‘Talk amongst yourselves’: designing and evaluating a novel remotely-moderated focus group methodology for exploring group talk

ABSTRACT The use of a moderator has become ubiquitous when using focus groups for social science research. While a skilled moderator can facilitate discussion, we argue that, in some instances, moderators can potentially hinder the generation of the types of group discussions that academic researchers may seek to access. In this paper we outline some of the challenges associated with moderated focus groups and propose a complimentary methodology: a remotely-moderated focus group that can help overcome some of the problems a physically present moderator might create, while still incorporating many of the benefits of moderation. Using two remotely-moderated focus group designs – one exploring dietary identity and the other exploring gendered experiences of sexual harassment – we provide evidence for the efficacy of this design in multiple contexts. We evaluate its ability to produce high quality conversational data and suggest directions for future research exploring the utility of this methodology.

. Since the 1940s, researchers inside and outside of academia have adopted focus groups as the primary methodological means of examining group discussions (Delli Carpini and Williams 1994;Liamputtong 2011;O.;Nyumba et al. 2018), with the focus group described by some as an 'established part of the methodological tool kit' for (qualitative) researchers (Barbour and Kitzinger 1998, 11).Traditionally, in a focus group, participants (typically between 5-12 people depending on the discipline of the research) are brought together to discuss a topic that is of interest to the researcher (Stewart and Shamdasani 2014).Focus groups may take place online or in-person, according to research circumstances and participant needs (Braun, Clarke, and Gray 2017;Watson, Peacock, and Jones 2006).A meta-analysis of the use of focus groups stresses that there are three primary elements to a focus group: the individual, the group, and the interaction (Cyr 2016).Cyr (2016) argues that the unique advantage of focus groups lies in the capacity for in-depth exploration of both the group, and the interactions had within that group.Often, participants are recruited such that they share some attitude or experience-based identity related to the project's research question, in the hope that this will allow the group to talk about an issue respectfully as equals (Acocella 2012;Morgan 1992).This group discussion is typically facilitated by a 'moderator': either the researcher themselves or a person aligned with the researcher who sits with the participants and is charged with asking the group of participants questions, probing answers, and facilitating discussion between group members (Gill et al. 2008;Grønkjaer et al. 2011).Moderators may also assist with the practical elements of the research process, for example ensuring that the group is recorded, gaining informed consent from participants, or ensuring that all participants are treated respectfully (Pickering and Watts 2013).
The moderator has long been considered fundamental to the effectiveness of focus group methodologies.However, in some early applications within research settings, focus groups involved a mixture of both moderated and unmoderated approaches, with the latter involving participants being encouraged to talk about a topic amongst themselves with no moderator present (Fern 1982;Itua et al. 2014;Morgan and Spanish 1984).The centring of the (physically present) moderator in more recent use of focus groups likely stems from the increasing popularity of focus groups within market research and business since the 1980s, where there is a need for groups to respond to very specific questions, and for groups to be more structured and directed in nature (McDonald 1993;Welch 1985).Some research argues that social scientists should 'break away' from norms surrounding focus group methods stemming from market research (Munday 2006).It is argued that direct moderation practices commonplace in market research may not be conducive for study in the social sciences where a moderator is typically used to 'facilitate' discussion rather than to direct it, and to encourage open and honest discussion among group members (Lezaun 2007).Even when adopting a more facilitative role, there may still be circumstances where the presence of a moderator in the room with participants is at odds with, or poses problems for, the specific research aims or the group discussion process.
In this paper, we wish to question the norms surrounding in situ focus group moderation as the default choice and understand how and under what circumstances it might be possible, and indeed even potentially desirable, to physically remove a focus group moderator from the room entirely.Recognising the value of focus group designs in general, we then outline and empirically explore the research utility of an alternative, 'remotely moderated', focus group methodology, which we believe could add another potential tool to the methodological toolkit of qualitative researchers.

What does a moderator offer?
In this section, we evaluate the normative roles of a moderator within focus groups and the challenges of moderation in different contexts.Throughout these sections we also consider whether and how researchers might be able to achieve the aims of a moderated focus group without the moderator present.

Successfully "managing" and "controlling" the discussion
One role of the physically and conversationally present moderator is to ensure the focus group discussion is 'successful'.A successful focus group has been conceptualised as 'a carefully planned series of discussions designed to obtain perceptions of a defined area of interest in a permissive, non-threatening environment' (Krueger and Casey 2014, 2).A successful focus group might be one where all participants are encouraged (and get the chance) to share stories and 'open up' to the group, including quieter participants (Cameron et al. 2005;Wong 2008).Morgan (1993) have stressed the vital role of the moderator in ensuring the success of a focus group methodology, and note that 'if the moderator, as the data-collection instrument, is not prepared, not attentive or not skilful, then the results will be just as bad as in a poorly prepared survey questionnaire' (p.6 (Morgan 1993)., Morrison-Beedy, Côté-Arsenault, and Feinstein (2001) make a similar claim, asserting that 'only a skilled moderator can provide the structure, stability, and freedom of expression that is required for successful discussions' (p.175) (Morrison-Beedy, Côté-Arsenault, and Feinstein 2001).
Much writing on focus group methodologies places a high responsibility on the moderator for ensuring the success of the focus group, both as an interaction and as a data-generating method.Agar and MacDonald (1995) argue that "it is not automatic that a group of strangers will have a 'lively conversation' about anything'" (p.78), and Morrison (1998) argues that the focus group is a wholly artificial environment where all discussion ultimately hinges on the moderator (Agar and MacDonald 1995;Morrison 1998).However, as much evidence from conversation analytic work demonstrates, as culturally competent members of a speech community, people are usually highly capable of following the 'rules' of conversation in their everyday talk (Goodwin and Heritage 1990;Kitzinger 2000;Sacks, Schegloff, and Jefferson 1978).Many norms relating to what is considered polite conversation are also prominent in a focus group.Indeed, many of the attributes that make a 'moderator' a successful discussion manager are familiar to us in the social rules and cues that lead to meaningful group conversations in our everyday lives: ensuring that everyone has a chance to speak, that everyone is respected and that no one person dominates (Onwuegbuzie et al. 2009).Which raises the question of why, as qualitative researchers, might we assume that a group of research participants is not capable of managing an amicable discussion without one of us being present in the room?While a moderator may be explicitly trained in these conversational skills, the assumption that participants do not have any conversational skills of their own may place undue pressure on the moderator to perform a task that may, in many cases, be just as successfully performed by participants themselves.
There is arguably also a risk of moderators 'over-managing' focus group discussions.In comparison to everyday conversations, where all participants share equal responsibility for the success of a social interaction (Gibson 2003), a moderated focus group may come to depend on the moderator as the primary force of the discussion, rather than them merely being a facilitating presence (Ryan et al. 2015).This can be a particular concern in more unstructured designs where free-flowing conversation is a priority; in contexts where resources or discussion time is limited; or where the moderator is not highly experienced in facilitating focus groups (Makosky Daley et al. 2010).Indeed, it is often noted that the quintessential 'sub-optimal' (moderated) focus group is one where the moderator asks questions and group members then simply answer in turn (Morrison-Beedy, Côté-Arsenault, and Feinstein 2001), thus rendering it somewhat akin to a 'group interview' (Gibbs 2012).While this might not be an issue for market (or other highly structured) research, where the focus group is a way of gaining multiple perspectives efficiently (Kidd and Parshall 2000), it could become a problem in more social scientific (especially social constructionist approaches) where the primary focus of the research is on the group's interaction (Freeman 2006).In this example, the moderator may risk becoming the sole mediator for the group's interaction, 'buffering' participants from speaking directly with each other (Gill et al. 2008), and preventing a group discussion from unfurling organically.This is particularly important if one of the researcher's aims is to study the very nature of that unfurling.A skilled moderator might be able to bridge participants' points and encourage discussion among the group, but this may not work (to the advantage of the research) if the group comes to depend too much on the moderator throughout the focus group (Fern and Fern 2001).Thus, the moderator's responsibility for the 'success' of the focus group interaction might actually change the dynamic and the talk generated in the group.
Issues of potential moderator 'interference' in the processes of interaction between participants aside, another key benefit of a moderated focus group is its ability for some control over the topic (or 'focus') of conversation, which can lead to fruitful insights to a research question and an analysis of the interactions between participants in a controlled environment (Smithson 2000).There is a potential for participants left completely alone to choose not to discuss the topic at hand, or discuss something else that is irrelevant, and this is a concern of many researchers (Krueger 1997).There are also ethical concerns to consider in so much as the moderator can play an important role in ensuring participants can provide informed consent to participate (Tolich 2009) and feel safe and able to be vulnerable and express their experiences (Sim and Waterfield 2019;Smith 1995).Therefore, when using focus groups, there is a need for a fine balance between over-management and under-management of the group.One must tread a line between a highly constrained conversation that may not be fully reflective of what participants want to converse about (and how), and the risk that the conversation veers off topic and fails to address the research question, or risks crossing ethical boundaries.Presently, available focus group methodologies offer the option of a (physically present) moderator, and no alternative.There is not yet a middle ground for those seeking to reduce the 'over management' of focus group discussions, while also allowing for some control that ensures participants are safe and can address the research question.
A common response to this dilemma -particularly from those interested in how a topic is discussed -is to discount the use of focus groups entirely and advocate for the use of 'naturally occurring data' (Potter 2002), such as talk generated on social network sites (e.g.Twitter and Facebook) or conversations originally recorded for another purpose (e.g., calls to customer complaint lines) (Kiyimba, Lester, and O'Reilly 2019;Potter and Hepburn 2007).Such data may sometimes also have ethical and practical benefits, particularly when studying 'seldom heard' groups (e.g., Drewett and O'Reilly 2021).However, for other research topics, such naturally occurring data can lack depth of explanation in the accounts captured, can suffer from a lack of the desired level of topical focus, and can pose ethical issues such as consent and confidentiality (Stommel and de Rijk 2021).It is also, by its nature, far more variable and uncertain (Lester, Muskett, and O'Reilly 2017).Access to naturally occurring data on the topic of interest may also be difficult to access and permission to download social media content could be revoked at any time (Hogan 2016).Smithson (2000) notes that 'focus groups . . .should not be analysed as if they are naturally occurring discussions, but as discussions occurring in a specific, controlled setting' (p.105).She argues that the level of control in focus groups allows for the examination of specific research questions and interactions between participants, which is advantageous for many research perspectives.Thus, there is clear utility in having a methodology available that allows a researcher to actively generate conversational data between participants on a specific topic of their choosing.The question we raise here, however, is whether such a methodology must necessarily involve the researcher being physically present as a participant in the conversation to successfully achieve this, especially given some of the challenges that a conversationally present moderator can involve.Thus, an alternative method that incorporates the benefits of moderated focus groups, while minimising the potential disadvantages of the method that we highlight above, may help to maximise the accessibility and utility of focus group methods.This is particularly so in research contexts where a researcher may have concerns about the effect that a physically present moderator might have on the nature of the talk generated, and thus, its ability to answer their research question.

