Serial Focus Groups: A Longitudinal Design for Studying Interactive Discourse

Abstract

Focus group methods specialize in the analysis of interactive discourse, but are only rarely employed as a stand-alone method to study such phenomena, owing to inherent limitations concerning the comparability and generalizability of findings. In this paper, we argue that focus groups undergo three kinds of transformations, involving changes in participants’ cognitive states, social ties, and discursive behavior, which raise both analytic challenges and valuable opportunities for the study of shared meanings and interactive negotiation processes in society. Introducing Serial Focus Groups, we extend familiar focus group designs as a method for studying interactive discourse in a longitudinal perspective, capitalizing on the analytic potentials raised by these transformations. Reviewing the methodological literature and drawing upon two large-scale focus group studies of socially interactive sense-making, we argue that serial focus groups can help overcome some of the limitations of cross-sectional focus groups and offer valuable new opportunities for analysis and validation.

Serial Focus Groups

A Longitudinal Design for Studying Interactive Discourse

Focus groups have a long tradition in social scientific research (Jarvis, 2011). Combining elements of individual qualitative interviews, ethnographic observation techniques, and quantitative surveys, they occupy a unique niche in research methodology that generates rich narrative data with an awareness of interactive group processes and social context (Barbour, 2007; Morgan, 1997). Especially for the study of social, political and cultural beliefs and attitudes, focus groups have been hailed as a prime avenue for examining “how meanings are constructed collectively and in situ” (Kamberelis & Dimitriadis, 2014, p. 318; see also Green et al., 2003).

At the same time, focus groups come with important limitations as a strategy for social inquiry. First, focus group discussions develop in heavily context- and path-dependent ways (Barbour, 2007). Observed patterns and processes cannot be readily compared or attributed to specific stimuli, variables or conditions, but remain at least in part achievements of unique group interactions. Second, focus groups respond strongly to the presence or absence of social ties (Kitzinger, 1994). Focus group researchers can either rely on groups of strangers, creating somewhat artificial communication settings (Brink & Edgecombe, 2003); or they can rely on pre-existing social groups, whose discussion behavior is co-shaped by unobserved norms and shared experiences (Bloor et al., 2001). Third, focus group interviews focus on relatively short episodes of interactive discourse and remain limited in their utility for longitudinal research (see Read, 2018, for a parallel argument about qualitative interviews).

In the present paper, we review existing scholarship and draw on the experiences from two large focus groups studies to introduce serial focus groups as a possible response to these limitations. We define serial focus groups as a research design wherein the same groups are reconvened multiple times over an extended period of time, periodically revisiting the same focal issue through interactive discussion (see also Krueger & Casey, 1991). Six groups of five to nine participants each were convened up to three times over a period of three months for a project investigating audiences’ strategies for evaluating contested ideological narratives in Eastern Ukraine. For another study, five focus groups were convened five times over a period of eight months to study voters’ future-oriented discourse over two rounds of Israeli general elections. For the present article, we examined the data collected by both studies to identify methodological opportunities afforded by the serial design, which we discuss in light of the methodological scholarship.

This paper is organized as follows: After a brief exposition of focus groups as a research method, we introduce the two studies that present the empirical base of our investigation. To develop our main argument, we posit that focus groups systematically transform participants’ knowledge, social identities, and discursive behavior. As groups of initial strangers are gradually transformed into realistic social settings, serial focus groups render these transformations observable to the researcher. As a result, we argue, serial focus groups not only introduce a longitudinal perspective to focus group research, but they also aid researchers in disentangling some of the complexity underneath the observed discourse, offering valuable new avenues for increasing the depth, systematicity and validity of analysis.

Focus Groups as a Method for Empirical Social Research

As an empirical research method, focus groups can broadly be defined as a method that convenes small groups of individuals to discuss some focal issue, moderated and observed by a researcher, in order to gain insights into those ideas expressed and discursive processes involved in the interactive discussion (e.g., Morgan, 1996; Vaughn et al., 1996). Following this definition, focus groups as a method for empirical social research can be delineated from (a) group interviews that focus on dyadic interactions with the moderator, (b) unmoderated group interactions (usually in ethnographic research; Stewart & Williams, 2005), and (c) group interviews conducted for purposes other than social inquiry (e.g., to educate or empower participants; Kamberelis & Dimitriadis, 2014).

Due to their focus on interactive discourse, focus group discussions cannot be conceptualized as the sum of individual responses, but are collaborative in a constitutive sense (Morgan, 1996; Wilkinson, 2004): While individuals and their views still matter in the discussion (Merton et al., 1956), their (discursive and other) behavior is embedded into the social interaction and responds to the contributions of others (Cyr, 2019) in ways that permit no straightforward attribution of observations to individual participants (Kamberelis & Dimitriadis, 2014; Lunt & Livingstone, 1996). Data is generated primarily by the social situation wherein participants seek to position and express themselves and negotiate their views within the group (Bonito, 2004; Kitzinger, 1994; Vaughn et al., 1996), thereby enabling researchers to observe the collaborative construction of meaning (Lunt & Livingstone, 1996). This perspective, which focus groups share with ethnographic methods (e.g., Spradley, 1979) or textual conversation analysis (e.g., Schegloff, 1984), is critical if the matter under investigation is best understood as collaborative achievement in a social setting (e.g., Gamson, 1992). Unlike deliberative polling, which focuses on the formation of shared opinions via guided social interaction (Fishkin, 2003), in focus groups, the interactive process constitutes the primary data. The prominent role of interactive discourse mitigates the extent to which the researcher-controlled focus dominates the data generation, separating focus group interviews – alongside unstructured and some forms of semi-structured interviews – from more structured interviews and surveys (Morgan, 1997). Accordingly, focus groups invite participants to ‘bring[.] forward their own priorities and perspectives’ (Smithson, 2007, p. 357) and guide the researcher toward uncharted insights (e.g., Nind and Vinha, 2016; see also Hughes and DuMont, 1993).

