A methodology for cross-national comparative focus group research: illustrations from discussions about political protest

Mass mobilization of citizens has influenced political decision-making in countries worldwide and on a wide range of topics. Recent examples include the Indignados movement in Spain and Greece, anti-corruption protests in Romania and Brazil, and student protests against rising tuition fees and education cuts in the United Kingdom (in 2010) and the Netherlands (in 2019). The amount of street protest has increased tremendously in recent years (Granberg 2013) and concerning a wide array of issues (e.g. Aelst and Walgrave 2001). Over the last few decades, social scientists have developed several theories to understand the dynamics of protest. The availability of the necessary resources (McCarthy and Zald 1977) and the political opportunity structure (Kitschelt 1986) have been stressed in mobilization for protests, and political process models have integrated several of these aspects (McAdam et al. 1996). Social psychological approaches (e.g. Klandermans 1997; Van Zomeren et al. 2008) have explored individual-level motivations for participation in collective action, such as perceptions of efficacy and group identification. Nonetheless, a certain element of surprise seems to be part and parcel of many instances of mass mobilization, especially given the contextual differences between various countries (e.g. Kuran 1991; Kurzman 2004). The majority of empirical research on protest departs from the assumption that people have stable, individually held attitudes about protest and participation in protests. However, the formation of these attitudes is rarely straightforward and people tend to change their opinions based on peer influence, current events, and socio-cultural developments (Fishkin 1991; Hollander 2004). In this paper, we develop a methodology for conducting focus groups cross-nationally and argue that this method not only improves our understanding of how attitudes are negotiated in group processes, but also provides an insight into how these attitudes are shaped by a country’s social or historical trajectory.

Social psychological theories claim that the formation of shared opinions within a group is more predictive of participating in collective action than merely identifying with a group (McGarty et al. 2009). Moreover, Schmitt-Beck and Lup (2013) demonstrate that the more people discuss politics within their social circles, the more they participate in politics. This line of research on political talk and political discussions has focused mostly on why and how often people discuss politics, and how this in turn can influence participation in electoral politics (e.g. Bennett et al. 2000; Eveland et al. 2015; McClurg 2006; Pattie and Johnston 2009). However, how citizens talk about protest, and especially the cross-national aspect of political talk has not received much scholarly attention. It is likely that the amount and nature of political discussion varies considerably across different countries. For example, Nir (2012) demonstrates the influence of different electoral systems on political talk and Eveland et al. (2015) show that people from collectivistic and individualistic cultures differ in their willingness to engage in political confrontation.

Conducting focus groups in a cross-national setting means finding a balance between wanting to make the data as comparable as possible across countries but also maintaining the context-specificity which this method provides. In this article we will elaborate on how standardization of the research design can be used to achieve this balance. We draw from data collected in the cross-national research project POLPART,¹ which conducted 80 focus groups in nine different countries, to show how the research design can influence comparability between countries. Moreover, we will show how the political, social, and historical context influence narratives about protest in several countries.

Focus groups have been used in the social sciences since the early 20th century (Merton and Kendall 1946) but have become increasingly popular in the late 1980s and 1990s as a tool for both marketing researchers and scientists (Morgan 1996). Since then, focus group research has provided many insights into how citizens think and talk about politics (e.g. Gamson 1992; Perrin 2009; Walsh 2004). On the topic of political protest however, focus groups have not been used as extensively despite some indications that other qualitative methods provide valuable insights into understanding the dynamics of protest (e.g. Drury 2002; Drury and Reicher 2000; Potter and Reicher 1987).