When might a moderator facilitate vs compromise the accessing of rich accounts within (minority) group talk?
In some situations, the presence of a moderator can pose problems for accessing the object of observation that the focus group intends to explore.This may be especially so where members of the focus group belong to a societal minority (e.g., vegans, minority ethnic groups, sexual and gender minorities, climate sceptics) and perceive either other focus group members or, as is often the case, the moderator, as outgroup members.Members of minority groups are frequently hypervigilant to how they are perceived by majority group members and are skilled at adapting their behaviour in daily interactions to reduce socalled 'social-interactional trouble' (Fallon and Brown 2002).Examples include code switching (Lo 1999) and techniques, such as 'face-saving', changing the subject, or simply withholding their perspectives from others (Barr and Chapman 2002;Derous 2017;Singh, Kumra, and Vinnicombe 2002).Thus, among homogenous identity groups who share a lived experience that is also shared by the moderator, discussion may flow more easily; people are more likely to be honest about their opinions and views and take more 'social risks' in their talk when they are surrounded by other people who they believe will 'understand' them (Finch and Lewis 2003;Madriz 1998).For example, Greenebaum (2012) talks extensively about how her own identity as a vegan allowed her to successfully network with, and develop rapport with, fellow vegans in her own qualitative research.However, when an identity (and associated lived experience) is not shared by a moderator, even the most talented of moderators may struggle to effectively mitigate these issues, and groups may orient their talk towards this 'outsider moderator' (for example in Allen's (2005) focus group study on masculinity).Indeed, literature on how asymmetries of power impact perspective taking suggests moderators may struggle to recognise how they and other majority group members are seen and so may miss when there is a problem (Ayrton 2019;Lammers et al. 2008).
The nature and composition of the groups studied, and their perception of the moderator's own identity group can, thus, be a facilitator or barrier to open group discussion.One solution to this problem of 'outgroup' moderators might be to always recruit moderators who are 'insiders'; who share identity on a dimension that is important to the focus group (e.g., ethnic identity in research on racial profiling in policing).Lived experience and own-voices researchers can benefit a research team dramatically at all points of the research process, including during focus group moderation, by providing an 'insider's perspective' on the work (Makosky Daley et al. 2010;Tracy 2010).Increasingly too, groups who have been marginalised have been calling for greater involvement in the production of knowledge -for instance, in an Australian context there is now an expectation that Aboriginal and/or Torres Strait Islander researchers lead, or are directly engaged with in, research with Aboriginal and/or Torres Strait Islander communities (AIATSIS 2022).However, the inclusion of lived experience researchers in moderation is sometimes not practical and can also be problematic.There may not be someone who is trained in the specific moderation skills required who also belongs to the group being studied.Placing the burden of group-based focus group research on members of said group can pose practical and political issues, especially when groups of interest are also underrepresented within academia (Lathen and Laestadius 2021).Additionally, projects may wish to simultaneously examine more than one identity group in a comparative sense (as one of the projects detailed in this paper did), and it may be difficult for a group of researchers to represent all the identities they would like to study.While it is undoubtedly advisable to include lived experienced researchers in a research project, particularly during the design and analysis stages (Newman et al. 2011), the data generation process of moderating focus groups is a large time commitment and requires skills that some researchers may not possess regardless of their group identity.
The issue of how moderators might hamper the generation of high-quality ingroup talk is complicated further still by the conversion orientation of some groups.Groups surrounding specific practices (Kurz et al. 2020), or opinions (McGarty et al. 2009) often mobilise around a 'conversionary collective action' goal (Wright 2009), wherein group members ultimately desire the widespread adoption of their identity-associated behaviour or opinion among the wider population.Thus, any participant in the discussion who is not already an ingroup member (such as a moderator) could be viewed by these groups as a prospective convert to the identity (Klandermans 2020).Therefore, moderator-oriented designs in this domain, such as focus groups or interviews, may increase prevalence of researcher-oriented responses within participants as they strive to make their identity and associated practices seem appetising to a currently non-identified moderator (Sneijder and Te Molder 2009).In essence, groups with activist orientations may be likely to see the research environment itself as an opportunity to convert observers to the cause.If a researcher is interested in this conversational process as their object of investigation, then clearly this is not a problem.However, this may also lead to exclusion of key stories or accounts in participant talk (e.g.Cherry 2015), which in some cases may be the very object of analysis that a researcher is interested in exploring.

