Version classiqueVersion mobile

Stratified Belonging, Layered Subjectivities

Zong Yao Edison Yap

3. Methodology

Texte intégral

1This dissertation is based on an ethnography of five refugees and ten other migrants who had lived in Berlin prior to the 2015 “refugee crisis”. To a lesser extent, it includes secondary sources from research and media publications to substantiate the claims I make. While the analytical focus is on the experiences of refugees in Berlin, I have intentionally included the voices of other migrants to inject greater nuance to the subjectivities of refugees.

  • 1 The inverted commas are necessary since I cannot claim to be an expert in Middle Eastern Studies. T (...)

2Before delving into an in-depth discussion on the methodology, it is important to note how my positionality played a significant role in the research process. My own struggles as a non-white, queer, ethnically Chinese Singaporean from a lower-income family as well as my advantage as someone with considerable1 knowledge of Iranian and Arab cultures and language have helped me connect with my participants and navigate the semi-structured interviews in multifarious ways.

3This chapter is organised as follows: the first section describes the sampling method, giving particular focus to the difficulties I faced when trying to secure interviews with refugees: this entails delving into my positionality and background and how they have facilitated or impeded my access to refugees. Subsequently, it discusses the semi-structured interview process and how my positionality has played a role in driving the process. The second section discusses grounded theory, its advantages, and the data processing method I used. The third section highlights the “mixed method” – of ethnographic and grounded theory methods used – and examines the limitations of my methodology.

3.1 The Sampling and Interview Stages

3.1.1 Language

  • 2 While all interviews were conducted in English, not all respondents were able to articulate their i (...)

4Despite having lived in Berlin for approximately three years at the time of research, sampling proved to be much more challenging than I had envisioned. I had expected my elementary proficiencies in Arabic and Farsi as well as my regional specialisation in Middle Eastern politics to help me secure a degree of trust and interest among refugees – but I was mistaken. Furthermore, my intermediate-advanced proficiency in spoken German at that time meant that I could not deftly conduct an interview in German, and that excluded from my sample refugees who could not speak English but who had reasonable proficiencies in German. I also did not have the requisite resources to hire an Arabic-English translator. As such, this introduces a degree of bias in my sample of refugees because all participants I interviewed spoke English2 – though my knowledge of German was instrumental in helping me – and that could mean that they were privy to certain privileges or advantages related to their family income, education, cosmopolitan outlook, etc.

3.1.2 Sampling Method

5At the beginning of my research, I frequented a queer club in Berlin called SchwuZ where LGBTQ refugees would visit weekly without fail; in the smoking room, I would use the smoke break or the offer of a lighter to strike up a conversation in Arabic or Farsi with some of them – as expected, my knowledge of Arabic and Farsi (and more so, as an East Asian-looking person who knows both languages) would take them by surprise and pique their interest in me. Following that, a conversation regarding my interest in the Middle East would usually ensue and I would be posed the question “Was machst du in Berlin?” (What do you do in Berlin?) – to which I would respond that I was on a study abroad programme at a local university and was also conducting my thesis research on refugee issues in Berlin. I did not ask these refugees for an interview the first time I met them because I wanted to build a degree of trust with them. It was usually during the second meeting at the club that I would ask if they would be willing to be interviewed for strictly no more than 90 minutes. All the refugees would enthusiastically agree to an interview – but to my disappointment, after having fixed a concrete time and place for our interview, the refugees would either block or delete me on social media and/or WhatsApp (a popular social networking and messaging application).

6Perplexed and disheartened, I confided in an Arab friend the following week and was told by him: “Arabs are afraid to say ‘No’. So they readily say ‘Yes’ to avoid rejecting you.” I corroborated that with my other Arab friends and was also told that there’s a general tendency in the Middle East.for people to do that. My friend’s remark reminded me of an Iranian civility – which I experienced on a daily basis while based in Iran for three months in 2015 – that is called “Tarof”. This means that “people refuse when they want to accept, say what is not meant, express what is not felt, invite when it is not intended, replace bad news with false hope. By doing so, they try to say what they ‘wished it were’ – without ever admitting that it isn’t” (Valle, 2016). Cognisant of that, I returned to the club the following week to check with other refugees whom I had not spoken to and to ask if there could be any other reasons for my streak of bad luck – to that, the refugees admitted that many of them had been interviewed many times (i.e. the problem of being over-researched) and they no longer wished to be reminded of their traumatic journeys to Europe or their experience of living in Berlin.

  • 3 All names used in this dissertation are fictional as the respondents wished to remain anonymous.

