consecutive days of discussions through both synchronous and asyn-
chronous means. We found that the use of WhatsApp as a means of conducting FGDs not only served as a means of generating
essential, time-sensitive data in the community, but also advanced the quality and quantity of data generated, democratized, and
enhanced the participatory nature of FGDs, and facilitated the communication of potential issues around data privacy between
facilitators and participants. Although challenges around privacy and confidentiality remain, this means of collecting data is novel
in terms of providing timely and relevant data during a pandemic and would be appropriate to be further utilized in the context
of other health-related research beyond a public health emergency.

Keywords
focus groups, digital data collection, qualitative methodology, COVID-19, Asia/Southeast Asia/Singapore, health behaviors,
health information

Introduction and Background In recent years, digital technologies focusing on text-based

data collection—be it through Zoom, Skype, Facebook or
Since the declaration of COVID-19 as a Public Health Reddit—have been increasingly used in qualitative research
Emergency of International Concern on 30 January 2020, we (Thunberg & Arnell, 2021; Vindrola-Padros et al., 2020) due
have seen a global escalation of public concern, panic, and to the ability of these platforms to enhance long-distance
misinformation around the pandemic. Unlike during SARS in participation, extract socio-cultural nuances through varied
2003, the advent and pervasiveness of social media and new
modalities of communication and information exchange like
WhatsApp, Facebook, Instagram, and Twitter have created new 1
dynamics in how the public interact with health and outbreak- Saw Swee Hock School of Public Health, National University of Singapore,
Singapore
related information. To investigate concerns around propaga- 2
University of North Carolina Project-China, Guangzhou, Guangdong, China
tion of misinformation, lay health beliefs, and hysteria through 3
Research for Impact, Singapore
social media in the context of a pandemic, we conducted a study
to explore the knowledge, attitudes, perceptions, and behaviors Corresponding Author:
Pearlyn HM Neo, Saw Swee Hock School of Public Health, National
around COVID-19 among Singapore residents using social
University of Singapore, Block MD1, 12 Science Drive 2, Level 8, Resource
media; specifically, focus group discussions conducted via the Centre, Tahir Foundation Building (TFB), Singapore 117549, Singapore.
mobile messaging platform WhatsApp. Email: [email protected]

Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons
Attribution-NonCommercial 4.0 License (https://creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use,
reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE
and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
2 International Journal of Qualitative Methods

modes of expressions like GIFs and emojis (Wong & Jensen,

2020), and capitalize on the increasingly omnipresent nature
of digital means of communication. With the changing nature
of research brought about by the on-going COVID-19 pan-
demic, WhatsApp focus groups have grown increasingly
popular as a form of text-based data collection tool with
numerous studies adopting this tool for public health research
(Anderson et al., 2021; Colom, 2021; Gibson, 2020; Singer
et al., 2020). Compared to traditional forms of in-person focus
group discussions, WhatsApp focus groups present several
advantages logistically and methodologically. For instance,
studies have shown that participants in WhatsApp focus
groups are more comfortable sharing sensitive information
and disagreeing with one another, creating space for more
inclusive discussions (Colom, 2021). WhatsApp focus groups
also allow for different modes of communication (e.g.,
through incorporating pictures and links), are more cost-
efficient and allow researchers to recruit a greater diversity
of participants (Wong & Jensen, 2020). However, the lack of
non-verbal cues in the online environment, security and
privacy concerns, and a possible breaching of confidentiality
could also be reasons why such forms of digital focus groups Figure 1. Screenshot of the official government WhatsApp push
might not work for all research settings (Reid & Reid, 2005). notification sent to subscribers daily.
An earlier study on digital focus groups using WhatsApp in
the Singapore context (Chen & Neo, 2019) found that Sin- conduct urgent data collection during a pandemic. As such, the
gapore was an appropriate place to conduct research using team found it appropriate to adopt WhatsApp as a data col-
digital means. WhatsApp focus groups had the potential to lection platform due to the familiarity of Singaporeans with
generate well-elaborated responses and group interaction, the platform and the pervasiveness of WhatsApp’s use as a
particularly among younger, digitally fluent participants. An- platform of everyday communication among Singaporeans.
other study conducted using WhatsApp focus groups in Sin- In this paper, we draw on our experiences conducting
gapore examining public perceptions of risk and trust also digital focus group discussions via WhatsApp during the
found WhatsApp to be highly effective at extract specific socio- COVID-19 pandemic to make recommendations on how to
cultural nuances due to the familiarity that participants in adopt this approach for qualitative health research.
Singapore had with WhatsApp as a text messaging platform
used by most people daily (Wong & Jensen, 2020). Singapore,
Methodology
where this study took place, is also ranked second (88.4%) in
terms of having the highest internet penetration rates in This study aimed to explore Singapore residents’ knowledge,
Southeast Asia. 91% of the population use a smartphone, with attitudes, perceptions, and behaviors around COVID-19 as
WhatsApp being the most highly used messaging platform, shaped by different information sources. In addition to un-
with 87.1% of the population between the ages of 16–65 years derstanding the propagation of pandemic-related information
old using it daily (We are Social, 2021). During the COVID-19 and misinformation, we also investigated the feasibility, ac-
pandemic, the government had also created a push message ceptability, and rigor of collecting focus group-based qualitative
dissemination service over WhatsApp, where residents would data using a common, trusted, free-to-download, free-to-use
be sent COVID-19 updates, key government announcements mobile messaging app. Subsequent data analysis focused on the
and clarification on widespread fake news regarding govern- similarities and differences in knowledge, attitudes, percep-
ment policies daily (see Figure 1). As of 1 February 2021, 1.22 tions, and behaviors across different age groups, with particular
million Singapore residents have registered for this service emphasis on how individuals share, post, and interact with
(MCI’s response to PQ on impact of change in WhatsApp’s health and outbreak-related information and how such infor-
privacy policy on Government communications, 2021). mation affects their behaviors. Evidence in the published lit-
However, as previous research was conducted pre- erature also suggests that there are differences in usage patterns
pandemic, it was still unclear how people would respond to for WhatsApp across age and educational attainment
the ongoing nature of such a focus group discussion during a (Rosenfeld et al., 2018). As such, eight WhatsApp-based focus
lockdown. The team was also interested to find out the extent groups were conducted with participants stratified by age
to which a person’s surroundings would affect the quality of groups, namely, 21–30 years, 31–40 years, 41–50 years, and 51
response, with WhatsApp allowing for the physical safety to years and above.
Neo et al. 3