How might groups function without a moderator being conversationally present?
It is clear that using a physically/conversationally present moderator, for some research designs, presents many methodological and theoretical issues.However, exploring the boundaries and bindings within and between different groups is extremely valuable in understanding their impact on social (inter) action within society.We argue that creative methodological innovation is needed to overcome these difficulties while still maintaining the benefits of focus groups as a methodology.
One tactic for researchers who are concerned with these issues might be to physically remove the moderator from the site of the focus group discussion altogether, opting for an entirely 'unmoderated' focus group.Indeed, focus group work that has considered unmoderated designs suggests reasons for optimism in this regard.Fern (1982) argued that there was no observable difference between the number of ideas generated in a moderated compared to an unmoderated focus group.More recently, Canipe (2020) proposed and tested an unmoderated focus group design with three elementary school teachers, demonstrating that a focus group can work without a moderator when exploring homogenous group identities.Canipe's (2020) research demonstrates that an unmoderated design could lead to rich generation of identity stories, as participants feel they are understood by other participants and may be more keen to open up in an unmoderated group discussion.However, fully unmoderated designs likely suit research contexts where flexibility and lack of structure is desirable, as there is no way to anticipate or control the data that the design will produce, and this could risk research questions not being addressed or complex questions being skipped over.While sacrificing control over the group in this way might be a good option for researchers with more flexible or broader research questions, for many researchers, maintaining some control over the group, but without the physical and conversational presence of a moderator in the room, might be seen as more desirable.
We aim to fill this methodological gap between fully moderated and unmoderated focus groups in this paper by developing and testing a novel 'remotely-moderated' focus group method where the focus group guide (question schedule) is moved through remotely by a moderator watching the group without being present in the room.We outline below the development of this method and provide evidence that these remotely-moderated focus groups effectively mitigate many methodological challenges that can be inherent in studying group interaction processes, while allowing for some of the benefits of moderation and control.In the contexts of meat-reducing dietary identities (i.e., vegan, vegetarian, and flexitarian) and gendered understandings of sexual harassment, we demonstrate how this novel methodology can lead to excellent data quality without sacrificing participant rapport or experience.

Research questions and aims
The goal of the remotely-moderated focus group paradigm is to provide a space for participants who share a topic-relevant identity to talk, while reducing the social-interactional difficulties a moderator might introduce in certain research contexts.Our intention was that it would allow participants to discuss, argue, and explore the bounds of what are normatively acceptable opinions and behaviours associated with their ingroup membership, as well as how they represent themselves to outgroups (Hogg and Rinella 2018).
We tested this remotely-moderated focus group design in two contexts where the presence of an in-situ moderator presented some of the problems discussed in the introduction.In the first design, we assembled (identityhomogenous) groups of vegans, vegetarians, and flexitarians to discuss their dietary identities (Study 1).In the second context, single-gender groups of either men or women were asked to discuss their attitudes towards sexual harassment (Study 2).We evaluated the success of the research method using multiple approaches that focused on three research questions: (1) Could participants manage the discussion effectively without a moderator in the room?(2) Did the remotely-moderated focus groups allow for the generation of group-level talk that was rich and in-depth enough for qualitative analysis?(3) Was the remotely-moderated group a positive experience for participants?
To explore our first research question, we performed a line-by-line discursive analysis in the tradition of discursive psychology (Wiggins 2016) of transcriptions of the group discussions, exploring the conversational patterns and techniques of the group, with particular attention to how (if at all) they compensated for the lack of a moderator in their discussions.To examine our second research question, we provide descriptions of the forms of data generated in each group to provide readers with insights regarding the data quality (with such quality ultimately being determined by its ability to explore relevant research questions through further substantive analysis using thematic analysis and discursive psychology).Our third research question is assessed through a descriptive reporting of qualitative and quantitative feedback that the participants themselves offered following the completion of the focus group to demonstrate the experience participants had in the group discussions.

Participants
Participants were recruited from a convenience sample, using flyer and social media advertisements in university social societies.There was a total of 40 participants in the dietary ingroup focus groups, including 11 vegans (people who do not consume meat or animal products), 13 vegetarians (people who do not eat meat) and 16 flexitarians (people who are actively reducing their meat consumption).There were 18 participants in the ingroup sexual harassment focus groups (half female/male), with between three and five participants in each focus group (see Table 1 for more details on the demographic composition of all focus groups).In total, 15 focus groups were run using the new method.All participants were either undergraduate or postgraduate students, or university staff.The gender imbalance in the vegetarian and vegan groups reflects the underrepresentation of men in these dietary identity groups more widely (Heinz and Lee 1998;Rothgerber 2013;Sobal 2005).Both studies were approved by the University of Bath Psychology Research Ethics Committee (reference number: 19-024).

Focus group procedure
To allow for close and thorough piloting of the method and observation of participants, the focus groups were conducted in a room equipped with live video and audio recording, provided via four video cameras mounted near the ceiling in the four corners of the room and a ceiling-mounted microphone (see Figure 1 for a photo of the focus group room and setup, taken via one of video recording cameras used).This was to ensure we could observe the group fully, although, at a minimum, our method could be used in any room with (mobile) audio recording available.In addition, a high-resolution 24-bit mobile digital audio recorder was placed on the table, which was used to capture the audio track used for transcription.The location chosen was also convenient for students and staff to attend during their lunch breaks.From a nearby, but concealed, space, the remote moderator (the first and second authors for the dietary and sexual harassment groups respectively) controlled a power point  slide presentation with a question/prompt on each slide that was displayed on the screen in the discussion room.This was designed to provide some structure to the focus groups and encourage discussion relevant to the topic.The remote moderators also watched/listened to the group in real-time via the audio-visual feed provided from the ceiling cameras/microphone to ascertain the optimal moment to advance slides.Participants were not told explicitly beforehand that the group would be remotely-moderated.They were, however, told that they would be engaging in a 'group discussion' with other group members about either 'diet and food' or 'sexual harassment'.We ensured that participants were aware of the discussion topic before they signed up to attend, and that they had the chance to give informed consent and ask questions before the group began.We did not give them advance notice of the remote moderation to preserve the novelty of the design, and pilot how participants might respond.We also wanted to ensure that participants were focused on the study topic and not the methodological quirks of the study, so we intentionally did not draw attention to the (lack of) moderator.
After informed consent was provided, participants were seated around a table with the mobile digital audio recorder in the centre.Participants were also not told of the role of the remote moderator in controlling the presentation of the questions; rather, they were told that the slides would advance 'automatically every five minutes or so'.This was to keep participants focussed on the discussion and avoid speculation about the location of the remote moderator.Participants were told before participating that the group was being recorded (and gave consent to this), but they were not told that they were being watched in real-time by a researcher.After the group, they were fully debriefed about the design, including this mild deception.These deceptions were intended to help us examine how participants would respond to this design without much prior information, and allow for a thorough pilot of the methodology.However deceiving participants in this way may not be necessary for future projects and any deception used in this design should consider ethical risks and potential harm to participants.
Due to the novelty of the design, participants were first given the opportunity to 'try out' the procedure.To start the group, the remote moderator left the room and advanced the screen display to an 'introduce yourselves" slide.Once participants had introduced themselves to each other, the remote moderator re-entered the room to answer any questions and establish ground rules for the group: that all participants should be respected and that anything said within the group was not to be shared outside of the group.At this point, the remote moderator also told the group that they would be waiting outside, and how they should contact them in case of an emergency (e.g., an uncomfortable conflict among the group or a medical issue), or in the event that they wanted to immediately withdraw from the focus group.After explaining this, the remote moderator then left, started the PowerPoint with the focus group schedule, and the session commenced.
The focus groups lasted between 50-65 minutes.This timing was decided on following piloting to ensure enough time for the participants to get to know each other, develop rapport, and discuss all of the questions in our guides.This was also the maximum length of time we could keep participants due to funding constraints in the dietary identity study and voluntary time constraints in the sexual harassment study.We recommend that researchers hoping to use this design in future pilot their groups to ensure there is enough time to cover the material desired because, in our experience, remotely moderated groups may take longer than expected.Over the course of the focus group, participants were presented with a number of questions.The researcher aimed to provide about 5 minutes discussion time for each question but had the flexibility to advance the slides sooner or later depending on the flow of the discussion.Small silences were ignored by the researcher, as we quickly realised that some silences led to subsequent interesting discussion.However, the slide was advanced for longer periods of awkward silence (over 60 seconds) to ensure group rapport did not suffer.The remote moderator monitored discussion for signs of potential distress and was ready to step in to stop the group in the event of any issues or emergencies.A second researcher was always available to help set up or address any issues.
We also adjusted the basic remotely-moderated design to suit the research topic of each group.Participants in the dietary focus groups were given lunch (that they had selected in advance of the group) to eat whilst talking.This was done to highlight the topic of diet in a casual fashion, as well as to cater for participant's chosen diet.In the sexual harassment focus groups participants responded to questions about online sexual harassment and its effects on victims.They also completed a task that involved collectively generating examples of sexual harassment, writing these examples on blank cards, and then collectively rating and ordering the examples in terms of perceived seriousness.
In the dietary focus groups, participants filled out a post-group feedback survey about the focus group experience.Participants rated the group on whether it was relaxed, civil, awkward, stimulating, and boring, using a sevenpoint likert scale from strongly disagree (1) to strongly agree (7).We also asked them to complete an open-ended question, which asked for feedback on the format of the focus group.Participants were not asked for feedback after the sexual harassment focus group due to time constraints.
Finally, participants signed the second part of the consent form to consent to their discussion being analysed by the research team and were reimbursed for their time.The dietary group participants were offered £8 to thank them their time, and the sexual harassment groups were thanked for their participation through course credits.