7The above insights drove me to reconsider my sampling method. Eventually I decided to use the snowball sampling method, which helped me secure five refugees and ten other migrants for semi-structured interviews. The snowball sampling method was the best method of securing respondents since a degree of trust had already been established between my respondents and my own friends. Despite being able to secure a reasonable number of interview respondents for this dissertation, the sample I gathered nonetheless contained significant biases, given that there exists a skewness in my social circle since most of my friends are liberal-minded, queer-friendly, and had completed higher education: firstly, all five refugees were between the ages of 22 and 29 years of age and are pursuing higher education in Germany (with the exception of Wathira;3 the other four refugees were pursuing their master’s degrees). Secondly, and as mentioned above, all refugees were fluent in German and could speak English reasonably well; thirdly, all refugees were reasonably integrated in Berlin or Germany and were all ambitious and had a positive outlook on their future in Berlin or Germany. Regardless of the sample biases, findings from my interlocutors still offered rich insights into their struggles (as well as that of other refugees) with discrimination and integration, as will be demonstrated in Chapters 4 and 5.

3.1.3 The Interview Process

8A total of 15 semi-structured qualitative interviews, each lasting 90 minutes (with the exception of Khaled, whose interview took three hours since he did not want his interview to be recorded) were conducted over a period of two months from March to early May 2019.

9Having conducted the literature review, I organised the interview into five broad themes: (i) settling down in Berlin, (ii) experience of discrimination in Berlin, (iii) social integration in Berlin, (iv) opinions about refugee or migrant integration in Berlin, and (v) contact with family back home. Each theme consisted of a few questions, but when respondents offered interesting and salient information, I would allow the interview to deviate from the planned structure so as not to render the conversation too rigid or mechanical. I also made it explicitly clear to refugees that I would not be asking any questions relating to their struggles back home or their journeys to Europe because my research focuses on their settlement issues after arrival in Germany and I also did not wish to dig up uncomfortable or traumatic experiences from the past. This however does not mean that their positionalities back home were excluded from our conversations, because at different junctures of the interviews, my interlocutors did reflexively offer information about their family life, beliefs, and so on.

10In addition to this, I deliberately left the meanings of “discrimination” and “integration” undefined during the interview because – in the spirit of grounded theory – I wanted to allow the respondents’ experiences to drive the conceptualisation and theorising process and not be confined to the boundedness of conventional definitions. Related to this is how my positionality has been key to driving the interview process. For instance, in the case of Wathira, when I asked if she had experienced any significant discrimination in Germany, a long silent pause followed; in this case, as with similar instances with other interlocutors I allowed the pause to endure as a means of a silent probe to offer them time to reflect. When this did not work, and only as a last resort, I offered anecdotes of my experience with racism and homophobia to see if it would trigger mental associations in the respondents’ memories. The silent probe was the most useful technique when it came to asking the respondents to reflect on how the composition of their social circle(s) had changed since their arrival in Berlin; most of my respondents expressed that they found the exercise rewarding for it helped them identify turning points in their settlement trajectories up till the present time.

3.2 Grounded Theory and Data Processing

11Grounded theory method “consists of systematic, yet flexible guidelines for collecting and analysing qualitative data to construct theories ‘grounded’ in the data themselves. The guidelines offer a set of general principles and heuristic devices rather than formulaic rules. Thus, data form the foundation of our theory and our analysis of these data generates the concepts we construct” (Charmaz, 2006, p. 2).

12Hence, by using the grounded theory method, I allowed data to drive the research and analysis. This avoids the situation where the paper could be defined and constrained by my own assumptions and preconceived notions of discrimination and integration as well as the insights gained from the literature review – in particular, the insights offered by Ruth Mandel, since this research to a large extent adopts her style of ethnographic study (albeit in a much shorter time frame and focusing mostly on refugees in Berlin). Hence, in short, a grounded theory approach offers room for the voices and subjectivities of the refugees and other migrants interviewed to challenge, refine, and reinforce the findings garnered from the literature review, thereby providing me with a more nuanced understanding of the subject matter.

