Homogeneity of Focus Groups as a Pathway to Successful Research Findings? Methodological Notes from the Fieldwork

Homogeneity of Focus Groups as a Pathway to Successful Research Findings? Methodological Notes from the Fieldwork concerning the efficiency of American war propaganda). Nowadays, it is also very often used as an instrument of research on political issues, in healthcare research, or while projecting solutions in policy-making processes. During the following decades this kind of study has been hailed and criticized by various scholars involved in methodological debates over the qualitative paradigm of social research.

Since the 60s, social sciences have witnessed a substantial increase in number of research efforts conducted and published using the qualitative methodology at the expense of quantitative studies (Fielding 2010:132). Along with this processes, Focus Group Interviews (hereafter: FGI) have become increasingly popular and an established technique among researchers, to the joy of some scientists and disappointment of others. From the beginning of the 90s, the bookshelf with elaborations and studies concerning unique methodology and all possible aspects of this kind of interviewing is getting more and more rich (Madriz 2000:835-836).
FGI is a research technique that is based on a direct interaction among participants of a group discussion on the topic defined by the researcher (Morgan 1997;Krueger 1994). The interview is described as "focused," firstly, because it is concentrated on one crucial theme, and secondly, it demands some specific collective actions of the respondents. The latter feature distinguishes this research technique from individual interviews. Due to the effects of synergy and group dynamics, FGI provides the researcher with an insight into respondents' attitudes, their language codes, priority values, and cognitive schemes as they are constructed in a di-rect and lively interaction. It also enables access to knowledge about opinions as they are formulated in a discursive manner during the exchange of ideas in the course of a group discussion (see : Kitzinger 2004;Kutsar, Strömpl, Trumm 2006;Daniłowicz and Lisek-Michalska 2007;Belzile and Öberg 2012).
Leaving aside interesting and stimulating discussions on various aspects of group interviews and usefulness of its applications in social sciences, this paper will concentrate instead on some elements of the research process which every scholar takes into account while planning his/her tasks. The process of selecting respondents and composing groups is important for the qualitative research. Precise design is essential for the quality and aptitude of the data gathered. In case of collective efforts like group interviewing improper composition of the sample could easily hamper the efforts. The procedure of purposeful sampling of specific informants, which is the core way for selecting the participants for qualitative inquiry, rarely allows the sampling with replacements, contrary to random sampling in a quantitative research which frequently uses substitutions.
Issues concerning homogeneity/heterogeneity of FG will be discussed in reference to other characteristics of this kind of research (effects of synergy, groupthink), as raised in the large body of previous research. Special attention will also be paid to the question of familiarity among the participants in the FG and the consequences coming from the nature of their pre-existing relations.
The particular study undertaken by the sociologists from University of Lodz will be used as an example

Słowa kluczowe
The paper approaches the topic of sampling in the qualitative research design. Particular attention is paid to the composition of the focus groups in two important aspects: the homogeneity of the participants in terms of their positions in socio-occupational hierarchy and the pre-existing relations among the participants in the single group. The latter issue is closely and directly intertwined with the former, in the case when the informants are approached in their institutional setting, and the research is conducted in a relatively small community. These problems are discussed in reference to the body of available methodological studies. The example of the research undertaken in aforementioned circumstances is subsequently presented to advocate the homogeneous composition of the focus groups while reconstructing the collective viewpoints of representatives of a particular level of public administrations. Avoidance of power relations within the groups proved profitable and effective. Substantial differences between the groups were easily identifiable in spite of the pre-existing relations among participants, which were unavoidable in this very case. Since the early 40s, thanks to Paul Lazarsfeld, founder of the Bureau of Applied Social Sciences (Barton 1979) and his most famous employee, Robert Merton, who is sometimes credited as the "father of Focus Group" (Kaufman 2003), group interviewing has been present in the methodological repertoire of social sciences. It has been firstly designed for the purposes of evaluative marketing research (also ©2014 PSJ Tom X Numer 1 8 study. Methodologists have various doubts concerning the latter issue. On the one hand, they claim pre-existing relationships among participants could sometimes hamper the rule of the same conditions for each respondent and impede analysis of intersecting influences of the competing opinions and thoughts of the respondents expressed during particular discussion (Kitzinger 2004 Warr was referring to her own research among people who not only knew each other privately and shared similar experiences but they have also originated from the homogeneous social and economic background. These features allowed for more dense interactions between the participants during the FGI, including very frank and personal references to common past. The familiarity between them resulted in lively discussions where they have mutually challenged their opinions in a way inaccessible for the researcher -outsiderand the trust between them was more probable due to the homogeneous nature of their positions. Thus, FGIs could be selected from the qualitative inquiry toolbox when looking for the reconstruction of some common attitudes and consensual perceptions among the participants sharing some common ground, as well as in the research which aims at challenging their assumptions and stimulating "argumentative interactions" (Kitzinger 1994). Kitzinger also underlines the fact that the presence of the others who share common experiences and knowledge on the other participants encourages the exchange of more trustful thoughts than in the case of individual data gathering with separate respondents: [n]ot only does safety in numbers make some people more likely to consent to participate in the research in the first place…but being with other people who share similar experiences encourages participants to express, clarify, or even develop particular perspectives. (1994:112) There is also a longstanding controversy concerning homogeneity in reference to almost all potential variables (Stewart and Shamdasani 1990:42-43;Morgan 1997:37;Rabiee 2004:656). Whether women and men should be gathered together while discussing gender-sensitive topics? Whether the discussion, or even an argument between them, of the research where specific sampling and composition of group for FGI sessions lead to profitable and interesting results. The results of this case will be used only to exemplify the consequences of ho-