Analysis
We evaluated the success of the focus groups in terms of three questions: whether participants could manage the discussion effectively without a moderator in the room, whether the remotely-moderated focus groups allow for access to highquality rich group-level talk, and whether the focus group was a positive experience for participants overall.Our analysis is presented below.

The effective management of the discussion by participants
The moderator plays the important role in traditional focus groups of keeping discussion focussed and managing group dynamics, a role that is typically considered to be critical to the quality and richness of the data produced.Our analysis focusses on how our groups discussed the questions posed to them remotely without a moderator present, and whether and how individual group members adopted the typical roles of a moderator.We found that both the dietary identity and sexual harassment discussion groups successfully managed their social interaction in a variety of ways to compensate for the lack of moderator, which we explore below using a line-by-line discursive analysis (Wiggins, 2018).

Avoiding the problem of the dominant talker
A clear advantage of a focus group design compared to an interview is that the multiple and potentially contradictory perspectives represented in the group can be discussed and negotiated, and this advantage hinges on no participants dominating the discussion.What we observed when we removed the moderator was that participants were very capable of filling this moderator role themselves, showing conscientious awareness of the rest of the group and ensuring everyone had a turn to speak and answer each question.This spontaneous turn-taking was particularly successful in the smaller, three-person, vegan and vegetarian groups, where participants regularly invited and encouraged quieter participants into the discussion using phrases such as 'how about you' and asking each other questions about their experiences, as we see in Extract 1 below.

Extract 1 (vegetarian group 1, lines 175-188)
CATH: ten years ago it was really bad (.)When I started (.)In extract 2, Mark describes his vulnerability around his veganism to the group.
When the discussion question changes on the slide, he realises that other group members did not have the opportunity to share their views, and asks Ella directly to expand on her point.Ella takes this opportunity and is met with a supportive 'mmhm' from Mark.Before these focus groups, participants were told that they should use the slides as a 'guide' and allow other participants to finish their point before moving on to the next topic.This extract shows Mark acknowledging this and inviting Ella to actively participate in the discussion before moving on.For participants, overruling the discussion guide in this way would likely be much harder in a moderated focus group, where doing so might require an element of confrontation with the moderator.Here, the group manages their own discussion, and allows room for other participants to share their full experiences in their own time.There were multiple occasions in the sexual harassment group discussions where the speaker attempted to pass the floor to the other participants.
Extract 3 (male sexual harassment group 4, lines 406-415) The consideration of others in these extracts shows the participants are aware of their own talk and are attending comprehensively to the conventions of both turn taking in conversation, and an equity of 'airtime' between participants in getting to express their views.Participants ensure that they are not taking up too much conversational space and are inviting quieter participants to participate.Drawing out the voices of quieter participants in focus groups can often be difficult, and participants in the remotely moderated focus groups recognised this.There was evidence in both the dietary and sexual harassment focus groups of specific awareness and consideration of quieter participants, as seen in David's talk below: Extract 4 (flexitarian group 1, lines 873-884) In extract 4, David tells the group a story about his childhood.While expressing his experience, he also attends to the needs of others and verbally seeks to clarify whether another participant was 'gonna say something'.He then apologises and again verbally clarifies that he wasn't interrupting, before posing a question to the rest of the group 'anything else that makes it easier'.In this way, we see how David moderates his own talk, ensuring that his experiences are described, but that he is not dominating the discussion, and that other participants also have an opportunity to contribute.In doing so, he fills what might have been the role of an insitu moderator as a group member.
In each group everyone got the opportunity to answer each question in depth, ensuring a high quality of data, and contrasting perspectives.There was only one group where there was evidence of some participants dominating the conversation whilst one participant remained silent (either through choice or submission).This was a flexitarian focus group comprising five participants (compared to three or four in the other groups).One explanation for this imbalance is that the size of the group meant less time to give each participant an opportunity to speak, suggesting that group size might need to be factored in when considering time provided for questions.Another explanation is that the group size provided more of an opportunity for a participant intent on free riding to do so (Puchta and Potter 2004a).Alternatively, this group might have just been composed of one or more, naturally quieter participants who may have found it more difficult to voice their thoughts or interject.This is a common problem for designs using a physically present moderator, and it was encouraging to see this only happen in one of our 16 remotely-moderated groups.

Encouraging and clarifying the discussion
In general, participants interacted in a supportive and collaborative way, encouraging, engaging with and clarifying each other's experiences and ideas.Even when the topics were sensitive, participants were supportive and recounted their experiences openly with other participants.Here we see two examples of this from the sexual harassment groups: Extract 5 (male sexual harassment group 3, lines 3-5) A: um it was with someone that she knew in real life(.) In extract 5, B reinforces C's contribution ('that's a really good point') before sense-checking and consolidating their shared understanding.In Extract 6, participant C's sharing of a friend's experience is met with interest and sharing of a similar story by participant B ('that's actually happened to my friend as well').Such displays of alignment across the focus groups were important to creating an atmosphere where personal opinion and sensitive experiences could be discussed.They are also examples of what is a highly-prized, but all too frequently elusive, quality of focus groups, namely, participants' orientation to each other rather than to an in-situ moderator.
Another typical role of an in-situ focus group moderator is to clarify or reword questions that participants don't immediately understand, often via the use of pre-planned 'follow-up probes' that sit below each question on one's focus group guide.While every effort was taken at the design stage to ensure that the questions presented to participants on the screen were clear, losing the ability to interactively clarify and draw attention to particular elements of questions risked a substantial loss of relevant talk.However, we found that participants also took this clarifying role upon themselves in the conversation, often verbalising and deconstructing questions for other group members, as demonstrated in the extract below: Due to the poorly defined nature of the flexitarian identity, this question was quite confusing for some participants in those groups, and there was often a moment of silence after the presented question progressed to this one.In the above extract, David verbalises the question, both to reduce situational awkwardness of the group silence and to clarify the question for confused others, just as a traditional focus group moderator would typically do.Participants also often spontaneously took up the responsibility to facilitate the continuous conversation by asking additional questions that branched out from the researcher's questions presented on the screen.These questions that participants posed to the group were typically phrased in a manner that stimulated discussion and experience-sharing from other participants, as we see below in the context of the sexual harassment groups: Without the moderator present in the room, the participants seemed particularly willing to go beyond their usual role of answering questions and took over the question-asking role (another outcome that is often highly-prized, but difficult to achieve, in a traditionally moderated focus group).Being able to document this question-asking process provided insight into the ambiguity and confusion surrounding the topic due to the varying definition of online sexual harassment behaviours.Indeed, the type and content of such questions asked by participants can provide researchers with insights regarding the issues within this domain that spark concern and warrant attention, from the participants' perspectives (and expressed in their own terms/words).Moreover, these questions can be generated in a way that is free from direction or restriction imposed by the moderator.This freedom may be particularly helpful for researchers who are studying groups that they don't identify with themselves because the focus group participants have the ability to shed light on parts of their experiences that may have previously been unclear to the research team.This methodology might also form an important part of a coproduction process, where unexpected questions might be expanded upon in more traditional methods for future research.Group members also extended the reach of questions in interesting ways in situations where discussion stalled.For example, in the vegetarian group featured below in extract 10, participants initially answered the question of 'how would you describe what a vegetarian diet is' with relative ease, owing to the strict non-meat-eating practice boundaries and norms of this dietary group.However, we see below how, after a lull in conversation, Mike then extends this question to an interesting end: This discussion exposed some of the flexibilities in dietary identity boundaries that our empirical research agenda was hoping to examine.Furthermore, not only does Mike act as an in-situ moderator might in this conversation, he also problematises the identity boundary previously drawn by the group, leading to a rich conversation that would be ideal for exploring using more discursive forms of analysis (e.g., discursive psychology-Wiggins, 2016).Typically, how questions are asked by a (researcher) moderator would be of little interest to many researchers, especially those coming from a more realist epistemological approach, and would often be excluded from analysis.However, in this unmoderated group no data is 'wasted'.Indeed, how questions become framed by participants themselves can become important theoretical features ripe for qualitative analysis.