representation across dimensions of ethnicity, gender, and so-

cioeconomic status. For groups that were overwhelmingly male
(i.e., 51 years old and above), we sought to purposively recruit
more female participants within that age bracket by repeated
calls on social media and snowball sampling. The team then
replaced participants in the existing digital focus groups to
achieve greater representation across these demographic
characteristics. All these were done internally, and the research
team only “created” a group chat a few hours before the first
synchronous discussion on the first day of the digital focus
group, which took place over five consecutive days.
Once groupings were finalized, the research team then
reached out to participants individually to share the participant
information sheet and obtain informed consent. This included
informing participants about study-related ethical concerns
and risks, Participants were informed about the risks sur-
rounding the study and ways in which they could mitigate
Figure 2. Recruitment poster for online and social media these risks. For instance, recognizing that a lack of anonymity
distribution. was a potential concern, the research team informed partici-
pants that they could change their names and display pictures
Participant recruitment and selection to protect their anonymity throughout the duration of the
discussion. However, while participants were aware of this,
Participants were recruited through the team’s personal net- very few participants changed their display pictures or gave
work via social media (Facebook, Twitter, Instagram, themselves a pseudonym during the actual FGD.
WhatsApp) using a standardized poster that was circulated Other ethical concerns and risks (see Figure 3) included
(see Figure 2). Participants who were Singapore citizens or participants being able to screenshot and share responses from
permanent residents, who were aged 21 and above and were the group chat on to social media platforms without consent,
WhatsApp users were included in this study. To indicate their having one’s personal details given out to other participants or
interest to participate in this study, potential participants having one’s phone number tagged to one’s personal details and
completed a pre-qualifying survey by scanning the QR code the potential of being contacted privately outside the confines of
provided in the poster, or to access the link given in the poster. the FGD and/or receiving unsolicited texts from other partic-
This pre-qualifying survey allowed the team to stratify people ipants. While the team screened participants before admitting
across demographics such as age, education level, housing them into the FGDs, private messages beyond the group chat
type, marital status, gender, and ethnicity to ensure greater where the FGD was being conducted was beyond the facili-
representation in the study. tation team’s control. Participants were asked to immediately
The research team launched the recruitment poster at about inform facilitators should they feel uncomfortable at any point
3p.m. on 10 March 2020.18 hours later at about 9a.m. on 11 in time, or if they had received unsolicited messages from other
March 2020, the team had received a list of about 109 sign- participants. Throughout the duration of the FGD, we had one
ups. Participants who signed up were largely skewed toward participant who had received an unsolicited text message from
those who were in the younger age categories (i.e., below 40 another participant in the same group. The facilitators then
years old). This is likely due to the personal social networks of privately messaged the sender to stop such behaviors and in-
the research team where the recruitment call was disseminated formed the receiver to block the sender on his/her end. The
to and consisting of people who were in the same age bracket facilitators also reminded the sender that if such behaviors
as the research team members, and due to a higher proportion persisted, he/she will be removed from the FGD entirely and no
of young individuals on social media channels. longer be permitted to participate.

Preparations for participants and researchers Data collection

Participants were first sorted by age group and eligible par- All eight digital focus groups were conducted over the course
ticipants were then approached individually to provide their of 3 weeks from end-March 2020 to mid-April 2020. Each
availability. Participants were then further sorted by time group consisted of eight participants, a lead facilitator, and two
availability to match their schedule with the fixed synchronous to three assistant facilitators from the research team whose
group discussion time slots. Participants were then randomized main role was to take note of main points that surfaced from
within each age group to get 6-8 participants per digital focus the group discussion in a separate “field notes” WhatsApp chat
group. The team then sought to assess if we had sufficient group that was created internally for the team to communicate
4 International Journal of Qualitative Methods

Figure 3. Screenshot of informed consent process where participants were informed about the ethical concerns and security risks of being in
a WhatsApp focus group discussion.