Despite the often-naturalistic flow of the group discussion (Kamberelis & Dimitriadis, 2014), however, focus groups cannot be understood as ‘natural’ observations (Brink & Edgecombe, 2003). The strategic creation of the group, its focus, and the moderation shape the interaction dynamic, ensuring that most interactions will pertain to a given research question (Morgan, 1997). This is especially important for studying phenomena that are not verbalized at predictable rates and venues and are therefore difficult to access without the researcher creating relevant situations (Bloor et al., 2001; Morgan, 1997). Unlike interviewing techniques that assume that probed views are (to some extent) already available to an interviewee (Zaller & Feldman, 1992), focus group data is deliberately co-created through the group interaction (Hughes & DuMont, 1993). Specifically, focus groups expose participants to new ideas and social contexts, and encourage them to engage with these through the group discussion (Kamberelis & Dimitriadis, 2014).

As a consequence, focus groups’ reliance on group interactions as a means of elicitation creates important challenges for the analysis. Unlike non-reactive methods, such as the study of textual discourse interactions (e.g., Meredith & Potter, 2014) or observational methods, which often go to considerable lengths to minimize the effect of the researcher’s presence (Spradley, 1979), focus groups give rise to path-dependent dynamics that constitute a meaningful intervention in the life worlds of participants. Over time, focus group interactions inadvertently transform the group setting in at least three key ways ¹ (Kamberelis & Dimitriadis, 2014): First, in what we term a “cognitive transformation”, focus groups alter the knowledge available to participants (Bonito, 2004), not only through the information exchanged within the group, but also through participants’ cognitive engagement with debated issues. Second, focus groups undergo a “social transformation”, as the group interaction invites participants to form social ties and position themselves in relation to one another, engaging in identity work (Munday, 2006). Third, and as a consequence of both aforementioned processes, focus groups also experience a “discursive transformation”, as the emergence of shared understandings and tacit social norms within the group alters participants’ discursive behavior: As the group gradually transforms into a richer social setting, participants tend to disclose personal information more readily, engage more with others’ views, and become enabled to build upon pre-established common ground, working toward the negotiation of a mutually acceptable position (Bonito, 2004). Throughout the evolving discussions, participants’ frequently updated beliefs, their densifying social relations as well as emergent communication norms inform, shape and filter their contributions (Kamberelis & Dimitriadis, 2014; Krueger, 1994), which are often born out of the very situation in which they are made (Kitzinger, 1994; Merton et al., 1956). Unlike experimental designs, which enable a concise attribution of observed changes to well-specified interventions, the evolutionary transformation of focus group interactions thus blurs the epistemic status of collected data. Any turn depends on a whole host of individual and collective, contextual and emergent factors, which cannot be readily disentangled (Morgan, 1997; Summerson Carr, 2010). Both for the generalization of findings, and for the comparative or longitudinal analysis of focus group data, this mode of interactive elicitation and transformation raises considerable barriers.

Taken together, focus groups can thus be characterized as a method that deliberately creates a somewhat artificial social setting in order to enable researchers’ access to interactive discursive processes and social phenomena (Kamberelis & Dimitriadis, 2014) that are not commonly verbalized, or expressed in settings that are not accessible, predictable, or sufficiently sustained to permit a meaningful analysis (Morgan, 1996). This unique capacity is obtained at the expense of diminished external validity; the filtering of personal beliefs and attitudes through the group process, which continually transforms the created social setting; and a somewhat vexing epistemic problem concerning the comparability and generalizability of collected data (Krueger, 1994; Millward, 1995).

Owing to these limitations, focus groups have been used primarily as methods for generating hypotheses (Merton et al., 1956), as well as auxiliary tools for validating instruments (Morrison, 1998; Vaughn et al., 1996) or interpreting findings obtained by other methods, such as surveys (Tavener et al., 2016), (semi-)structured interviews (Peek & Fothergill, 2009; Reay, 2004), observation (Munday, 2006), or textual analysis (Schrøder & Phillips, 2007). Only few, large-scale studies have inspired sufficient confidence to formulate theoretical conclusions based on focus group research alone (e.g., Gamson, 1992; Just et al., 1996). As the path-dependent group process complicates the identification of systematic patterns, focus groups are largely absent from comparative and longitudinal research, save for a short list of studies relying on pre-existing groups, closer to an ethnographic tradition (e.g., Cramer, 2016; Eliasoph, 1998; for exceptions, see Bronstein & Mason, 2013; Dolan et al., 1999; Jonsson et al., 2002; Nind & Vinha, 2016). In the methodological literature, we found passing references (e.g., Krueger & Casey, 1991, p. 192; Wilkinson, 2004, p. 178) but virtually no discussion of longitudinal focus group designs. In the following, we will introduce two studies that relied on serial focus groups, which inform our present argument.