The management of group conflict
When designing this methodology, we were concerned that participants might mostly agree with each other to alleviate any social awkwardness or tension in the group.This is also an issue in traditionally moderated focus groups, which is addressed largely by the moderator inviting disagreement (Myers 1998).We worried about whether disagreements would still emerge among the remotelymoderated groups, and how any conflict would be handled by participants themselves.We found that, without a moderator present, there was some conflict in opinion throughout the groups, which then necessitated management by the participants themselves.In a traditional focus group, a moderator would typically fill this role and manage conflict within a group by changing the subject or moving on to another question (Braun and Clarke 2013) Mark's comparison of discrimination against vegans to racial discrimination is a very contentious view and many would consider this an offensive remark.This statement could pose problems for the group's subsequent interaction.However, it could be argued that the flat hierarchy of the focus group without a moderator present in the room, allowed Rhona to strongly contest Mark's point here, and indeed perhaps facilitated her doing so.Myers's (1998) work examining focus groups from a conversation analytic perspective demonstrates that, in traditionally moderated focus groups, participants are much more willing to explicitly disagree with a moderator than with other group members.He demonstrates that participant-to-participant disagreements are often softened and hedged to avoid offending other participants.It is interesting to see direct conflict occur in the remotely moderated group here, without this initial hedging response or talk going via the moderator as the source of disagreement.In a traditionally moderated focus group, participants may well have heard this controversial point and looked to the moderator to see if/how they were going to respond to potentially contentious content, effectively deferring to them as the person 'in charge' of 'smoothing' the interaction.However, here, as the group themselves are responsible for managing conflict, we can gain a sense of some the ideological differences present within this group, and how the group manages them collaboratively.A traditional moderator might have invited disagreement among the group here by encouraging group discussion, for example by saying 'what do others think' or 'do the rest of you agree?' (Myers 1998).While these utterances are designed to encourage others to participate, they also have a potential effect of indirectly 'shutting down' the original participant.This interaction plays out differently without a moderator present in the room.In extract 13 below, the group calls back to this initial instance of disagreement, expanding their arguments even further as a group, and shedding important light onto how ideological differences between vegans are (re)negotiated in group talk.Here we present a longer extract to comprehensively demonstrate how the vegans tackle the disagreement together as a group: This extract illustrates how the group is impacted by, and continues to successfully manage, this difficult disagreement in their talk later on in the focus group discussion.Not unlike a traditional moderator, Mark explicitly calls back to the initial disagreement as it comes up in later discussion 'I know you really disagree . . .', which gives Rhona a new opportunity to expand on her argument.Even the quieter participant Ella is drawn in to offer her mixed opinion, and through this interaction we gain a sense of how group members have reflected on and managed this potentially difficult ideological disagreement in their discussion together.
Allowing this possibility for conflict may be seen by some critics of this design as unethical.However, this experience of interactional conflict is not dissimilar to the discussions people may have within their social circles on a daily basis (Chuck, Fernandes, and Hyers 2016).The way this method facilitates, rather than hampers, more direct argumentative conflict between participants (rather than 'through' a moderator) may be advantageous for its utility in exploring research questions surrounding a range of issues, including, in this case, group boundaries and motivations.However, researchers hoping to use this method should be mindful of the types of conflict that may occur within their groups, and make plans for how to intervene in the event of discriminatory or hateful speech and actions.Offering participants a discreet way to alert the researcher to their discomfort-for example by giving them remote buttons they can use to 'bring the researcher in' if need be-may be effective.For this reason, we argue for a remote moderator, who observes the focus group from a distance and can still step into the group if truly needed to diffuse conflict.
It could be argued that conflict is harder to manage in sensitive topic groups, such as in our groups discussing sexual harassment.However, we found that these groups also managed conflict productively.In one group, as we see below, one participant (B) expressed disagreement on a comment made by another participant (D) when rating the severity of online sexual harassment behaviour examples.The participant (B) conveyed with apparent ease his concerns regarding the underplaying of the emotional impact of sexual harassment acts on the victim while putting the focus on the perpetrator's intention: Extract 14 (male sexual harassment group 3, lines 146-163) Observing this interactional conflict within the group shed light on typically subtle expressions of differences in values among individuals with shared gender identity.These differences were significant enough to the participants that they felt the need to contest and address their disagreement with other in-group members, and the way they deal with this conflict is ripe for discursive analysis.One notes, for instance, the socialinteractional work that B performs ('I know this isn't necessarily what you mean') to 'soften' what might be potentially read as an accusation levelled at D (Edwards 2000).Also interesting is the way in which B constructs this as a collective problematic assumption ('I worry that we run the risk of'), which works to deflect the notion that it should be seen as a specific problematic position ascribed to D as an individual.In the absence of a moderator, participants' abilities to handle disagreements respectfully became key to a peaceful and ethical discussion experience.While sexual harassment is typically a sensitive topic of discussion, participants in this study were perfectly able to express disagreements in a respectful manner.
The identity homogeneity and incorporation of a group task in the design may have also helped participants develop rapport, shaping them to act collaboratively and handle subsequent disputes with each other respectfully.
It should be noted here that the conflict introduced in these sessions is not dissimilar to the conflict one might experience in everyday life.While leaving participants alone in this environment might be seen by some as potentially irresponsible, or a situation that the use of an in-situ group moderator would try to prevent, the lack of immediate moderator-participant power structure allows the group to express their disagreement in a way that sheds light on the research questions at hand.While one of the identified roles of an in situ moderator in traditional focus groups has been to try to elicit a wide variety of opinions from the group, it has been shown (Myers 1998) that their presence does tend to inhibit participants from engaging in direct disagreements with one another (preferring to instead express these 'via' the moderator).This arguably hampers the collection of data that allows one to examine how such disagreements might be conversationally managed and negotiated between persons in more (hierarchically-flat) everyday international contexts.We hope to have empirically demonstrated how the remotelymoderated focus group can mitigate some of these difficulties by opening up spaces where such disagreement can occur in more direct ways that may be of interest to the research enterprise, but without creating significant ethical concerns.