Selecting Respondents for the FGI -Methodological Questions and Doubts
In most cases researchers point out the following features of the group as the specific social entity: direct communication, reciprocal influence, interaction, interdependence, mutual reactions, self-identification, structure, psychological meaning. The most important ones, while planning the FGI research, are the two remaining: similarity of the members of the group, which in case of meaning has at least one common characteristic and common experience. The latter one has two meanings: experience gained before the group was formed and "common fate" experienced by the participants during FGI session.
Sampling for the FGIs is always purposeful and theoretically motivated, determined by the topic of the study (Morgan 1997:35). As qualitative research does not try to provide results which could be extrapolate for a wide population, researchers are able to imply "convenience sampling," to use the phrase of Stewart and Shamdasani (1990:53).
Instead of any kind of randomization, groups have to be homogeneous in reference to at least one characteristic, which is indispensable for the proper procedure of recruitment for the group session.
The common variable may be connected with any kind of socio-demographic characteristic, as well as with the individual characteristic of the potential participants regarding the topic of the meeting (Bloor et al. 2002:21-22).
The common assumption that in order to achieve advantageous results the groups must be composed of total strangers is one of the myths describing qualitative interviewing. To the contrary, Bloor and colleagues prove that rarely there is a need for this kind of approach and avoiding the pre-existing relationships in FG research as a general rule is both unnecessary and improbable (2002:22). In this respect, he follows Kitzinger who claims: "[a]bove all, it is useful to work with pre-existing groups because they provide one of the social contexts within which ideas are formed and decisions made" In the case of focus groups which are homogeneous and the participants know each other in advance of the session, these issues should be taken into ac-2 One of the potential solutions that could be implemented in such a situation is introducing into the group a provocateur, collaborator of the researcher, who, by expressing intriguing statements, will attempt to stimulate discussion. This kind of technique was used in an interesting study on the tolerance among Polish students, conducted at the Warsaw University, which revealed the real anti-stranger prejudices and bias well-hidden under the facade of tolerant and politically correct statements (Lisek-Michalska 2007:57). count ex ante, while planning the research, and ex post as the potential explanation for some specific attitudes, opinions, and behaviors presented by respondents during the session. 3