Two Studies Using Serial Focus Groups

This paper draws upon two separate applications of serial focus group interviews in very different settings. One study was conducted in Eastern Ukraine during summer 2019 to understand residents’ strategies for navigating the contested information ecology shaped by the ongoing Russian-Ukrainian conflict. For this study, the use of focus groups was motivated by the intention to understand the negotiation of tacit norms and practices in engaging with contested information, which are socially embedded, rarely verbalized and difficult to tap in hierarchical, individual interviews (Bloor et al., 2001). Serial focus groups were conducted specifically to examine the implications of sustained social interactions upon shared norms and practices (Kamberelis & Dimitriadis, 2014; Read, 2018). For this purpose, six groups (U1–U6) were reconvened up to three times over the course of three months. Specifically, the second author convened three groups of younger (19–35; U1/U3/U5) and another three of older participants (50+; U2/U4/U6), recruited through a combination of snowball sampling and screening to ensure diversity with regard to gender, political orientation, political interest, and age. Meetings took place in a major city (U1/U2), a medium-sized town (U3/U4), and a village (U5/U6) in Eastern Ukraine.

The other study was conducted in Israel during the running-up to two consecutive rounds of general elections in 2019 and focused on voters’ expectations regarding the prospective election outcomes. The use of focus groups was motivated by the need to elicit underlying reasons and considerations informing participants’ expectations (Morgan, 1997) and gauge the extent to which expectations were shared among participants (Kitzinger, 1994). Serial focus groups served to track the evolution of expectations over time, identifying changes in projection strategies as the elections approach. For this study, five groups – one of young (18–28) Jewish voters (I2), one of Arab voters (I5), and three groups of Jewish voters with mixed age (29+; I1/I3/I4), recruited from a commercial panel using stratification criteria to obtain groups diverse on political and religious identification, education, gender, and age – were reconvened five times over the time span of eight months.

In both studies, group size varied between five and twelve participants. Participants were encouraged to identify themselves using pseudonyms and were not previously acquainted with one another, save for two instances in the Ukrainian study, and one instance in the Arab voters’ group in Israel. Only the latter group experienced significant attrition over time and required an addition of new participants after the third wave. To ensure continued participation, the Israeli study, but not the Ukrainian one, relied on gradually increasing remuneration; however, participants’ growing intrinsic commitment to subsequent meetings, which was observed in both studies, suggests that this may have been unnecessary.

Focus groups followed a similar structure in both studies, commencing with a broad discussion of the situation, before focusing on specific themes raised by the group, and exposing participants to selected media contents. In both applications, moderation guidelines mandated restraint and encouraged interactive discussion among the participants. The Ukrainian sessions were moderated by the second author, a doctoral student supervised by the first author. Discussions took place in both Ukrainian and Russian language, of which this researcher is a native speaker. Meetings of the four Jewish Israeli groups were conducted jointly and in Hebrew by the third and fourth author, both native speakers of Hebrew. The meetings of the Arab Israeli group were conducted in Arabic by two associated researchers, both native speakers of Arabic. All sessions lasted around 75–90 min and were video-recorded and transcribed.

In both applications, group sessions were followed by brief questionnaires to record participants’ media use habits, political attitudes and identities. In the Israeli study, participants were asked to reflect on the process of the reconvened groups in the final meetings. In the Ukrainian study, group sessions were followed by individual interviews with selected participants to elicit their private reflections upon the preceding group discussion. In total, 15 such interviews were conducted, probing participants’ perceptions concerning group dynamics and power relations within the discussions, as well as personal disagreements. Interviews lasted between 20 and 60 min and were audio-recorded and transcribed.

Both studies relied on a Grounded Theory approach to analyze the data (Strauss and Corbin, 1997): In the Ukrainian study, analysis focused on media-related lay theories and attitudes, information behavior, and epistemic norms (Pasitselska, 2022a; 2022b). In the Israeli study, analysis focused on participants’ strategies for predicting electoral outcomes and the likelihood and desirability of projected outcomes (Tenenboim-Weinblatt et al., 2022; in print). Both analyses emphasized the dynamics of interactive negotiations, as well as the social embedding of the group interaction. In the Ukrainian study, also non-verbal behavior was analyzed. Critically, in both applications, intermittent real-world developments – e.g., in the election campaigns, including the calling of renewed elections in Israel; in the violent conflict and accompanying media narratives in Ukraine – constituted key factors in the longitudinally comparative analysis.

For our present argument, we draw upon a meta-analysis of regularities and trajectories observed across all 11 groups. The entirety of data drawn upon this study comprises full transcripts of 39 focus group sessions (14 in Ukrainian/Russian, 20 in Hebrew, five in Arabic), as well as 15 interview transcripts (from the Ukrainian study). Specifically, we were interested in the implications of reconvening groups for the mode and content of interactive discourse, contrasting groups’ initial meetings (which represent the design of one-off focus group meetings) against any subsequent meetings. Our analysis focused on (a) transformations in the knowledge negotiated in the group, both in terms of presumed common ground and the contents, presentation and justification of claims; (b) transformations in the social configuration of the group, such as participants’ self-positioning and identity negotiations; and (c) transformations in the discursive norms and behaviors in the interaction, including the mode of disclosure, collaboration and contestation, as well as the policing of communicative norms.

Serial Focus Groups as a Transformative Social Setting

Owing to the endogeneity inherent to focus group discourse, focus group interviews raise more profound concerns in a longitudinal perspective than most other methods. If participants’ views, social relations and discursive behaviors already undergo continuous transformation over the course of a single focus group session, any subsequent reconvening of the same group inevitably extends existing path-dependencies. Serial focus groups give participants additional opportunity, time and incentive to learn, form social identities, and adjust their discursive behavior. Participants who entered the first meeting as strangers encounter one another in later meetings as acquaintances, sharing a history of common knowledge, social ties and ongoing discussions. Despite these important challenges caused by the ongoing transformation of the group, we argue that the repeated and prolonged observation also offers valuable opportunities for empirical social research. In a nutshell, we propose that the longitudinal reconvention of focus groups enables the formation of a reasonably realistic social setting under the eyes of the researcher, which affords a much richer range of comparative opportunities than one-off focus group meetings.