The core feature of focus groups is that data is collected through the interaction between individuals (Morgan 1996); multiple individuals engage in a dialogue focused on the research theme which is guided by a moderator. The intersubjectivity, or shared meaning, expressed in the focus group constitutes the primary unit of analysis (e.g. Stanley 2016), although several intra-group processes can be studied as well (e.g. Krueger and Casey 2014; Morgan 1996; Stewart and Shamdasani 2014; Wiggins 2016). There is considerable flexibility in how focus groups can be used within different methodological traditions depending on the research question, ranging from observation of natural group discussions to very structured discussions or even adding experimental elements to the design (e.g. Della Porta 2014; Duchesne et al. 2013; Gamson 1992; Myers 1998).

The narratives provided by focus group participants stem from a range of sources, including personal experience, news items, and popular wisdom (Gamson 1992). These narratives are situated in a specific socio-political context (e.g. Lup 2015; Stanley 2016) which is more apparent in focus groups than in survey research. Nonetheless, the vast majority of focus group research has been conducted in a single country (see Perelli-Harris et al. 2014 for an exception). Are there specific historical events that shape current narratives about protest? Are discussions more or less consensual due to diverging cultural norms? Studying focus groups cross-nationally allows for comparison between the sources of the narrative that are developed in focus group discussions.

The core aim of our proposed methodology for cross-national focus group research is to allow for meaningful comparisons between countries. Any country comparison in social science is notoriously hard (e.g. Burkhauser and Lillard 2005) and the context-dependency of focus group discussions (Hollander 2004) adds a layer of difficulty. How do we maintain the richness of the data that focus groups provide without treating these insights as unique instances of political discussion? Our methodology is based around standardization of the design across countries. Standardization of the research design allows for interpreting differences in focus group discussions between countries as stemming from differences in attitudes, the political culture, the social context, and other influences than the research design itself.

In making country comparisons based on survey data, one of the key challenges is to ensure equivalence in what is measured by a certain item in different countries (e.g. Burkhauser and Lillard 2005; Dubrow and Tomescu-Dubrow 2016). However, focus group standardization provides additional challenges: not only do the question items need to be understood the same by participants in different countries but the selection criteria and moderating style also need to be similar across countries in order to produce comparable data.

While the main focus of this paper falls on elaborating which methodological decisions can ensure cross-national comparability between focus groups, we believe the guidelines presented here can also be fruitfully applied in other types of comparative research contexts (e.g. across language groups, across regions within one country).

The main objective of the POLPART project is to investigate the issues that motivate citizens to participate in politics and whether they would engage in institutionalized forms of action—such as voting—or non-institutionalized forms of action—for example street demonstrations (Klandermans 2013). In order to investigate how political action is contingent on a country’s historical trajectory or political culture, the POLPART project conducted focus groups in nine democracies. There are several ways to select cases in designing country comparisons, but typically the decision rests on the balance between similarity versus differences between countries (e.g. Seawright and Gerring 2008). The nine countries in the POLPART project all have a democratic political system in common but differ in whether these are old or new democracies. Moreover, among the new democracies, there was a divide between post-authoritarian and post-communist countries, providing sufficiently different socio-political contexts to compare. The Netherlands, Switzerland, Germany, and the United Kingdom represent ‘old’ or ‘mature’ democracies. Hungary and Romania represent a post-communist context, whereas Argentina, Brazil, and Greece represent new democracies, having changed from authoritarian regimes to representative democracies in recent decades. We opted to organize the focus groups in the capital or other major city in each of the nine countries to ensure comparability in the political opportunities (e.g. Koopmans 1999) for the people who were participating in the focus group discussions. Despite targeting a more specific urban population with this strategy, Goodman et al. (2013) argue that in survey research there is similar bias towards urban rather than rural participants.

The unique aspect of focus group data is the interactive nature of the discussions therefore the most important methodological considerations are the size and composition of the groups (Gamson 1992; Morgan et al. 1998). Compared to one-on-one interviews, focus groups promote self-disclosure to the extent that social homogeneity among the participants is ensured (Duchesne et al. 2013), making it more likely that the participants feel comfortable revealing information to similar others (Morgan et al. 1998). Too much similarity between the participants however will lead to less cross-cutting discussions (e.g. Mutz 2002) where different points of view are being exchanged. Therefore, a balance must be struck between having differing opinions represented, while maintaining social homogeneity within a focus group. This balancing act is complicated in cross-national focus group research where social homogeneity within groups should be similar between countries.