with each other while the focus group discussion was going active and have more messages sent per hour, the facilitator
on. Each focus group lasted five consecutive days with a also needed to summarize points that were being mentioned by
different prompt given each day, taken from a topic guide that various participants and help participants have checkpoints so
was developed by the research team around themes that this that the conversation would flow according to what the topic
study was trying to elicit (see Table 1 for a summary of the list guide intended for the discussion of that day to be.
of prompts, and Supplemental Appendix 1 for the complete On the first day, the lead facilitator started the ball rolling
topic guide). On days 1 and 5, groups were asked to be online by welcoming all participants to the group. Following a script
at the same time for a synchronous discussion, while par- that was developed by the research team, the lead facilitator
ticipants were free to chat asynchronously from days 2–4 of also reminded the participants of the ground rules of the
the focus group discussion. discussion (see Supplemental Appendix 2) and gave partici-
In each FGD, there was one lead facilitator from the re- pants a two-hour window to introduce themselves. The fa-
search team who guided the flow of conversation throughout cilitator also reminded participants to return to the WhatsApp
the 5 days of discussion. The main role of the facilitator was to chat group at the pre-agreed time for the day’s synchronous
follow the topic guide and ensure a conducive environment for discussion. A similar format was used for the discussion on
participation throughout the duration of the study. Day 5, where the facilitator also reminded participants of the
Each member of the research team took turns to facilitate at specific time for the synchronous group discussion. On days
least one FGD. We realized that the facilitator also required 2–4 when asynchronous discussions were taking place, the
other “softer skills” such as managing the different person- facilitator would type the prompt of the day at 9a.m. and
alities in the group conversation, managing, and defusing participants were told that they could respond anytime be-
heated disagreements among participants, finding ways to get tween 9a.m. and 6p.m. that day.
those who are quieter to speak up and being on the ball to During the focus groups, participants primarily responded
always monitor the chat. Especially for chats that are more to questions and interacted with the facilitator and their fellow
Neo et al. 5

Table 1. Summary of topics for daily discussion.

Day Type of Discussion Topic

Day 1 Synchronous Introductions

Knowledge, attitudes, and perceptions of COVID-19
Day 2 Asynchronous Government sources of COVID-19 information
Day 3 Asynchronous News media as a source of COVID-19 information
Day 4 Asynchronous Fake news and unofficial sources of COVID-19 information
Day 5 Synchronous Pandemic preparedness, social distancing, and panic buying

focus groupmates in text format. However, recognizing the Data analysis

importance of other forms of expression that are not strictly
text-based, the research team took note of other cues shared by All WhatsApp chat messages from the digital FGDs were
participants, including emoji, images (e.g., photos, memes), considered primary data. Chat logs were exported directly
web links, and videos, that provided insights into their sen- from WhatsApp in. txt format using the app’s inbuilt “Export
timents on the topics being discussed. Data” function and included web links, photos, videos, and
any other media that were shared by participants during the
FGDs. Exported WhatsApp chat logs were uploaded into
Consideration of evolving data collection context (software) and thematically analyzed. Analyses drew on both
Notably, the team kept track of Singapore’s evolving COVID- inductive and deductive approaches. As most entries con-
19 landscape throughout the data collection period. In every tained colloquial and abbreviated terms, looking at the data in
focus group discussion, we made sure to incorporate discussion their original form ensured that social and cultural contexts
of recent key pandemic-related events, particularly major policy were maintained (Hymes, 2013). Analyses were conducted
changes (e.g., mandated use of contact tracing apps and masks, iteratively by all members of the research team. Three
physical and social distancing rules, lockdowns, government- members of the team (PHMN, JML, RKJT) are local Sin-
rendered financial assistance). This ensured that the data col- gaporeans and one team member (SEO) is a long-term Sin-
lected were not only reflective of changing pandemic condi- gapore resident, ensuring that interpretation and analysis of the
tions, but also that we were gleaning participants’ responses to, data was grounded in in-depth understanding of the local
attitudes toward, and concerns around these changing pandemic socio-cultural context.
conditions in real time. This included emotional responses to
new policy announcements and perspectives on how new Reflexivity
policies would impact participants’ daily lives.
Table 2 below summarizes the key pandemic-related events As argued by Fox et al. (2007), qualitative researchers who use
that took place over the course of data collection and the novel methodological approaches must engage in reflection
conduct of our focus groups relative to the timeline. and reflexivity to make the research experience transparent
and critically explore the viability of the method being used
(Fox et al., 2007). Reflexivity was built into the data collection
Post-FGD procedures and feedback gathering and analysis process in two main ways: the presence of an
At the end of the discussion on Day 5, participants were given observer and parallel discussion, which functioned as a
instructions on how to collect the food or rideshare vouchers “living” analytical memo to supplement and inform our
as remuneration for participating in the entire duration of the thematic analyses.
study. We also collected feedback via an online survey that Apart from the lead facilitator who would drive individual
sought to understand the feasibility and acceptability of the FGDs, another 2–3 members from the research team were also
WhatsApp platform for conducting research (see included in each FGD to play an observer role. These
Supplemental Appendix 3 for survey questions). Facilitators members were also introduced at the start of the FGD, so that
then proceeded to thank participants for their time, informed all participants are aware that they were a part of the group
participants that they could contact the respective groups’ chat. The main role of the observer was to gather field notes
facilitators should want to share detailed feedback on a one-to- during data collection and to consolidate all these findings in a
one basis, removed all participants from their respective separate “field notes” WhatsApp chat group. This “field notes”
WhatsApp chat groups, and deleted all groups from the app. chat was crucial in allowing the team the space to discuss
Voucher incentives were distributed to all participants on an difficult situations while the actual focus group was happening
individual basis within 2 weeks of the end of the FGDs; concurrently.
participants were given the option to receive these vouchers Observers were also invited to point out biases and
either electronically or via physical mail. bring to the facilitator’s attention areas that may have been
6 International Journal of Qualitative Methods

Table 2. Timeline of key pandemic-related events over the course of the data collection process.