From a Group of Strangers to a Realistic Social Setting

For focus group research, both composing groups out of strangers and relying on pre-existing social settings raises distinct challenges, which have been widely debated in the scholarly literature (Silverman, 1993). Among strangers, especially initiating and facilitating interactive discussions often requires considerable effort. Both the unfamiliar setting and the lack of shared knowledge, mutual trust and common purpose can present a major obstacle for the discussion, which often remains notably artificial for a while (Warr, 2005). In pre-existing social groups, discussions are usually easier to initiate, as they can draw upon prior social ties and common knowledge. At the same time, acquaintanceship-based groups tend to take the debate up from a pre-existing base of common knowledge, tacit consensus and social identities, which are unlikely to be verbalized in the discussion, and thus unavailable for analysis (Morgan, 1997). By contrast, strangers need to explicitly develop and negotiate such common ground, offering researchers a valuable window into the underlying transformations that permit an analysis of shared ideas, subject positionings and collective identities.

In serial focus groups, group discussions commence as interactions among strangers, enabling the observation of emergent common knowledge, social ties and discursive norms. Over time, groups gradually transform into more naturalistic social settings underpinned by rich common ground, sustaining ever deeper, more meaningful exchanges among participants. In this way, serial focus groups exploit both participants’ initial unfamiliarity and their later familiarity with one another to generate rich, contextualized data. In addition, their distinct design helps avoid some constraints that may be imposed upon group discussions among acquaintances by their need to maintain social relations that exist outside the group, as pointed out, for instance, by Zohar (I3, Session 5): ‘...at the workplace, I’m not talking too much about politics. […] I don’t, don’t want, it’s too loaded. […] And you also don’t want to create [tensions] with your colleagues’. Accordingly, focus group participants that share membership in existing social groups are often reluctant to enter into certain discussions (Krueger, 1994), especially where these touch upon sensitive issues and external power relations. By crafting rich social settings artificially out of initial groups of strangers, serial focus groups enable trustful conversations without the risk of spill-overs into participants’ pre-existing life worlds: ‘I could say everything that I thought’ (Artem, U1, Session 3). Intimate discussions are facilitated both by the growing sense of acquaintanceship and trust, and by participants’ relative anonymity beyond the group setting (Krueger, 1994) ² . By transforming initial strangers into later acquaintances within a constrained, researcher-observed social context, serial focus groups thus combine the advantages of both settings. In particular, the capacity to directly observe the ongoing underlying transformation of participants’ cognitive states, social ties and discursive behaviors renders valuable additional information accessible for the analysis. In the following, we discuss each of these transformations in turn.

Cognitive Transformation

While all focus groups tend to transform what participants know and think about (notably, due to the researcher-controlled focus and mutual learning; Bonito, 2004), opportunities for cognitive transformation are notably extended in serial focus groups. Participants are not merely invited to consider issues that they have no strong, pre-formed knowledge about, but they gradually form confident opinions through repeated discussion and reflection in the times between sessions:

Lidia: And sometimes you even made us also think about something,

Zinaida: Yes! Absolutely right. Yes-yes-yes-yes.

Lidia: about the things that we haven’t considered, and you made us ponder these questions.

U 4, Session 3

Serial focus groups create additional incentives for participants to access available information in the times between sessions, fostering prolonged engagement (Jonsson et al., 2002). For instance, Nathan explained, ‘these discussions that were very energetic brought my interest back, like thanks to the first group meeting, I became more interested, I started to read’ (I4, Session 5). Some participants deliberately prepared for group meetings, to the point that Grigoryi (U4) brought extensive documentation and video materials to the sessions. Through participants’ reflection upon their acquired knowledge, the process of mutual learning becomes observable to the researcher:

Yuriy: Prepare your arguments!

Artem: Watch the news! [laughs]

Yuriy: Watch the news, yes. I will write down my arguments, you will write down yours, and we will debate.

U1, Sessi on 2

In addition, the prospect of repeated interactions motivates participants to consider what claims are accepted or contested in the group, to carefully position their own views and, if necessary, to try and persuade other members of their legitimacy:

It made me think once again about the… about my opinions. That I should justify them. […] So you rethink, you think maybe it does not make sense, and maybe there are also things [on the other side that] I totally accept. Nothing here, nothing is black and white!

Hani, I3, Session 5

Over time, participants generate shared understandings, which are iteratively refined through the group interaction. Where views remain contested, the prolonged interaction deepens participants’ recognition and knowledge of alternative perspectives (Kamberelis & Dimitriadis, 2014). For instance, Oksana observed that ‘In my circle there are few people with opinions like Yuriy’s, so it was also interesting to listen’ (U1, Session 3); and Gil noted: ‘It also gives you other perceptions [...] and I want to hear some of the people, even the left. Even the extreme left. Also the opinions of the Arabs, it is important.’ (I3, Session 5). As the participants position beliefs as shared or contested, also the researcher is enabled to reconstruct the group’s evolving set of commonly recognized ideas.

Social Transformation

While focus group participants generally enter into social relations with one another through collaborative interaction, serial focus groups rapidly develop a dense mesh of social ties that structures subsequent meetings. Participants initially position themselves vis-à-vis their fellow group members, forming coalitions by signaling or expressing support or resistance to one another’s contributions, organizing around key cleavages in the discussion:

Well, it seemed to me that the group was very much divided into those who took some ideological position. Like there was a man, who with such overtly … not anti-Ukrainian, but with distrust toward the Ukrainian side. There was a fervent more like nationalist, a kind of a philosopher, and the rest were more like … like regular citizens, the people, so to say.