Remotely-moderated focus groups give access to rich group talk
It was vital that the groups felt safe to explore the intricacies of their group identities and perspectives without fear of outgroup mockery, derogation or invalidation.In both the dietary and sexual harassment contexts, groups quickly established rapport, largely based on their shared identities and experiences.This process was facilitated by the focus group guide which (similar to a traditional, moderated, focus group) started off broad-to give the group time to get to know each other-and became more focussed over time to allow deeper discussion.The remotelymoderated methodology was effective to the extent that the participants appeared comfortable and were actively sharing personal experiences throughout the discussions.
We also hoped that this design would encourage participants to talk about the less (socially) desirable aspects of their identities, such as occasions where they may have slipped up on their dietary practices despite their commitment to their identity label, or times they may have engaged in, or experienced sexual harassment.The space of the unmoderated group indeed allowed these conversations to occur.Particularly in the vegan group, which, based on existing literature, one would ordinarily expect to be the strictest in observing dietary identity-associated practices, we saw open and frank conversations.Below Rhona recounts her experiences of a drunken slip-up to the group, and her experiences of shame upon becoming aware of her actions.

Extract 15 (vegan focus group 1-lines 354-361)
RHONA: because I (.) have had slip ups (.)And I was even telling her before that (.) um (.)I got really drunk (.) and then ordered loads of dominos (.)And then the next day (.)I say I woke up (.) and I looked in the fridge and I was judging my flatmate like (.) ugh classic Cara (.) she's ordered herself loads of dominos!ELLA: ((laughs)) RHONA: And then I opened the box and I was like NO I ORDERED DOMINOS (.) THIS IS AWFUL!And I just obviously felt (.) so bad (.) because I did something that was out of line with my principles (.)This is a personal story for Rhona that, in a group with non-vegans (either other participants or the moderator) could become grounds for potential derogation (on the grounds of 'hypocrisy') and identity invalidation (Rothgerber 2013).In a traditionally moderated focus group, Rhona may have anticipated such negative responses, and not shared her story as a result.However, in this remotely-moderated focus group with only sympathetic members present in the room, we see Rhona develop trust with the group that allows her to share the story without fear of reputational consequence.As shown in the next turn of the discussion (picked up in extract 16 below), the group is very supportive in their response and we see how this shared anecdote then opens up a further conversation with the group about how to manage identity slipups, generating a rich account that advances our understanding of the inner workings of vegan identity.Whilst sexual harassment is often gendered in nature, meaning that women are commonly expected to be victims, gender identity homogeneity in the sexual harassment groups allowed male participants to disclose subjective experiences of being sexually harassed.The following extract provided an example of a participant (A) being encouraged to speak about his experience respectfully by one of the other members (C).
Extract 17 (male sexual harassment group 4, lines 70-77) A: okay yes I sort of experience but not sure if it was even about sexual harassment.

C: you don't have to share if you don't want to.
A: no no no no I mean it's just because I was just having messages with a lady and she asked me to go for drinks and I didn't want to but she was trying to persuade me and then I just like say bye but i mean it's not about sexual harassment (.) she wasn't offending me or something like this basically, she was just trying to uh have a dinner with me I guess (.)

C: I think there are different levels of sexual harassment
We see how A is demonstrably hesitant to share his account, as is taken up by C ('You don't have to share if you don't want to').However, A precedes to share his experience with the group and we see how this sharing is met with alignment by C in so much as A's decision to share this experience in the context of the discussion of sexual harassment is legitimised by C's response ('I think there are different levels of sexual harassment').It is likely that in the presence of outgroup members, in this case female participants or moderator, A may not feel encouraged to verbalize his personal example at all out of worries of it seeming trivial in comparison to the female experience.It is also worth noting that the male participants' knowledge of the (currently conversationally absent) researcher being a woman did not seem to compromise the intended effect of having identity homogenous groups.Although participants were fully aware that the discussions were being recorded (and listened to later by the researcher), we found that they did not appear at all constrained or inhibited by this, in (conversational) practice.Our approach also aimed to create an environment where participants could freely discuss their opinions of outgroup members and so shed light on how groups construct and situate ingroup-outgroup boundaries.It was apparent that the lack of a moderator (in combination with homogenous groups) successfully facilitated this talk, with group members openly discussing their disdain for other groups.In the extract below, Mark uses the group environment as an outlet for his anger at outgroup members, and in this passage clearly separates and outlines the boundaries between the Vegan 'us' and the Non-Vegan 'them': Extract 18 (vegan focus group 1, lines 255-269) MARK: And like you were saying (.) watching the documentaries?(.) going FUCK this is the reality?I think (.) this is what unites us as vegans (.)We (2.0) not only are aware of and understand the horror?(.) but we're also willing to do something about it?There are so many people who have seen those horrible movies and documentaries and everything (.) who know what's happening in the industry (.) and who still go >oh it tastes so nice?< Weak motherfucker (.) How dare you (.) and I know that's (.) and I'm sorry like that's my angry vegan coming out RHONA, ELLA: ((laughs)) MARK: but it's so (.) and I just don't understand (.) how do you (.)I don't know how you can be aware (.) but not make that change in your life?(.) ELLA: I agree RHONA: mmm That's really interesting (.) Mark's use of the collective pronouns 'we' and 'us' here constructs a clear distinction between the aware and active vegan on the one hand, and the more reprehensible, aware yet inactive, non-vegan.He openly derogates this group, utilising a common trope of the 'angry vegan' (Guerin 2013) to demonstrate why and how his construction of veganism is at odds with his judgement of the non-vegan wider population.Mark here speaks directly to the imagined outgroup member, derogating them in a way that he acknowledges is representative of various vegan stereotypes.Mark is surrounded here by sympathetic vegans who show their support verbally and through their laughter.However, what is perhaps most telling, and methodologically important, is that, (as mentioned above) despite knowing that the group is being recorded, he feels comfortable to be able to say something which may cause significant social-interactional trouble in a larger, mixed identity group.Indeed, Rhona here also fulfils a traditional role of the in-situ moderator, by encouraging Mark and validating the relevance of his experience to the group: 'mmm that's really interesting'.The vegetarian group also constructed clear boundaries between themselves and 'the non-moral other'.In the following extract, Mike speaks about his difficulty in refraining from overt judgement of meat-eaters in daily life, describing a desire to judge meat-eaters as a 'temptation'.Cath concurs with Mike's moral judgement, arguing that 'eating beef is one of the worst things you can do', and that avoiding 'resistance' from meateaters is difficult: Rather than engaging in the face-saving practices seen as common to vegetarians -where overt moral derogation of meat-eaters is avoided by vegetarians to minimise interactional trouble (Greenebaum 2012) -here Mike openly expresses his disdain for those who eat meat and his struggle to resist the temptation to express such moral judgement in his daily life.
In this paper, we focussed primarily on the efficacy of the methodology in our analysis.But we also note here that, to answer the empirical questions posed by each design, the group transcripts were further analysed (by the first and second authors) using discursive psychology (Wiggins 2016) andthematic analysis (Braun and Clarke 2022).In both circumstances, we found that the remotely moderated focus groups generated rich data which was suitable for qualitative analysis using these analytic approaches.Thus, we can see how the remotely-moderated focus group method is able to generate very rich data that allows a researcher (ethical) access to a rather novel form of 'fly on the wall' data for understanding intragroup processes and discussions.

The experience of the group for the participants themselves
As well as being successful from a methodological, and data-quality standpoint, participants themselves also responded well to the setup of the focus groups in their post-group feedback survey.We only collected formal participant feedback from the dietary focus groups, as that was the first pilot of this methodology.We had tighter time constraints in the sexual harassment groups, as these were compensated using course credit.Thus, we prioritised time for the discussion and sorting task alone in those groups.In an online survey after the dietary focus groups, participants rated the group positively on a variety of attributes, reporting the group discussion experience as highly relaxed, civil and stimulating while not being particularly awkward or boring (see Table 2).
Participants also commented on the enjoyable nature of the homogenousidentity focus group in their responses to the open-ended question on the feedback survey, which asked them for any written feedback they had on their experience.