Date Event

20–03-2020 Launch of smartphone app TraceTogether for contact tracing; more events canceled (e.g., IT show, PC show)
21–03-2020 MOM revoked 89 work passes for breaching entry approval and SHN requirements
22–03-2020 Spike in imported cases of COVID-19; ban on all short-term visitors arriving or transiting through Singapore starting 23rd
March
23-03-2020 FGD1A, FGD1B start
24–03-2020 Announced that from 26th March, any resident returning from US or UK is required to serve out their SHN in dedicated hotels
24–03-2020 Returnees to Singapore charged full hospital rates if they left Singapore from 27th March and were admitted for treatment of
COVID-19 within 14 days of their return
24–03-2020 All entertainment outlets, nightclubs, bars, places of worship, attractions and tuition centers to be closed from 26 March, and all
mass events are canceled regardless of size. Groups limited to 10 people at any time. Public spaces required to reduce crowd
density to one person per 16 sq. meters of space. 1m social distancing at all places in Singapore. Measures to last till 30th April
25–03-2020 73 new cases—highest number in Singapore up to that point in time
26–03-2020 New regulation for people who broke social distancing or SHN rules. Punishment include jail terms of up to 6 months, fines of
up to $10,000 or both
26–03-2020 DPM HSK announced second stimulus package, a $48bil Resilience Budget (first package, the $6.4bil Unity Budget, announced
on 18th Feb)
27-03-2020 FGD1A, FGD1B end
28–03-2020 Government issued advice via WhatsApp that people should stay at home and avoid malls with the exception of buying
essentials such as food and groceries
29–03-2020 All long-term pass/long-term visit pass/student pass holders have to get in-principle approval for entry before they arrive in
Singapore; someone’s passport gets canceled for not abiding by SHN rules
30-03-2020 FGD2A start
03–04-2020 Prime Minister’s speech: One-month “circuit breaker” starting 7 April—all preschools and kindergartens to close (with limited
services for children of essential workers who cannot find alternative arrangements), schools and institutes of higher learning
to move to online learning, all who can work from home must do so, essential services (i.e. food outlets, markets,
supermarkets, clinics, hospitals, utilities, transport, key banking services) and economic sectors (strategic or part of global
supply chain) to remain open, all residents advised to stay at home as much as possible + avoid socializing beyond immediate
household members + go out only for essentials (e.g., groceries, buying food home, exercise in parks at safe distance)
03-04-2020 FGD2A end
06-04-2020 FGD2B, FGD3A, FGD 3B start
06–04-2020 Third budget: $5.1bil Solidarity Budget aimed at saving jobs and protecting livelihoods during the 4 weeks when schools and
non-essential businesses have to be shut as part of the “circuit breaker” distancing measures.
In total, all three budgets come to $59.9bil, or about 12% of GDP
07–04-2020 New parliamentary bill: all social gatherings of any size in both private and public spaces have been banned, including private
parties or gatherings with families or friends not living together, at home or in public spaces such as HDB void decks
10-04-2020 FGD2B, FGD3A, FGD 3B end

missed out in the “field notes” chat without directly un- Results
dermining the authority of the lead facilitator. As part of
the research team, observers were also brought in to Engaging participants
mediate any disagreements that occurred between par- The techniques used to engage participants in the FGD process
ticipants, or if any participant complained about receiving are shown in Table 3. All participants were reminded daily
unsolicited text messages. Most importantly, observers about the importance of participation and the ground rules of
helped to capture analytic memos around themes that were engagement via the WhatsApp platform. The daily prompts
being generated during the discussion and highlighted addressed five key topic areas, namely:
main points that the team could revisit during analysis.
Themes that were identified from the memos were dis- 1. Knowledge, perceptions, understanding, attitudes to-
cussed within the research team after each FGD. In our ward COVID-19
analysis, field notes were considered a secondary data 2. News-sourced information
source to acknowledge the reconstruction of meaning and 3. Government-sourced information
context that resulted from the research team’s real-time 4. Lay health beliefs, misinformation, and fake news
discussion process. 5. Outbreak preparedness and hysteria
Neo et al. 7

Table 3. Steps taken to engage digital FGD participants throughout data collection period.

Techniques Examples Effectiveness

Daily group-based welcome messages “Good morning everyone, and hope you had a Participants echoed our ‘Good Morning’
good rest last night! Please be reminded that the messages and that also served the function of
ground rules of ‘Confidentiality, anonymity, signaling the official start of the group chat,
respect, consideration, and safety’ that we which was useful in initiating and facilitating
outlined in yesterday’s chat will also apply today conversation for the day
and throughout the rest of the week.”
(Facilitator from FGD2A—aged 21–30 years
old)
Messages to encourage quieter “What about the rest of you, what do you think? Soliciting responses from the remaining
participants to speak up and share Do you agree/disagree with XXXXX?” participants was effective in getting their
their views (Facilitator from FGD2A—aged 21–30 years feedback on the topic. A more effective
old) “Thanks @XXXXXXXX for these approach was to “tag” using the “@” function
insights! I’d like to ask you now - to what extent specific participants who had been quiet or
do you trust the content of the ST article, and for whom we needed further substantiation
why?” (Facilitator from FGD1A—aged 51 years or response from
old and above)
Thanking participants for their “No worries - please feel free to weigh in on We noted that several groups had continued
participation at the end of every day, anything that we have discussed earlier at any sharing information even after the official
including reminders to participate time! This is a benefit of WhatsApp focus 6p.m. end time for the FGD. Some
the next day groups.” (Facilitator from FGD1B—aged 51 participants who were also busier in the day
years old and above) were able to substantiate some of their
answers and responses further after they had
more time to respond
Feedback channels, including ability to “To ensure the safety of all participants, we would The team received a private message from one
privately message research team like to request that you do not approach or participant who had received an unsolicited
members to share comments and send content to another person without their text message from another participant in the
questions explicit prior consent. Please approach the same group. The facilitators then acted by
group facilitator or moderator should you feel messaging the sender privately to stop such
harassed or unsafe throughout the course of behaviors and informed the receiver to block
the focus group.” (Included at the start of all the sender on his/her end. The facilitators
FGDs) also reminded the sender that if such
behaviors persisted, he/she will be removed
from the FGD entirely and no longer be
permitted to participate