Igor, U1, Personal Interview

In subsequent meetings, then, participants expressly recognize and often anticipate one another’s stances, and relate to one another as familiar, thus revealing key parts of the group’s emergent social structuring:

I think we are now at a stage where we know what everyone, more or less, is going to say, everyone’s opinions, where is it going, so there is no longer the part of ‘confronting’ so to speak, and telling everyone what I think, or what he thinks, because we already know. And there is some understanding already that this is the situation.

Tova, I1, Session 5

As a growing sense of empathy and trust underpins participants’ recognition of divergent positions, solidarity and mutual support grow among participants (Bronstein & Mason, 2013), facilitating disclosure and engagement. In the words of Gil, ‘It seems to me that we all understood that we are the same nation, leftists and rightists, and in the end, everyone wants the same thing’ (I3, Session 5). Even where participants find themselves at opposite ends of a controversy, the collaborative, civil mode of focus group discourse enables the negotiation also of contested issues:

It is like we are sitting now, we have different points of view, but we talk normally. […] No one throws chairs at others.

Valeriya, U1, Session 2

Eventually, this mutual recognition extended to the point that one group insisted on waiting for two delayed right-wing participants before commencing the discussion:

Moderator: Hello everyone and thanks for coming…

Tal: But we do not have the contra [points at Avraham and Gil’s empty chairs]!

[…]

Dvir: [as Gil enters the room] Here you are! You! You are insane! What's up with you?

Zohar: You left us [the right-wingers] in the minority!

Dvir: We delayed the session until you arrive.

Hani: We were under pressure that there was no one to talk to!

I3, Session 4

Through the artificial creation of the group around a focal issue that most participants do not normally discuss, group session came to be recognized by the participants as a valuable venue for discussing specific issues of common interest (Warr, 2005, see also Rodriguez et al., 2011) generating both enjoyment and intrinsic motivation to continue:

In my circle I don’t discuss with anyone such topics, and in general the topic of politics I’m trying to avoid as much as possible. And here it was a kind of fun! I could say everything that I thought. And in any case my opinion is not judged right or wrong.

Artem, U1, Session 3

Most groups developed collective identities revolving around their shared discussions (Kamberelis & Dimitriadis, 2014). Expressly negotiating these identities while referring to themselves as ‘we’, group discussions enabled researchers’ direct access to shared positions and the group’s collectively constructed sense of purpose:

In this group, between seven people, we have five opinions that we have all invested in and discovered, and this is something that you generally don’t find in other groups or other settings.

Wardi, I5, Session 5

Especially during the conclusion of the later sessions, participants regularly expressed their looking forward to the next meeting, to the point that participants jokingly wished for Israeli coalition talks to fail again, so there would be new elections and thus, additional group sessions. In another Israeli group, a participant went so far as to suggest inviting the entire group for a barbecue.

Discursive Transformation

While the group interaction acts as a key motivation for participants to share and contribute in any focus group interaction (e.g., Morgan, 1997), serial focus groups gradually shift through somewhat different modes of interaction, revealing different aspects of participants’ more self-centered and more socially embedded discourses. In both studies, participants reflected on these dynamics toward the end of the first round of meetings: ‘We are only warming up!’ (Artem, U4, Session 1).

In their first session, most groups commenced with a focus on strategic self-positioning, eliciting mostly personal accounts accompanied by other participants’ expressions of agreement, empathy or contestation:

Ksenia: I raised my two daughters alone and I have been a single mother for almost 20 years. I didn’t rest, not even for a day! […] I worked in two jobs, and I still work today. I stopped buying vegetables last year because it’s so expensive […] I don’t buy meat […] nor fruit. […] It’s impossible to support a family, the prices here are insane!

[…]

Ronen: I know that if we meet again in a year and a half, there’s a chance I'll hear Ksenia say again that she can’t afford buying vegetables and it breaks my heart. I want her to buy vegetables!

I4, Session 1

Already during the first session, but especially during the second, the growth of common ground and social relations gradually shifted the focus toward more interactive discourse, reflecting participants’ deepened desire to understand and engage one another’s views:

Raisa: Oh girls, I will again talk about something that I shouldn’t

Maria: No, why, talk, for God’s sake

Liliya: Raisochka, you must talk. Raisa, talk, your opinion is very important for me.

U6, Session 2

Over time, developing group norms raised not only an expectation to share, but also a growing sense of being safe to do so (Bonito, 2004) – exemplified by Elizaveta (U6), who only disclosed during the third meeting that her family had recently moved out of the Russian-occupied territories. Especially in the second and third sessions, we registered entire sequences where participants ‘interviewed’ one another.

As collective group identities gradually emerge, discourse shifts once again toward identifying shared positions and actualizing the group’s purpose vis-à-vis the researcher: ‘You called us, simple people, and sat, and listened.’ (Lidia, U4, Session 3). At the same time, the expanding range of jointly elaborated norms and perceptions nurtures a shared sense of what are acceptable contributions, which may at times discourage participants from voicing deviant views (Kitzinger, 1994; Smithson, 2007): ‘I’ll better be silent, or they’ll beat me up! [laughs]’ (Alla, U2, Session 2). Accordingly, two complementary dynamics interact over time. One supports increasing disclosure and engagement, especially where most participants can find themselves in the emerging group norms:

Aliza: From meeting to meeting I felt freer to express myself. […]

Sami: […] the level of openness has risen in an amazing way.