Extract 20 (anonymous vegan participant)
"Good to discuss veganism with other vegans!"Some participants in the dietary groups directly referenced the absence of outgroup members in the group as a core factor for their enjoyment, and ability to speak freely: Participants in the flexitarian groups also mentioned the value of having a space to talk to similarly identified others about issues they couldn't typically talk about with their friendship group: Extract 22 (Anonymous flexitarian participant) "It was a topic that I don't get to speak about with friends very often, and it was nice to know that we all agreed to some extent." Vegans, vegetarians and flexitarians are typically in the minority, and it is not common for these groups to encounter homogenous group environments such as these in this manner (at least by chance alone) in daily life.Manufacturing this environment for this design therefore took advantage of the opportunity to record and analyse bonding of the group over their identity in real-time and the discussions that group members may not be able to verbalise often in their daily lives.Due to time, topic, and funding constraints, limited feedback on the design was gathered from the sexual harassment groups, however following the debrief, one participant noted: Extract 23 (anonymous male participant) "I actually think if you had actually been sitting in the room with us it would be a totally different kind of discussion I think it was probably a bit more natural and spontaneous thank you for having us participate." Thus, the design seemingly succeeded on the third criterion of also being an enjoyable experience for participants, as well successfully alleviating many of the methodological difficulties in-situ moderators might present for studying groups in these research contexts.

Discussion
Taken together, these two studies demonstrate the efficacy of a remotelymoderated focus group method and evaluate how this new methodology might fit as a tool in the qualitative researcher's toolkit alongside traditionally moderated focus groups.Our analysis demonstrates that these focus groups function well with a remote moderator, they generate rich data, and that the discussion sessions are enjoyed by participants.The exploration of this design in two different topic areas demonstrates its flexibility and its ability to handle more sensitive topics like dietary identities and sexual harassment.The dietary identity focus groups asked explicitly about experiences and opinions of an identity, whereas the sexual harassment groups asked about a broader topic where views and experiences have been shaped by a gendered experience of living as a man or a woman.

Ethical considerations
Ethical considerations were a large concern for us throughout this project, and the ethical issues present in trusting the group alone with the conversation should not be understated.In these studies, we chose not to tell participants that they were being observed by a remote moderator.However, in future iterations it may not be necessary to deceive participants in this manner.
Understanding the impact of this deception would require further evaluation of the method comparing remotely moderated focus groups where the remote moderator was made obvious to, or concealed from participants.It would be necessary to ensure that participants do not attend to the remote moderator more when they are made explicitly aware of their presence.Alternatively, one could consider trialling the impact on the unfolding discussion of literally setting the questions to change on the screen using predetermined timings (although, based on our experience, we suspect this may not work as well).We would argue that any deception should be considered with reflections on how it may impact participants' wellbeing and ability to provide informed consent, and that this should be judged accordingly on a study-by-study basis.
Similarly, informed consent is a vital component of this methodology.To gain rich conversational data, it is important that participants are aware that they will be discussing a topic with other people (strangers in our designs), without a moderator present in the room, when they sign up to participate.As this design is not conventional, participants should ideally be fully informed about what the group will involve before they choose to participate and must be given the opportunity and ability to withdraw as they were in these studies.Furthermore, participants should be made aware of the topic of the discussion so that they can avoid any topics that may feel uncomfortable discussing with others.Informing participants that it will be an unsupervised group discussion, as well as the topic of discussion, ahead of time may also be vital to ensure the inclusion of neurodivergent participants, who may need time to adjust their expectations of the method, or who become uneasy if the study context/ topic is different to what they expected.There is a chance that the nature of the group may discourage participation from some, but this is an issue common to focus group methodologies overall, not just this remotely-moderated design.
As we note in the methods section, researchers looking to adopt this method should have a clear plan in place for both how and when to intervene should the groups pose a threat to participant wellbeing and/or safety.Researchers should have a clear process for how and when a conversation should be halted.In our examples, we ensured that at least two members of the research team were available at all times during research to assist and manage the group if required.However, the amount of support should be tailored according to the research context.In our studies, participants were told that the researcher was waiting outside and to go outside if they had any problems.However, if the topic were even more sensitive in nature or more likely to distress people, participants may need an alternative route to contact the researcher discreetly without drawing the potential ire of the rest of the group.For example, participants could be provided with discreet buttons to an alarm that they could press to bring the lead researcher into the room if they feel uncomfortable.Of course, a detailed and well-considered plan on what to do should participants press this alarm would be necessary to protect participants and minimise awkwardness of an intervention.
As well as having a clear plan of how to intervene, researchers should also be clear about when they need to intervene.What are the ground rules around participation and safety: for example, if a discussion becomes too heated or threatening to participants?These ethical concerns led to our decision to make the groups 'remotely-moderated' rather than completely unmoderated.The lead researcher in both studies was always watching the conversation and set ground rules for the group to make it clear that disrespectful or derogatory language and behaviour would not be tolerated.In both studies detailed here, the remote moderators never had to intervene, and all conversations were respectful and were in line with the guidelines given.Though it should be noted that this may be primarily due to the 'common ground' participants had in both contexts with group members holding the same identity to them.Future remotely-moderated designs bringing different groups together may need to be more cautious of group conflict arising.More challenging though is for the research team to consider whether there are any rules around which topics are not permitted for discussion (where conversations begin to touch on illegality or protection issues, for example; or when the team considers the risk to participants of a conversation outweighs the benefits to the research project).Participants may struggle with the ambiguity in setting these boundaries considering that part of the design's aim was to create a space in which participants would feel comfortable disclosing things that might have otherwise been inhibited by the presence of an in-situ moderator.Providing participants with clear instructions and a discreet way of alerting the remote moderator when they feel uncomfortable to continue the discussion, including when they feel emotionally violated or distressed, may help prevent risking participants' well-being over the richness of data collected.Furthermore, starting each study with a brief 'warm up' or 'training' session where guidelines and boundaries are clearly established by the remote-moderator may be helpful to address this.This may be particularly important when the study is examining more controversial or potentially upsetting topics than explored here, and we would encourage research exploring particularly sensitive issues to consider other methodologies that would allow for more direct participant support and/or appropriate levels of safeguarding.However, it should be noted that these are essential considerations that all kinds of qualitative research face, which are not limited to this design alone.We encourage researchers to seriously consider the ethical implications of this remotely moderated focus group methodology in their specific research context and make adjustments to the method accordingly to ensure participant wellbeing is protected.

The benefits of being willing to 'let go'
As Walters (2020) notes, the experience of relinquishing control in qualitative research can be daunting.This remotely-moderated focus group is perhaps particularly challenging in this regard, especially for researchers who are used to conducting traditional interviews and focus groups, in so much as it puts the research process much more in the hands of participants.Many may see this as a risky approach, and we recognise that this design will not be appropriate for all research questions and contexts.However, when the use of an insitu moderator is problematic for research, we argue these risks are outweighed by the high quality of generated data where talk is shaped primarily by the group, and less (both directly and indirectly) by the presence of an insitu moderator.
Furthermore, it could be argued that this approach works to democratise the research process, giving the group more of a chance to discuss what they think is most interesting and important about the topic at hand.This approach also allows for natural progression of a conversation that may cover topics that the researchers may have not been aware of in their planning.Some may argue that this is also a feature of semi-structured interviews and moderated focus groups, to the extent that these approaches do allow for some departure from the schedule/guide (Smithson 2000).However, we would argue that there is often a tendency for the researcher (when a present participant in the conversation) to still 'mark out' these departures in ways that sometimes actually shut them down.Even somewhat ubiquitous phrases like 'Oh this is really interesting, that's not something I'd ever considered', arguably sometimes risk such an outcome because of the hierarchical power structure of a moderated group.Firstly, there is the risk of the group suddenly perceiving that they have gone 'off piste'.Secondly, such utterances by the moderator also work to reinforce and remind of the hierarchical nature of the knowledge sharing that is going on -if they, as the 'all knowing expert' hadn't thought of it before then it must be important/interesting.The comparatively flexible and autonomous format of the remotely-moderated approach gives more room for conversations to flow and new insights to emerge organically, without being conversationally passed judgement upon by the researcher in any (even indirect or unintentional) way.If one of the goals of qualitative methods can be to enable further understanding of participant experiences (in their own words and on their own terms), then giving participants the opportunity to generate their own insights and take control over the conversation direction arguably helps to alleviate issues of power held between participants and researchers.The participants are entrusted with the research and their representation of the group.This approach is particularly useful when understanding group constructions or representations.After having had the chance to 'warm up', these groups even felt comfortable enough to problematise themselves and their opponents.The level of rapport developed among the group's shared experience also benefited the quality of data.