Successful engagement with participants was defined as the Summary of thematic findings
effectiveness of these topics and prompts in eliciting frequent,
insightful, and rich responses from participants, including per- The textual data collected from the FGDs fell into five broad
sonal accounts and experiences, sharing of information sources, themes: perceptions of COVID-19, sources of COVID-19
and encouraging healthy and constructive back-and-forth debate information, determining the trustworthiness of information
and discussion within groups. Groups were also kept open sources, differentiating between news sources, and definitions
throughout the week so participants could continue their dis- of social responsibility in a pandemic. Table 4 provides il-
cussions with one another beyond the scheduled FGD slots; this lustrative quotes for each theme.
was deemed helpful to developing within-group trust and rapport.
Recognizing some participants’ privacy and confidentiality
concerns, throughout the week of FGDs, participants were also
Feasibility and acceptability of the digital FGD format
given the opportunity to send private messages to members of Feedback gathered from the participants via the post-FGD
the research team to ask questions and/or provide feedback that feedback survey affirmed the growing acceptability of holding
they may not have been comfortable sharing within their al- focus group discussions digitally. When asked “how appro-
located groups. At the end of the week, all participants were priate do you think WhatsApp was as a platform to conduct the
thanked by the research team and provided with incentives in focus group discussions?”, 62.1% of all participants indicated
the form of grocery shopping or rideshare vouchers. that it was “very appropriate” and 34.5% of all participants
8 International Journal of Qualitative Methods

Table 4. Key themes and illustrative quotes.

Theme Illustrative quote

Perceptions of COVID-19 “I rely on social distancing, make sure I have enough sleep, work from home, wash hands, etc.; since
nobody seems to provide definite answers [on what works best to prevent transmission].”
(Participant from FGD1A—aged 51 years old and above)
Sources of COVID-19 information “For the now, midnight numbers, ST AND today online, which I seek out. Everything else, frankly,
whatever comes up on Facebook that doesn’t look like a suspicious website, I’d take a look. My
friend group is pretty diverse in viewpoints, so I feel like what I get to read is pretty well rounded”
(Participant from FGD4A—aged 41–50 years old)
Determining trustworthiness of “There was a post online that had a title along the lines that COVID was engineered. I shared it with
information sources my friends with the intent to fact check and FB called out that it was fake. My friends also
commented about the low trustworthiness of the source.” (Participant from FGD3B—aged 31–
40 years old)
Differentiating between news sources “International news tends to be more anecdotal and politically driven. I’ve noticed that depending
on the political leanings of the news outlet, they tend of criticize certain country’s measures
more.” (Participant from FGD2A—aged 21–30 years old)
Definitions of social responsibility “Do our part in following the guideline set out so that we don’t spread the virus, don’t spread fake
uses, use the Trace Together app, stay home, etc.” (Participant from FGD3B—aged 31–40 years
old)

said that it was “appropriate.” Across all age ranges, comfort, Digital data collection also allows the research team to
flexibility, and the lack of time-sensitivity were cited as collect data anytime, anywhere (Colom, 2021; Singer et al.,
reasons behind why participants enjoyed participating in focus 2020). This app-based data collection method was also lo-
group discussions digitally. This is similar to other empirical gistically favorable given the quarantine and lockdown
research on online/digital focus groups suggesting that such measures that have been implemented in many countries
online spaces allowed for greater comfort and inclusive in- during the COVID-19 pandemic. Conducting discussions over
teractions among participants due to the relative anonymity WhatsApp allowed us to generate data in a way that was
and disembodied experience of these online spaces (Colom, rigorous, thoughtful, and perhaps most importantly at this
2021). Table 5 shows examples of participant feedback by time, protective of participants’ health and well-being due to
category. the lack of physical exposure/contact involved. Additionally,
in post-FGD feedback forms, participants also highlighted that
the convenience of participating in a discussion via an app was
Methodological Insights a major reason motivating their willingness to participate in
our study.
Strengths Collecting qualitative data with a widely used (i.e.,
We found that the use of WhatsApp as a means of conducting WhatsApp had over 2 billion users worldwide as of February
focus groups not only served as a means of generating es- 2020), free-to-download, free-to-use app process democra-
sential, time-sensitive data in the community, but also ad- tized and enhanced the participatory nature of FGDs. Barriers
vanced the quality of data generated, vis-a-vis traditional focus to participation, such as the need for participants and facili-
group discussions, for several reasons. tators to download and familiarize themselves with a new and
Like other forms of data collection that adopt digital means, unknown app, were eliminated. Also, participants’ familiarity
collecting data on WhatsApp is easy and low cost compared to with the WhatsApp platform allowed us to consider and fa-
traditional in-person ways of collecting focus group data cilitate the communication of potential issues around data
(Rosales & Fernández-Ardèvol, 2016). In our study, the only privacy, such as the accessibility of participants’ profiles and
costs incurred were printing costs, researcher time, per- other meta-data among other participants, and the participants’
participant cost of participation incentives, and costs of ability to screenshot and share discussions with others.
purchasing mobile SIM cards which the research team used to Compared to traditional focus group discussions, app-
as facilitator accounts from which WhatsApp focus group based focus group discussions allowed for the scope to
discussions were conducted. There were no costs incurred cover a much wider breadth of issues. With the discussion held
from items typical of in-person focus group discussions, such over 5 days at a comfortable pace, the team could explore a
as facility/room rental, participant transportation reimburse- range of at least five different topics with minimal participant
ment, snacks, audio transcription, or refreshments. These cost burnout risk, as compared to exploring the same breadth of
benefits have also been highlighted by other researchers in the issues in a traditional 3–4-hour session. In addition, other
wider literature (Anderson, et al., 2021) empirical research in this area suggests that online/digital
Neo et al. 9

Table 5. Participant feedback on feasibility and acceptability of the digital FGD format.