I4, Session 5

The other dynamic progressively establishes limits to discriminate between permissible and deviant views, endowing the emerging social texture with a potential to silence deviant contributions:

Harel: In the US, the ultra-Orthodox work, pay taxes, don’t exploit the country, although they live an extreme lifestyle. There, you have a complete separation between religion and state and they accept it, only here [not].

Benny: You and your agendas! […] It’s not related to the discussion here! What do you want, […] to delete sections in the Torah? […] This is not the place to spread your own hatred of religion […]

Harel: First of all, if you’re upset […] then I apologize. […] I have no hatred, not of the ultra-Orthodox, and especially not of you.

I1, Session 3

Analyzing Serial Focus Groups

Through the gradual progression of interaction modes, serial focus groups permit a much richer and more nuanced analysis of the observed group discourse, which enables a differentiation of varying degrees of personal and collective narration. They do so, of course, at the cost of exacerbating observations’ dependency on countless interacting factors, as subsequent sessions are additionally contingent upon previous ones, upon intermittent external events, and upon participants’ changing engagement with the focal issue and the group. At the same time, similarly to comparative validation strategies commonly used in ethnographic research (Spradley, 1979; Summerson Carr, 2010), the sustained observation also offers three key opportunities for the analysis: the observability of transformative processes, the combination of cross-sectional and longitudinal comparisons, and the added richness of data. In the following, we will discuss these opportunities in turn.

Observability of Transformative Processes

Serial focus groups accompany an initial group of strangers throughout the extended process of forming common knowledge, social ties and discursive norms, which are characteristic of existing social settings and enable a much more committed, collaborative and informed discussion. Unlike conventional uses of focus groups, which have to choose between using strangers or existing social groups, their design thus combines the advantages of both variants. Witnessing the same groups develop into ever-richer social settings, enables researchers to examine how changes in the social and cognitive configuration of the group shape both the discursive behavior of individual participants, and the dynamics of interactive group discourse. Any shared understandings and social identities present in the group were by necessity constructed in prior interactions, in the presence of the researcher.

For the analysis, this means that the entire body of prior common ground becomes available to interpret subsequent interactions, which is valuable for delineating emergent shared understandings of key constructs. For example, in the Ukrainian study, group U3 completed the entire first session without mentioning ‘propaganda’. After the construct was introduced in the second session (by the moderator) and immediately debated in the group, participants later applied it throughout their discussions. While such uses largely eschewed further elaboration, they remained transparent as references to shared definitions negotiated earlier in the group process.

Veronica: No, I’m not against Ukrainian [language]. I am against the – any kind of propaganda. I think these are terrible things.

U3, Session 2

Similarly, group U2 early-on introduced a range of analogies between the Russian-Ukrainian conflict and the Nazi occupation of Ukraine in the 1940s, thus establishing a powerful normative framework for subsequent evaluations, invoked in the following exchange:

Anna: If I can buy cheap gas, [swears], why should I buy more expensive gas from Europe? And if I’m buying it from Russia, then I surrendered. Why did I surrender?!

Vitaly: Why didn’t you surrender to the Germans, then?

[...]

Anatoly: In 1942, if Germany gave you the best price for Soviet metal, would you agree to sell metal to Germany?

U2, Session 2

Likewise, observing the formation of social ties offers invaluable context for interpreting subsequent group behavior. In the Israeli study, for instance, participants recurrently raised claims that they knew other group members to agree with in order to mobilize support for their views. For instance, in the following exchange, Yoel relies on the group’s shared familiarity with Ksenia’s personal situation (quoted above) to demonstrate the importance of turning out to vote:

Yoel: You [Ksenia] talked in one of the meetings about the super expensive prices in the supermarket, right?

Ksenia: Yes!

Yoel: The only ones who talked about this in their campaign are the Blue and White guys. […] Bibi [Netanyahu] did not even mention it, and if he stays, there is no chance that prices will go down. […] Really, your only chance that something will happen with this...

Ksenia: Yes, is to exercise your influence at the ballot box, yes […]

I4, Session 5

Similarly, we were able to trace how participants’ personal, prior views were transformed (and occasionally suppressed) by the group – such as Peleg’s initial perception of one Israeli party as orthodox, which the group convinced her to revise:

I do not know who I am going to vote for [...] I am confused [laughs]. That’s all I know. How is it possible that every time I come here with an opinion and leave without an opinion, explain to me! If this is your goal you have succeeded! [...] This is amazing!

Peleg, I2, Session 2

Through the observation of the participants’ gradual cognitive, social and discursive transformation, the analysis gained considerable confidence both concerning the attribution of claims to more personal or collectively negotiated positions, and concerning the depth of implied meanings and pragmatic achievements.

Cross-sectional and Longitudinal Comparison

Through their repeated observations over extensive periods of time, serial focus groups introduce a longitudinal perspective to focus group research. Unlike conventional uses of focus groups, whose powerful, path-dependent group dynamics tend to undermine meaningful comparison, this addition of a second dimension of systematic variation enables some degree of both cross-sectional and longitudinal comparative analysis. In a longitudinal perspective, participants independently witness the same external events in the time between meetings, without the interference of group discourse ³ . As a consequence, we can interpret consistent over-time changes that recur across groups as an effect of either the passage of time or the occurrence of external events. For example, in the Israeli study, the failure of coalition negotiations following the first round of elections diminished participants’ confidence in their forecasts across all groups, which shifted from predicting the future coalition and its policies to more proximate concerns; however, the same events did not diminish participants’ overall optimism about a favorable outcome (Baden et al., 2021; Tenenboim-Weinblatt et al., 2022).