Achieving social homogeneity is usually done by selecting participants on certain comparable criteria. Usually the more countries involved in the cross-national focus group research, the more observable these criteria have to be to maintain the comparative nature. The reason behind this is related to the availability and quality of data on participant characteristics which can be used in a functionally equivalent manner across countries. Selecting certain age groups for example can be easily managed in different countries, and standardized measures for education level (ISCED; OECD 2007) similarly provides a basic measure of attaining social homogeneity. If the research is conducted in different countries where certain sets or clusters of occupations represent a similar social group, selecting participants based on their occupation can also be a fruitful way of creating comparable and socially homogenous groups (Duchesne et al. 2013). Nevertheless, selecting homogeneity criteria remains research topic dependent, especially as it can often involve a trade-off between standardization on some dimensions deemed relevant with regard to the topic of discussion versus non-standardization on others.

Another important consideration is whether to work with a group of strangers who are invited for the purpose of the research or to work with natural discussion groups—for example people who regularly interact such as a group of friends or co-workers. Using natural discussion groups increases the chance that pre-existing power relations and group dynamics, that are usually hidden from the researcher, influence the content and the nature of the discussions. (e.g. Morgan 1996). Studies using natural discussion groups are valuable in cases where the research topic is related specifically to relations within these pre-formed groups. However, when it comes to using pre-formed groups in cross-national focus group research, researchers should pay particular attention to standardization and to making these groups as functionally equivalent as possible.

Group size is mainly dependent on how complex or contentious the topic of discussion is. Topics that are likely to create a lot of discussions and (emotional) involvement from the participants require fewer participants compared to more neutral topics (Morgan 1996). An additional consideration is whether the goal of the research is to get a wide variety of opinions versus in-depth opinions, with the latter requiring smaller groups that do not impose time constraints. The lower and upper limits in terms of group size in general will not go lower than four or exceed ten (e.g. Duchesne et al. 2013; Krueger and Casey 2014; Morgan 1996). Depending on how the groups are composed, there needs to be a decision on how many focus groups will be conducted in each country. Reaching saturation, the point where collecting more data does not add anything new (Bowen 2008), is normally used as a criterion in qualitative research for the amount of data that needs to be collected. There are rules of thumb, for example using between four and six groups usually will provide saturation although the complexity of the topic and desired depth of opinion have to be considered as well (Morgan 1996). An iterative process of adding more focus groups to reach saturation is less feasible in cross-national comparative research. Based on the research question, the relevant selection criteria (e.g. age, socio-economic status) for the participants should be determined, which in turn will provide the initial division of groups. For the number of groups, “better safe than sorry” provides a good rule of thumb, with additional data being less problematic compared to not reaching saturation.

Recruitment of the participants forms the bridge between the sterile, theoretical decision about the size and composition of the focus groups and implementing these decisions practically. More often than not, recruitment will prove to be a tedious task requiring some degree of flexibility between methodological rigor and pragmatism. In general, potential focus group participants can be reached in many different ways, ranging from convenience sampling through existing networks or groups of people, to highly stratified sampling of individuals with very specific characteristics (e.g. Stewart and Shamdasani 2014). Whereas many practical concerns regarding recruitment such as (self-) selection bias, contacting the participants, scheduling, and sending reminders are important, they are extensively covered in existing literature about focus group methodology (Morgan 1996; Stewart and Shamdasani 2014). Standardization again is the leading principle when applying all of these considerations in cross-national focus group research. For example, if incentives are used to attract participants, the value of this incentive has to be adapted to local standards in the country where the focus group is being conducted.