PROFIT Project -Background Information
This paper is not an attempt to adjudicate the controversies concerning the sampling procedures.
Showing the example of particular research where homogeneous focus groups were conducted and purposeful sampling was applied, I wish to present: • how the sampling was connected with the general conception of the study; • what assumptions stood behind the specific composition of the focus groups; • what results were obtained.
It is based on the research findings collected during the realization of the international project PROF-IT, 4 which dealt with the issue of intergenerational 3 In the case of the topics which are not sensitive (e.g., in marketing research), homogeneity in composition of groups increases the comfort of participants, stimulating more free-flowing discussions and allowing researchers to make more reliable and justified comparisons between data collected during diversified sessions. If the groups are more heterogeneous, the number of criteria which define the participants is smaller and therefore, it is advisable to conduct more group interviews to get a more detailed picture of the studied subject (Daniłowicz and Lisek-Michalska 2007:25). 4 PROFIT (Policy Responses Overcoming Factors in the Intergenerational Transmission of Inequalities) is financed by 6 th Framework Program of European Commission under priority 7 -Citizens and governance in knowledge based society; contract no: CIT2-CT-2004-506245. It was coordinated by the University of Lodz and personally by Professor Wielisława Warzywoda-Kruszyńska. More information about the project, as well as reports and publications based on its outcomes could be found at: http://www.profit.uni.lodz.pl/. allows the researcher to draw the line of distinction between the ways of thinking dominating in each category? Or, is it more effective to search for some patterns of thinking typical for each of them while interviewing each gender separately? Should we bring together representatives of various social strata representing social differentiation within a given population, or shall we rather comfort respondents allowing them to speak with people of similar backgrounds? It is particularly important to remember that the researcher is not always (or never is) able to recognize ex ante which topics could prove sensitive for the participants (Bloor et al. 2002:22). There are important and justified arguments for each of these positions, ranging from ethical (uncomfortable situations among participants) to strictly practical ones (unity in experiences and conformity of thoughts could diminish the intensity of the discussion and too differentiated group could lead to the chaos or conflicts during the session). Recruitment of participants via a pre-existing formal group could also be easier and more convenient due to reducing attrition rates.
According to other scholars, researchers can benefit from the homogeneous FG as they: can scrutinize how the participants are developing certain themes in the discussion and how they are reflecting on and developing their understanding and anchoring of individual experience against the sum total of the other participants' arguments, experience, and knowledge. In the analysis, the researcher may explore how the elaboration of individual accounts helps in forming a web of socially shared knowledge emerging through group discussion. … The group is a think-group, in which cognition is going on in the minds of members, but this happens largely in and through the interaction. Individuals with some kind of common background stimulate each other to develop thoughts and arguments. In this process, ideas interpenetrate and often contradict each other…expressing disagreement may also be part of the learning process, as participants challenge each other, defend their arguments, and at times modify their viewpoints. (Wibeck, Abrandt Dahlgren, Öberg 2007:255-260) It is especially fruitful when the research is con- PROFIT was a comparative research project providing deep, idiographic insight into various aspects of inequality reproduction, however, for the purposes of this sketchy paper, only small excerpts of the data collected will be presented.
As it is presented in the table above, FGIs were used in the project to study opinions, knowledge, and beliefs regarding the issue of inheritance of inequalities possessed by the local stakeholders from the purposely selected middle-sized town (50-80 inheritance of social inequalities. The study was multi-level and multi-method and the research team collected data at diverse levels, using various techniques (for more data about the project objec-tives and methodology, see: Warzywoda-Kruszyńska and Rokicka (2007). A general outline of the project content and its phases is shown in the scheme below. Most participants knew one another already; thus, they sometimes referred to common experiences.
They recognized themselves well in their professional roles, knowledge about private encounters and informal relations pre-existing prior to the meetings were inaccessible to the researchers. Homogeneity allowed researchers to identify certain features emblematic for the specific work ethos and way of thinking in each of these professional groups covering specific conscientious convictions, as well as unspoken stereotypes, language codes, and clichés. Thanks to the effect of group synergy some elements typical for the groupthink, as described by Irving L. Janis (1982), were also found during analysis of the transcripts, some of which will be presented separately for each of the FG.
Hardly ever did a controversy occur among the interlocutors, with an exception of the group of councilors whose discussion clearly reflected political divisions and conflicts. These "argumentative interactions" and differences in opinions plainly corre- Interviewees also pointed to more and more frequently occurring pathological phenomenon, which they describe as "purposeful social inefficiency": with an intention to claim that -because of benefits and social services -parents deliberately deprive their children of intellectual and social development possibilities: [w]e have many such cases where children are not in fact sick, or we can see that they are not really sick, but they are neglected because the mother fails to work with the child, to teach him to sit, or to walk. There are two-three-year-old children who hardly walk, hardly sit, or talk… Not because they are mentally or physically retarded, but just because they are neglected. This is simply nurturing the sickness. (Social Worker, FGI 1) The analysis of the participants' statements allows concluding that the clients' demanding attitude, highlighted by the respondents, is coupled with a specific "demanding attitude" of the institutions implementing state social policy. Their employees, dealing directly with clients on a day-to-day basis, often rate their professional group underpaid. They complained that their salary was not much higher than the social benefits their clients received: "[i]t's just like us, we get our pay to our accounts and the lady in the bank asks if we are social services clients.
That's right! But, this is a common thing for us. This embarrassment in the bank" (Social Worker, FGI 1).
Lack of satisfaction with work is not only due to insufficient pay but also the sense of inefficiency, helplessness, and lack of possibility to use the knowledge possessed in adapting measures to the specific local conditions (the feeling that the social worker's role boils down to mere distribution of the state funds). Frustration and professional burnout were also perceptible among other "frontliners": [b]ut, we are never appreciated. Teachers are always the worst, the worst caste of people just because they are teachers. Only that all these wise people were pre-