Category Feedback example

Comfort and “I get to see how others view the same virus issues from the comfort of my smartphone” (Participant from
convenience FGD1A—aged 51 years old and above)
“Flexibility of replying at my time. No need to dress/go out. Very encouraging responses by the moderator which
makes me more willing to share.” (Participant from FGD3B—aged 31–40 years old)
Flexibility “I enjoyed how flexible it was time-wise, when it came to answering the questions, so I was able to participate during
the free times I had from work in the day, rather than specially carving out dedicated time for this purpose. I also
felt that our moderator professionalism in facilitating the discussion was to be commended.” (Participant from
FGD2A—aged 21–30 years old)
Lack of time-sensitivity “Interesting to hear other participants’ opinions on the topic. I can re-read or scroll through all the replies, it is not
time sensitive” (Participant from FGD4A—aged 41–50 years old)

focus groups encourage participants to be more open, be more analyzing FGDs over WhatsApp faced similar challenges as
comfortable candidly sharing potentially sensitive or con- this study when it came to the data collection process (Colom,
troversial information or opinions, and express disagreements 2021; Singer et al., 2020). Against this backdrop, we present
more openly compared to in-person focus groups (Mann & the challenges we faced while conducting this study and re-
Stewart, 2000; Stewart & Williams, 2005). flection points on how we overcame them.
App-based data collection also allowed us to simulta- Part of the digital FGD design was to compare the dif-
neously collect two types of data—textual data from partic- ferences between asynchronous sessions and synchronous
ipant’s typed responses, as well as media data such as images, sessions and assess which was more suitable for the setting of
videos, GIFs, and web links. Participants could also forward an app-based focus group. We found that asynchronous
messages to the group chat that they had received from the sessions were more convenient for most participants across all
family and friends as well. The benefits of this form of data age groups, but conversations on days of asynchronous dis-
collection were manifold. First, we were able to save time and cussions were also more scattered. This is similar to what
money in converting traditional audio recordings into ver- Singer et al. (2020) and Colom (2021) found in their studies
batim transcripts. Next, by encouraging participants to share which spoke about a similar tapering off of participation to-
their thoughts through various means, they were able to ward the end of the FGD and how the elongated timeframe for
“show” instead of “tell” the group what they had meant, a WhatsApp FGD allowed for an increased quality of data
ensuring that there is minimal data loss in translation. Finally, collected. For instance, in some groups, there would be long
the team was also able to concurrently analyze textual data pauses in the middle of the day until the facilitator prompted
alongside media data, resulting in a richer and more well- the participants again for their contribution. Asynchronous
rounded analysis. discussions, however, worked for mothers with young chil-
Methodologically, WhatsApp-based focus groups had also dren who required flexibility and could only respond when
proved to facilitate a more democratic data collection process their children were taking a nap. Asynchronous discussions
as compared to traditional in-person focus groups. The lack of could then consider opinions from this group of participants.
in-person relationship between the facilitator and participants Synchronous sessions however, provided the team with more
in the WhatsApp focus group meant that power structures are targeted discussions. Participants mostly stayed on point and
not as concretely put into place. In WhatsApp focus groups, answered the required prompts.
participants have as much power to speak up as facilitators and Compared to traditional focus groups, it was not possible
have as much control to shift the flow of conversation. for the facilitator to read and access a participant’s body
language or tone of voice in app-based focus group discus-
sions. Without these non-verbal cues, it might be hard for the
Challenges
facilitator to pick up certain nuances in a participant’s response
While using WhatsApp as a means of conducting focus groups and increase the chances of misunderstanding among par-
had many strengths especially during a pandemic, there were ticipants. The lack of such bodily cues also means that fa-
also challenges that the team had to tackle and learn from cilitators sometimes find it hard to nudge quieter participants
while concurrently facilitating the various WhatsApp groups. to respond, or to encourage more dominant participants to
At the time of our study conception and conduct, to our best share the space.
understanding and knowledge of the current published liter- When a participant is not responding on an app-based chat,
ature, there were no other research studies using WhatsApp to the facilitator does not know whether it is because the par-
collect qualitative data as part of exploring the impacts of the ticipant is facing any difficulties, is disengaged by the con-
COVID-19 pandemic as it unfolded. In fact, other studies versation or simply did not have the time at that moment to
10 International Journal of Qualitative Methods

Table 6. Summary of modifications made.