Similarly to other forms of qualitative sampling, it is important to define the sample universe and sampling strategy before turning to the specific recruitment strategy (e.g. Robinson 2014). A specific question that arises with conducting focus groups in multiple countries is whether different recruitment strategies can result in focus groups that are similar in terms of certain key participant characteristics. Some form of pre-screening is needed, based on the sampling criteria, but also on other characteristics that could influence the discussions. For example, on the topic of political participation it is possible to create social homogeneity by forming groups based on education level, but screening here should include additional relevant characteristics such as political interest, political ideology, or past political behavior. If this longer list of selection criteria is administered in each country in the same way, and decision rules for the selection of the participants on the different criteria are established, the recruitment of the participants can be done online or offline, using recruitment agencies, or advertising the focus group discussions in public places.

The decisions about who participates in focus groups and how to recruit these people are crucial in a comparative cross-national focus group project. Nevertheless, more practical concerns in the implementation of this design can have a substantial effect on the quality and standardization of the focus groups. Stewart and Shamdasani (2014) provide an excellent overview of the practical considerations regarding arrangement of the groups and recording the discussions. For cross-national focus group research however, every practical decision should stem from a two-step decision making process: how does it follow from the research question or goal of the research, and can this be implemented in the same way in each country?

The second aim of this article is to illustrate how the proposed methodology for using focus groups in a cross-national comparative way can be useful for studying protest. We argued that focus groups provide an insight into the process of intersubjective meaning-making of attitudes (Stanley 2016), adding to quantitative measures of motivations to join political protest. Moreover, in line with Lup (2015), we argued that cross-national focus group research allows for an examination of cultural differences in how these attitudes are negotiated. We will illustrate both of these points with excerpts about street protest taken from the POLPART focus group discussions in different countries, where participants discuss the same question in each country about what citizens can do about important issues in society (see “Appendix 2”, question 2). All the names that are used below, are fictitious names and “M” is used to indicate the moderator.

In a political arena where protest is becoming an increasingly conventional strategy to influence politics on a wide variety of topics, focus group discussions provide us with a better understanding of how citizens’ attitudes towards protest are being negotiated and shaped in social interactions, which in turn could influence their propensity to participate in protests. We proposed a methodology for cross-national focus group research, serving two related goals. First, focus groups have seldom been used cross-nationally and we argue that especially due to the contextual influences on how people form attitudes (Hollander 2004; Stanley 2016), using focus groups cross-nationally enhances our understanding of meaning-making and attitude formation across cultures. Secondly, we illustrated the methodology we put forward with findings from a cross-national research project, in which 80 focus group discussions were conducted in nine different countries in Europe and Latin America. We used excerpts from the focus group discussions to demonstrate that attitudes towards protest are not always stable (Fishkin 1991), but sometimes shaped by social interactions and the country’s political culture and historical trajectory.

We see our proposal for a cross-national focus group methodology as a starting point for increased scholarly efforts in this direction and we believe that the principles we outlined can be applied more broadly in comparative focus group research. Despite the focus on attitudes towards political protest, research questions on many other topics can be answered using a cross-national methodology that allows for a meaningful comparison on complex topics of discussion. Similarly, although we focused on the comparability of the focus group participants across groups and countries, we believe the guidelines introduced here are also of interest for researchers seeking comparisons in other domains, for example across language groups, ethnicities and regions. Standardization of the research design can in these instances warrant a meaningful comparison between the group discussions without losing the depth of understanding of social phenomena that focus groups offer.

Future developments of cross-national focus group research could benefit from considering the transformative effect that engagement in focus groups can have on its participants. How much does participation in these collective group discussions alter the participants attitudes, beliefs, or norms? Do people change their behavior after engaging in critical discussions with their peers? We think that using focus group research has tremendous potential in capturing these processes that are relevant for different disciplines in social science. In a time where comparative and, in particular, cross-national research is becoming more and more important, we hope that our proposed methodology provides a first step in applying focus groups comparatively across diverse contexts more regularly.