13As mentioned in the preceding sub-section, with the exception of Khaled, all interviews were audio recorded and were subsequently transcribed verbatim. The first few interview transcripts were uploaded onto the qualitative data analysis software NVivo. Using NVivo, I coded the transcripts as I ran through the data line-by-line and developed categories (i.e. initial coding). Using the insights gathered from the first iteration, I returned to my interview themes and questions which I had developed based on my literature review and refined their organisation to render them more targeted and coherent. I then proceeded with the remainder of my interviews using the updated interview questions and transcribed and coded them in a second iteration of coding. At the end of this second iteration, the codes reflect that a reasonable degree of saturation has been attained, i.e. no newer concepts or insights emerged in the later interviews. In my last iteration, I organised the analytical codes – deleting some, adding others – and organised them into broader themes of “discrimination” and “integration”. Within these two general themes, sub-themes were generated, such as “factors fuelling discrimination”, “discrimination by Germans”, “discrimination by older Arabs”, and so on. With the broad structure in place, I was then able to structure my findings section according to the broad conceptual framework developed using NVivo.

3.3 Limitations of the Methodology

14The previous sub-sections have highlighted some of the significant sample bias that pervades this research. Here, it is necessary to highlight additional shortcomings in terms of what this research has missed out on.

15Given that only five out of fifteen of the interview respondents were refugees from Syria – of which two, Khaled and Sam, were Palestinian refugees who grew up in Syria – the research would have significantly benefited from (i) a larger number of respondents and (ii) refugees from countries outside of Syria, e.g. Iraq, Iran, Afghanistan, and non-Middle Eastern countries such as Eritrea. While I consciously sought refugees from outside of Syria, their numbers are low since most of the successful asylum seekers are Syrians (Deutsche Welle, 2017). The lack of insights and subjectivities of refugees outside the Syrian refugee community is especially salient since it leaves out perspectives which could be related to their nationality, ethnicity, language, or even religion. This skewness in nationality also applies to the migrants: most of them were in their twenties or thirties – with the exception of Paul, who was in his forties – and they were liberal-minded, reasonably integrated residents in Berlin.

16Equally important is the lack of voices of referents (Germans, Turks, or older Arabs) with which refugees associated experiences of discrimination. The findings have documented the experiences from the point of view of some refugees and migrants, but have left out the voices of those who have been blamed for discrimination – this is a significant shortcoming for, if the voices of refugees reflect a high degree of heterogeneity and complexity, perspectives gathered from Germans, Turks, and older Arabs would likely offer the same nuance and heterogeneity. Future research would benefit immensely by filling this lacuna.

  • 4 Due to the restrictions in time, resources, language ability, and the scope of this paper, I was un (...)
  • 5 Terrorism suspects have posed as refugees to enter Germany and many refugees still experience signi (...)

17Reflecting on the research process,4 I would have probably conducted more interviews with the staff and volunteers working at refugee-related organisations in Berlin, for as Dana’s insights have shown (see Chapters 4 and 5), they possess a wealth of information on the struggles refugees face. Moreover, given the higher number of refugees and plethora of subjectivities that they deal with on a regular basis, they offer a bird’s-eye view of the subject matter which could have rendered my findings much more robust when juxtaposed to the individual insights offered by refugees. I would also have layered individual interviews with focus group sessions: the interactive aspect of this method would allow competing perspectives to surface; it also juxtaposes one participant’s positionality with that of other participants – which might elicit very different insights from the juxtaposition of my positionality with that of refugees. However, refugees might feel apprehensive about revealing too much information about themselves for fear of reprisal from other participants who could be more conservative5 or who belong to a particular sectarian group.

18Lastly, it is important to note that this dissertation was conducted using a “mixed method” of ethnography (as a data collection method) and grounded theory and NVivo (as a data processing method). While ethnography enabled me to obtain as nuanced an understanding of the subject matter as possible, grounded theory and NVivo allowed me to present patterns and themes in a systematic fashion. Admittedly, forgoing entire narratives does erode the depth of my “thick description”, but this I believe is a necessary compromise to accommodate as many viewpoints as possible in a short paper (Geertz, 2008).


1 The inverted commas are necessary since I cannot claim to be an expert in Middle Eastern Studies. The many months I spent in Haifa, Israel, and Tehran, Iran, as well as my years of working at a Middle East research institute in Singapore, have however provided me with a relatively more nuanced understanding of the cities and their politics and cultures.

2 While all interviews were conducted in English, not all respondents were able to articulate their ideas in this language. For a couple of them, German – other than Arabic – was the language of choice to express certain words. During the analysis and coding stage, I had translated these German words to their English equivalent.

3 All names used in this dissertation are fictional as the respondents wished to remain anonymous.

4 Due to the restrictions in time, resources, language ability, and the scope of this paper, I was unable to conduct a much more extensive research.

5 Terrorism suspects have posed as refugees to enter Germany and many refugees still experience significant fear of extremists living among them (Nabeel & Bhatti, 2016; Smale, 2016).


Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search