Similarly, parallel transformations in participants’ discursive behavior across different groups can be fruitfully interpreted. Following the same failure of coalition talks in Israel, for instance, some supporters of former prime minister Netanyahu reduced the certitude with which they predicted a renewed Netanyahu-led government, while others’ certainty remained undiminished, revealing important differences in projections’ reliance on participants’ political identities. Exploiting this opportunity for longitudinal comparison, we could develop a typology of personal projection styles, characterized by different degrees of wishful thinking, defeatism and responsiveness to changes in available information (Tenenboim-Weinblatt et al., in print).

In a cross-sectional perspective, we would be cautious to attribute differential outcomes of one-off focus group discussions to differences in group composition or specific events within the discussion, owing to the powerful influence of accidental path dependencies within the interaction. By contrast, persistent differences between groups that recur across multiple sessions may be interpreted with some confidence. For instance, the Israeli group I3 consistently stood out as showing the most heated political disputes, which became enshrined in the group’s communication norms. Following media reports that right-wing activists had deliberately lied to polling institutes about their voting intentions, in an effort to foster false hope and a rude awakening among left-wing voters (Tenenboim-Weinblatt et al., 2022), the group responded much more aggressively than any of the other groups:

Hani: The Likudniks [supporters of the right-wing Likud party] cheated!

Abraham [a right-winger]: [next time] I will lie too! […]

Hani: Maybe [we] leftists should also lie.

Hadar: Yes, I will also [lie and] say that I am voting for the Likud.

I3, Session 4

However, against the backdrop of the group’s strongly polarized and emotional discussion style, we could recognize this exchange as a performance of ingroup loyalty, rather than actual hostility toward supporters of the opposing camp.

In other cases, also the fact that some groups developed along distinct, but continuous trajectories can be fruitfully interpreted. This was the case, for example, when some groups in the Ukrainian study appropriated key categories from available media narratives, while other groups progressively abandoned these in favor of collaboratively constructed narratives and identities (Pasitselska, 2022c). The addition of a second, longitudinal dimension of comparison to the juxtaposition of groups thus facilitates the delineation of common factors shaping discussions across different groups from those idiosyncratic dynamics that arise from the specific path-dependencies inherent to each group or specific meeting.

Richness of Data

Through the re-convention of the same groups at multiple points in time, serial focus groups permit the repeated observation of similar discursive situations, while mitigating the impact of path-dependent interaction dynamics. Unlike conventional uses of focus groups, which make it difficult to probe deeper without calling considerable attention to the intervention, serial focus groups enable the researcher to revisit key interactions (including those identified only ex post) in a less obtrusive manner, adding nuance and drawing out complexity and contradiction (Kamberelis & Dimitriadis, 2014; Read, 2018). For instance, when the group of Arab voters in Israel (I5) initially discussed the possibility of voting for Zionist parties (Session 4), the controversy focused narrowly on economic interests versus Palestinian identity politics; however, when the group returned to the issue one session later, it added a broader focus on (in)effective representation, which had no place in the preceding interaction.

In addition, participants would occasionally reflect upon their own stances and their position within the group, offering valuable corroboration for the analysis. For instance, in the following comment, Yuriy clarifies that only his pro-Russian comments constitute his personal views, while other, discrepant comments merely reflect his familiarity with competing narratives:

Yuriy: [in Russian] No, I can take a pro-Ukrainian position from the TV, I have seen this position, I can also say that there should be Ukrainian language, and [switching into Ukrainian], I can do it while speaking in Ukrainian, you understand?

U1, Session 2

Furthermore, the possibility to compare the same participants’ statements across different sessions offers a valuable validation of the presented viewpoints. In the Ukrainian study, this strategy enabled us to distinguish provocative statements that most likely represent deliberate trolling behavior, from others that appeared to reflect genuine, extreme political views. Similarly, observing the group discourse among Arab voters in Israel (I5) both in the presence and absence of a salient activist participant helped us distinguish between largely shared political attitudes and positions introduced primarily to accommodate a specific ideological discourse.

Finally, the availability of repeated observations proved instrumental for differentiating between participants’ strategic self-positioning, the influence of emergent group consensus, and the disclosure of more idiosyncratic, personal views. For instance, the second author could show how Eastern Ukrainian participants’ distancing of the ongoing conflict reflected less of a genuine disengagement, but responded to shifting communication norms negotiated in the group (Pasitselska, 2022b). Especially at the beginning, but also later following exchanges that positioned simple people as powerless victims of the conflict, participants tended to echo common public narratives that encouraged cynical media-related attitudes:

Maria: There is no truth on the TV channels. Me and my sister, we don’t watch it, that box. Our mother watches it, some series, due to her age. We don’t trust it.

U6, Session 1

At later stages, however, and especially when present communication norms legitimized a more committed stance, the same participants often presented much more differentiated views:

Maria: It [pro-Ukrainian channels’ coverage] is more real. [...] If not for [specific channel], we wouldn’t know many things.

U6, Session 2

In the same vein, the availability of recurrent observations also acts as a safeguard against over-emphasizing deviant behavior observed within the discussion. Without the benefit of multiple meetings, it would have been near impossible to notice which participants might have had a bad day (e.g., following their party’s defeat) or required more time than others to start feeling safe or “warm up” in the initial meeting, and which were truly reluctant to share (e.g., in the case of Koren [I4], whose conspiracist views made him wary of discussing media coverage). Through their capacity to generate multiple observations of similar discursive situations, serial focus groups reveal important consistencies in individual stances and behaviors, interaction sequences, situation-specific responses and other regularities across groups and meetings, permitting a systematic validation of interpretations: ‘It’s like the Big Brother here, we’ve seen you in several situations!’ (Jasmine, I1 Session 5).