Aspect First digital FGD study Second digital FGD study

Recruitment Facebook, personal networks Facebook, Telegram; moving beyond just personal
networks
Conduct of digital FGDs Combination of synchronous and asynchronous Asynchronous
Interval of asynchronous Asynchronous discussions were conducted from Discussions will start later at 11a.m., and will end at
discussions 9a.m. to 6p.m. 8p.m.

participate in the discussion due to other commitments. Our conversation also moved faster in FGDs that involved par-
team noted that in the eight FGDs that we had conducted, there ticipants of an older age range. One facilitator noted that if she
were two groups that faced inactive participants. The team looks away for 10 minutes, the FGDs that involved the older
then had to come together in the separate “field notes” chat we age group could have 60 unread messages, compared to the
had created for each FGD to discuss how we could better FGD with younger participants that might have only eight
engage these inactive participants—which included messag- unread messages. Our team learned the importance of alter-
ing these participants privately to check if they were facing nating between rapid and slow paces of discussions. We
any difficulties, or in some cases, to signal them using the @ discussed the importance of engaging participants so that there
function in the WhatsApp chat and specifically asking for their would be no long silent pauses, especially during asynchro-
response. nous discussion days, and knowing when to step in such that
Also, we found that our most enthusiastic and perhaps most no one dominated the chat for too long. We also found that
dominant participants tended to respond to facilitator ques- older participants used emojis a lot more as compared to
tions more quickly than the rest of their group and tended to younger participants who used gifs or animated stickers to
respond to facilitator questions with lengthy text responses or express themselves instead. At the end of the discussion, one
bursts of consecutive short messages, which could be per- participant from FGD2A (from 21 to 30 years age range) gave
ceived as “crowding out” the views of their discussion feedback that “I would say if it was on Telegram, it might have
groupmates. In these situations, facilitators would thank the been better, because Telegram allows you to pin messages, so
dominant participants for sharing their experiences and re- we don’t have to scroll to find the questions in the flood of
direct the flow of discussion toward other participants who had messages. Other than that, I think the experience was pretty
not yet had a chance to share their views. Where needed, smooth).” Such comments also suggest the possibility of
facilitators would also message these dominant participants conducting FGDs over other app-based chat platforms with
privately to gently encourage them to allow other participants more features compared to WhatsApp that younger partici-
to share their views. pants might be more familiar or comfortable with.
Additionally, we noted that despite having distinct discus-
sion topics on each day of discussion, different days’ themes
Ways forward
sometimes organically overlapped with each other as partici-
pants shared their views. This has also been highlighted as a At time of writing, the research team is preparing to conduct a
challenge in other scholarship in this area (Anderson, et al., new digital FGD study to explore the Singapore public’s
2021). To manage this, the facilitators made deliberate efforts to experiences and understandings of the COVID-19 pandemic
ensure that each day’s discussion guide was adhered to as over a year in, with a focus on three thematic areas: 1)
closely as possible and gently signaled to participants during healthcare and information-seeking behaviors throughout the
discussions to let them know that the topic they were foraying pandemic, 2) experiences with work and day-to-day activities
into would be discussed in greater detail at another session. during the pandemic, and 3) thoughts and perspectives on
Methodologically, our team also realized that when fa- maintaining social interactions and mental/emotional well-
cilitating such focus groups, different methods are needed to being throughout the pandemic.
engage different age groups and continue establishing and Taking into consideration the lessons learnt and participant
maintaining participant rapport online. What worked well for feedback from our maiden experience conducting digital FGDs
one group did not necessarily work the same for another age during the COVID-19 pandemic, we have made some modi-
group. For instance, we noticed that older participants tended fications to the way we plan to carry out our next round of
to talk in colloquial language a lot more than younger par- FGDs. A summary of these changes is shown in Table 6 below.
ticipants, who typed in fuller sentences. This could be due to
younger participants being on WhatsApp using the WhatsApp
Lessons for future research
Web function where they could sync their phones to their
laptops, allowing them to type in longer sentences. Older Our experience with digital FGDs using the free-to-use, free-
participants also stayed online a bit more and hence the flow of to-download mobile communication platform WhatsApp
Neo et al. 11