Remaining questions
We hope that the presentation of this methodology sparks a new discussion around remotely moderated focus group methodologies as a complimentary tool, and their potential for use in different research contexts.However, there are remaining questions about this design that should be investigated in future research.While we made some observations based on our own experiences as a research team and focus group interactional processes previously identified in the literature, these two studies did not offer a direct comparative analysis (within a single study design) of how remotely-moderated focus groups might differ to in-situ moderated focus groups.Further studies wishing to investigate qualitative differences between the designs could compare and analyse them directly.This may be of interest when it comes to who takes on the moderating role in each focus group type, and how the type of talk generated differs between focus group types.
Furthermore, both studies were conducted in a university context, with university students and staff, located on a university campus.It would be interesting to investigate how this design could be flexibly used in other environments, for example within local communities or schools.We chose our study location based on what we felt would make participants feel most comfortable, and what was most convenient for participants to travel to.We encourage other researchers to also change the design to best accommodate their participants.We used a 'classroom' layout with chairs surrounding a table in both designs reported in this paper as participants needed space to eat their lunch and do the sorting task.However, this configuration could and should be changed to suit the research question and context.There may be additional benefits to using this design in more personalised (lounge-like) contexts, as this might make participants feel more comfortable and able to open up about their experiences.It would also be interesting to see how this design works for different participant groups with different requirements.We argue that the design should be adapted to best accommodate participant's preferences and needs and recognise that accounts of how this might work in practice could help to improve the use of this methodology outside of the university context.
These groups were all held in person, and we do not yet know if this design might also work in an online environment.Online discussion platforms such as Zoom and Discord are becoming increasingly common, and may work well with this design, but thorough examination is needed to confirm this.An online version of this design would, of course, require further ethical and practical considerations that may need to expand on those already noted in this paper.
We used data from these studies for further thematic analysis (from a critical realist and realist ontology) and discursive psychological analysis (from a social constructionist epistemology), and we can confirm this data was rich for analysis using both methods.However, further research could determine if this data collection method works for other analytic and epistemological perspectives.This method should not be assumed to be relevant to all research questions, just as some research questions answered through focus group designs would not be appropriate to examine in an interview (and vice versa).Overall, we believe that this design is likely to be most appropriate for contexts where processes of group interaction are of most interest in the analysis, where the ethical risks surrounding discussions are minimal, and where participants are assumed to have strong experience of the topic in question that will be easy and agreeable for them to recall in a group discussion.We expect, for instance, that accounts that are highly indepth and personal (such as those used in Narrative analyses or Interpretive Phenomenological Analysis) might not be best produced using this method.Additionally, this method is still fundamentally researcher-generated, and the data cannot be classified as 'naturally occurring', potentially making it less palatable for other perspectives such as conversation analysis.

Concluding thoughts
Our proposed new approach fills the gaps between a completely hands-off approach to qualitative data collection (i.e., naturally occurring data) and hands-on approaches where the researcher directly guides the discussion as a conversational participant (i.e., interviews or traditional focus groups).The variety of data produced in these two remotely moderated focus group studies demonstrates the use of this paradigm for a number of research topics within the social sciences.However, it remains to be seen how well our proposed new method might work in a range of different kinds of settings.For example, whether it might prove useful in the context of Participatory Action Research (Clavering and McLaughlin 2007;Fine and Torre 2019) or when conducted in community-based settings where the researcher has less control over the physical surrounds.However, in presenting our approach here, and providing some empirical evidence for its efficacy in the contexts in which we have used it thus far, we hope to potentially inspire others to consider utilising it within their own qualitative research programs and that they might also find it similarly useful.undermine) social cohesion and community well-being.She is also interested in the diverse forms of organisation through which people pursue social change and how inter and intragroup processes shape their identities and strategies.Jan Willem Bolderdijk (PhD) is a Full Professor of Marketing & Sustainability at the University of Amsterdam, and director of the Amsterdam Centre for Responsible Consumption.Jan Willem also works at the University of Groningen, as an Associate Professor at the Department of Marketing.His interests include social influence, moral decision making, social tipping points & marketing.Tim Kurz is a Senior Lecturer in the School of Psychological Science at the University of Western Australia.His research focuses on the social psychological dimensions of attempts to foster more environmentally sustainable societies, with a particular focus on the role of morality and social identity and the ways in which both play out within social interaction.

Figure 1 .
Figure 1.Authors' image of discussion room and participants in the dietary identity focus groups.
0) 'Do you label your diet when in public' (reading from screen) I guess that means (.) are we flexitarians?(laugh) (1.0)In public?
Saffron O'Neill is a Full Professor in Climate and Society in the Geography Department at the University of Exeter, UK.Her research explores the social science dimensions of climate variability and change, particularly focusing on communication and public engagement.Her research specialism is the visual communication of climate change.

Table 1 .
Demographics of each focus group.
We see here how, after spending time outlining her own experiences, Cath directly invites Mike to comment on the question, asking 'how about you?'.While Sarah is commenting throughout, Mike is quieter, and Cath attends to this by giving him an easy opportunity to participate in the conversation, one that he quickly takes up.This was also displayed in the vegan groups, as outlined in the following interaction between Mark and the (quieter) Ella.
Extract 2: vegan group 1 (lines 363-373) MARK: You know?Right?Because it feels so vulnerable doesn't it?Because of all that stigma (.) and all that (.) crap (.) that gets put onto (.) (Q4: How do you think others feel about your diet?(Click-slide changes)) MARK: I'm sorry sorry we didn't get-what's your opinion?What was your (.) ELLA: Um (.) what was it (.) about labels?Oh (.) just that I think that labels can be positive thing?Because they create that sense of community?(.) and I think that it may be encourages people to think (.) 'oh I wanna be vegan because that looks like a good thing' (.) but it can also be a negative thing (.) because it can be associated with a certain stereotype MARK: mmhm

female sexual harassment group 1, lines 291-310)
B: that's a really good point, do you mean like there's always sort of the kinda emotional element to the harassment beyond (.)Extract 6 (C: it happened to one of my friends, so um(.) . . .

Extract 8 (female sexual harassment group 1, lines 453-454)
C: Do you guys think like online sexual harassment is more severe than real life sexual harassment?

Extract 10 (vegetarian focus group 1-lines 16-21)
CATH: so is it then vegetarian?Are you then vegetarian if you eat(.)bugs?Or not?SARAH: I feel like no? MIKE: I agree (.) not strictly (.) you're not strictly being vegetarian if you eat bugs (.) SARAH: yeah (.)I don't have the same (.) like affinity for a bug that I do a cow (.)I just feel like (.) they have more like (.) emotions (.) and I feel like they're more like a being (.)I know that's (.) CATH: yeah MIKE: are we making an arbitrary cut off though?Like between living things?SARAH: exactly (.) it's so arbitrary (.) that's where I draw the line

Extract 13 (vegan focus group 1, lines 837-848; 856-862)
MARK: but I totally get I understand I'm not taking that away at all about marginalisedgroups and everything absolutely (.)ELLA: It's funny I feel quite in-between with that at the moment between your two points cuz I understand both parts of it?And it's more that that's quite a good metaphor?Like we need a metaphor in a sense to communicate our idea to people who aren't vegan so they can have understanding of it(.)

Table 2 .
Participant ratings of dietary focus group qualities on likert scales of 1 (not at all)-7 (very).