Conclusion

In the present paper, we have argued that beyond extending the limited time horizon of single focus group sessions to enable longitudinal research, serial focus groups offer important new opportunities for analysis and validation in social scientific research. Permitting researchers to deliberately ‘grow’ reasonably realistic social settings under closely observed conditions, serial focus groups expose the gradual unfolding of social ties and common understandings, enabling a well-informed analysis of ensuing group discourse. Where focus groups of strangers are lacking in social structuring and groups of acquaintances bring along much unobserved prior common ground, serial focus groups constitute a third option that combines the advantages of both formats. Separated from participants’ usual social contexts, serial focus groups offer both a certain degree of anonymity and a sense of familiarity and mutual trust, facilitating the discussion of complex and touchy issues. Prioritizing depth of discussion over the breadth of perspectives, serial focus groups offer valuable opportunities for social research, especially where the rich interactive processes of discursive negotiation are in focus.

Practically speaking, serial focus groups may be particularly useful for the study of issues that are closely interwoven with social, political or cultural identities (e.g., ideological beliefs, anti-elite resentment, intergroup conflict; Cramer, 2016; Ron & Maoz, 2013) or embedded in discursive-cultural dynamics of consensus and dissent (e.g., norm/opinion formation; polarization processes; Eliasoph, 1998; Just et al., 1996). Likewise, for the study of deviant experiences or otherwise touchy issues (e.g., deprivation; paranormal belief; Lamont, 2007; Spradley, 1979), as well as for complicated matters that require deeper consideration by participants (e.g., information evaluation strategies, as in our Ukrainian study; ethical dilemmas; complex decision-making processes; Pasitselska, 2022a; 2022b), serial focus groups offer valuable advantages. The same is true where discursive negotiations accompany concurrent real-world developments (e.g., election campaigns, as in our Israeli study; developing crises; Tenenboim-Weinblatt et al., 2022) or when the gradual transformation of knowledge and social relations within the group are relevant for the object of study (e.g., consensus formation; empowerment processes; Kamberelis & Dimitriadis, 2014), making use of serial focus groups’ longitudinal design.

To conduct serial focus groups, the key point of departure is to carefully conceptualize the role of the progression of time in the intended design. Researchers are well-advised to plan out the over-time trajectory of group meetings, discussions and returning issues well in advance, while remaining alert to relevant real-world and within-group developments that might require an adjustment of plans. Besides updating moderation guidelines, interview schedules and materials in light of unfolding group discussions and events, this includes identifying items and discussions that might need to be revisited in future meetings, or even refinements in the pursued research interest that arise from initial insights and necessitate edits in the planned design. Consequently, at least some preliminary analysis of completed sessions should be conducted prior to successive meetings of the same group, and the scheduling needs to permit sufficient time for this. Serial focus groups require meticulous preparation during both data collection and analysis, as special attention must be paid to the tracking of individual, group-specific and situation-specific idiosyncrasies, which feed into the multi-dimensional, comparative analysis.

The need to reconvene the same groups places also important practical constraints on the planning of group meetings, which are sensitive to timing, changes of venues, and changes in the composition of the group (e.g., due to attrition or participants’ availability). While in both our cases, most participants developed a strong intrinsic commitment to their participation in the group’s sessions, every effort should be made to ensure members’ continued participation. Likewise, we recommend staying with the same moderator over successive sessions, which minimizes disruption in the group’s social fabric and enables the moderator to rely on familiar common ground established in previous meetings. Additional demands are placed on protecting participants’ data and privacy (see Kamberelis & Dimitriadis, 2014), as the recurrent format generates ample opportunities for breaches of anonymity, both through the process of organizing the meetings, and through the participants themselves. Other than these, most practical recommendations for one-off focus group research (e.g., concerning group composition, setting and moderation style; Bloor et al., 2001; Cyr, 2019; Morgan, 1997) remain valid also for the conduct of serial focus groups.

Among the key limitations of serial focus groups, first, even the increased naturalism of artificially grown communities remains but an approximation. While groups may come to increasingly resemble naturally existing social settings as they develop, they remain relatively free from the interference of goal-directed interactions, dependencies and power relations that characterize most real-world social settings (Silverman, 1993). Over the course of successive meetings, the formation of social ties and common ground decelerates, approaching saturation – in our cases, after the third or fourth meeting – well below the density of social structuring common in naturally grown communities. Beyond this, additional sessions are useful only if concurrent developments outside the group continue to evolve the group discourse. Another limitation concerns groups’ capacity to develop along idiosyncratic trajectories, which become harder to control over time, and are not easily balanced by convening a few more groups. The rising complexity and sheer size of the data likewise poses important challenges for the analysis, which requires careful organizing (e.g., using qualitative data analysis software) to ensure a systematic appraisal of findings.

In conclusion, serial focus groups extend existing focus group designs, opening up new avenues for social scientific research. Beyond establishing a qualitative analogue to longitudinal panel designs in quantitative survey research, serial focus groups extend serial qualitative interviewing methods (Read, 2018) by adding a focus on interactive group processes and collaborative meaning. Compared to the study of longitudinal group interactions in digital ethnographies or observational methods, serial focus groups add a facility for the researcher to guide and probe the group interaction. Critically, by adding rich opportunities for validation and systematic comparative analysis, serial focus groups contribute to emancipating focus group from their common role as auxiliary method for explorative analysis or the contextualization of findings obtained through other methods, to constituting a stand-alone method in its own right. Especially with its new affordances for studying socially embedded discursive processes, this novel method will hopefully open up new avenues for social inquiry.