offers several lessons for those considering this approach. publication of this study was generously funded by the Saw Swee
Firstly, that WhatsApp-based qualitative data gathering is of Hock School of Public Health, National University of Singapore and
high utility during a pandemic situation, where social/physical the National University Health System.
distancing and safety of participants and the research team are
paramount. Secondly, that there is potential for applicability of ORCID iD
the conduct of digital FGDs in other contexts to Singapore Pearlyn HM Neo  https://orcid.org/0000-0002-7797-8942
which report similarly high levels of mobile penetration and
mobile communication platform use. Third, that future re- Supplemental material
search in this area should recognize differences in use patterns,
Supplemental material for this article is available online.
openness and transparent sharing, and utility of digital
communication platforms between age groups, depending on
technology familiarity, uptake, and adoption levels. Fourth, References
future research in this area should also reflect on the extent to Anderson, E, Koss, M, Castro Luque, AL, Garcia, D, Lopez, E, &
which online/digital spaces are truly disembodied and in what Ernst, K (2021). WhatsApp-Based Focus Groups Among
ways this ambiguous space impacts data quality and rigor, Mexican-Origin Women in Zika Risk Area: Feasibility, Ac-
especially when compared to traditional in-person qualitative ceptability, and Data Quality. JMIR Formative Research, 5(10),
data collection. While some researchers might point out the e20970. https://doi.org/10.2196/20970.
disembodied nature of online interactions (Dreyfus, 2013; Chen, J., & Neo, P. (2019). Texting the waters: An assessment of
Marin, 2022), others argue that online spaces are not dis- focus groups conducted via the WhatsApp smartphone mes-
embodied as they continue to be spaces where emotions, saging application. Methodological Innovations, 12(3). https://
social action, and experiences of togetherness and community doi.org/10.1177/2059799119884276.
are experienced (Coffey & Kanai, 2021; Osler, 2020). Finally, Coffey, J., & Kanai, A. (2021). Feminist fire: Embodiment and affect in
there is a need to strike a balance between the need to elicit managing conflict in digital feminist spaces. Feminist media studies
high-quality data and placing potential stress on participants’ (pp. 1–18). https://doi.org/10.1080/14680777.2021.1986095.
time, schedules, and well-being, especially during a pandemic Colom, A. (2021). Using WhatsApp for focus group discussions:
where mental health concerns are paramount. Ecological validity, inclusion and deliberation. Qualitative
Research. https://doi.org/10.1177/1468794120986074.
Dreyfus, H. L. (2013). On the internet. Routledge.
Conclusion Fox, FE, Morris, M, & Rumsey, N (2007). Doing Synchronous
Digital focus group discussions via a mobile communication Online Focus Groups With Young People: Methodological
platform are a promising, flexible, and adaptable means of Reflections. Qualitative Health Research, 17(4), 539–547.
collecting qualitative data from a diverse range of respondents https://doi.org/10.1177/1049732306298754.
sampled from the public. The ability to maintain frequent and Gibson, K. (2020). Bridging the digital divide: Reflections on using
convenient two-way interaction with participants and re- WhatsApp instant messenger interviews in youth research.
searchers, facilitated using WhatsApp, a free-to-download, Qualitative Research in Psychology. https://doi.org/10.1080/
free-to-use, widely used chat app in the Singapore context, 14780887.2020.1751902.
offers the potential to safely collect timely and relevant data Hymes, D. (2013). Foundations in sociolinguistics: An ethnographic
during a pandemic, during which social and physical dis- approach. Routledge.
tancing are crucial. Additionally, this study supports findings Mann, C., & Stewart, F. (2000). Internet communication and
in the wider literature that using WhatsApp for the conduct of qualitative research: A handbook for researching online. Sage.
focus groups is effective, efficient, increasingly ubiquitous, Marin, L. (2022). Enactive Principles for the Ethics of User Inter-
and highly accepted by study populations. Although privacy actions on Social Media: How to Overcome Systematic Mis-
and confidentiality concerns remain, it is imperative that such understandings Through Shared Meaning-Making. Topoi,
digital solutions for qualitative data collection are further 41(2), 425–437. https://doi.org/10.1007/s11245-021-09792-9.
explored in the context of other health-related research MCI’s response to PQ on impact of change in WhatsApp’s privacy
questions beyond a public health emergency context. policy on Government communications (2021). Ministry of
Communications and Information. Parliament Sitting on 1
Declaration of conflicting interests February 2021 (2021) (testimony of S. Iswaran) https://www.
mci.gov.sg/pressroom/news-and-stories/pressroom/2021/2/
The author(s) declared no potential conflicts of interest with respect to
mci-response-to-pq-on-impact-of-change-in-whatsapp-privacy-
the research, authorship, and/or publication of this article.
policy-on-govt-comms.
Osler, L. (2020). Feeling togetherness online: A phenomenological
Funding sketch of online communal experiences. Phenomenology and
The author(s) disclosed receipt of the following financial support for the Cognitive Sciences, 19(3), 569–588. https://doi.org/10.1007/
the research, authorship, and/or publication of this article: The s11097-019-09627-4.
12 International Journal of Qualitative Methods

Reid, D. J., & Reid, F. J. M. (2005). Online focus groups: An in- Stewart, K., & Williams, M. (2005). Researching online populations: The
depth comparison of computer-mediated and conventional use of online focus groups for social research. Qualitative Research,
focus group discussions. International Journal of Market 5(4), 395–416. https://doi.org/10.1177/1468794105056916.
Research, 47(2), 131–162. https://doi.org/10.1177/ Thunberg, S., & Arnell, L. (2021). Pioneering the use of technologies
147078530504700204. in qualitative research – A research review of the use of digital
Rosales, A., & Fernández-Ardèvol, M. (2016). Beyond WhatsApp: interviews. International Journal of Social Research Method-
Older people and smartphones. Romanian Journal of Com- ology, 1-12. https://doi.org/10.1080/13645579.2021.1935565.
munication and Public Relations, 18(1), 27. https://doi.org/10. Vindrola-Padros, C, Chisnall, G, Cooper, S, Dowrick, A, Djellouli,
21018/rjcpr.2016.1.200. N, Symmons, SM, Martin, S, Singleton, G, Vanderslott, S, Vera,
Rosenfeld, A., Sina, S., Sarne, D., Avidov, O., & Kraus, S. (2018). N, & Johnson, GA (2020). Carrying Out Rapid Qualitative
WhatsApp usage patterns and prediction of demographic char- Research During a Pandemic: Emerging Lessons From COVID-
acteristics without access to message content. Demographic 19. Qualitative Health Research, 30(14), 2192–2204. https://
Research, 39, 647–670. https://doi.org/10.4054/demres.2018.39. doi.org/10.1177/1049732320951526.
22. We are Social (2021). Digital 2021 Singapore. https://wearesocial.
Singer, B., Walsh, C. M., Gondwe, L., Reynolds, K., Lawrence, E., & com/sg/digital-2021-singapore.
Kasiya, A. (2020). WhatsApp as a medium to collect qualitative Wong, C. M. L., & Jensen, O. (2020). The paradox of trust: Perceived
data among adolescents: Lessons learned and considerations for risk and public compliance during the COVID-19 pandemic in
future use. Gates Open Research, 4, 130. https://doi.org/10. Singapore. Journal of Risk Research, 23(7–8), 1021–1030.
12688/gatesopenres.13169.1. https://doi.org/10.1080/13669877.2020.1756386.