Focus Group Interview 4 -Local Politicians
Main Feature of the Session: "Indifference/Lack of Concern" The main features of session three were also observed during the session with local politicians. Their attitude was characterized -as showed the analysis of transcripts of the interviews -as "indifference/lack of concern," which may also be understood as a "lack  Again, a high level of trust and common experiences allowed the reconstruction of the viewpoint of those involved in the third sector in the local environment under scrutiny and partial recognition of the circumstances leading to low levels of various actors' engagement in this kind of activities.

Main Feature of the Session: "Official Optimism"
The analysis demonstrated differences in attitudes of Thirdly, the composition of focus groups could be one of the main factors defining differences between the outcomes of the interviews within one research project. Therefore, various patterns of participants' behavior deserve particularly careful attention of the researchers. These issues should be taken into account not only while designing the FGI research but could also be treated as an important separate topic of investigation into socio-psychological aspects of communication within the group.
All respondents were following argumentation typical for the national political elites, rarely referring to the unique and concrete examples from their own areas of activity. The in-depth analysis provided some evidence that groupthink presented by members of this category could be influenced rather by the political discourse at a top level of national policy than local experiences. Certainly, this impact is connected to the way media are addressing and defining social problems.

Summary and Conclusions
The outcomes of the study gave some important answers to the questions concerning the quality of the governance at the local level and factors influencing effectiveness of the social policy actions undertaken by various actors at the local level. Assessment of these results does not belong to the purposes of this paper. However, it has to be said that notwithstanding differences noticed in each of the groups and a common strategy, which could be described as "blame the victim," were observed among all of them. It underlines psychological and cultural features, typical for the group of poor people who are not willing to cope with their situation using socially accepted means and generally deserve their own status quo. This kind of attitude presented by people responsible for projecting and implementing the social actions aimed at supporting all needy groups does not allow being optimistic about the quality and outcomes of these operations.
As to methodological issues, decisions on separating the four categories of local stakeholders into differentiated groups turned out to be beneficial for the study.
It allowed researchers to point out and enumerate some crucial and important distinctions between the groups, and also to analyze their way of thinking. All positive features of homogeneous FG, as described in the first part of the paper, were observed. Hence, it was possible to identify some features typical for the specific groupthink of each of the categories under study. Drawing a clear division line between various groups of stakeholders was found as a recommendable solution in this kind of policy-oriented FGI study, where methodological issues are often neglected, not being the main sphere of researchers' interest.
It seems evident that while planning qualitative research within the institutional setting in a relatively small community where a dense network of formal and informal intertwined interrelations is affecting them, the special attention needs to be drawn to the composition of the FG.
In the aforementioned study, due to the nature of the researched phenomena (local social policy), it was impossible to compose groups of the participants who would not know each other prior to the group interviews. However, avoiding power relations between those invited for each of the groups allowed increasing the comfort of the participants, which probably also enhanced the chances to gather rich and reliable data provided by participants who felt more self-assured and confident about their expertise in a debated field and lack of threat coming from the eventual disagreement with those located higher within the institutional structure.
There is always an unavoidable risk that the pre-existing mutual acquaintanceship of the participants, even in the absence of the power relations, could increase their consistency with the group. Knowing