This overview is aimed at academics, practitioners, and policymakers to help with understanding refugee integration research. We describe the methodology of the RIT project, why we use certain methods over others, and the advantages and disadvantages of each approach in order to support other researchers in this field.
Case studies are a type of research that aims at providing a “comprehensive description of an individual case.” A “case” could be any unit of analysis: an individual, a group, or a phenomenon. In RIT’s cases, we look at the unit of the urban area experiencing the phenomenon of forced migrant integration.
Case studies have a range of different temporal scopes. RIT’s case reports are “diachronic,” meaning they look at changes in urban spaces over time, using original data collection from the present supplemented by archival data and communities’ recollections of past events. In this way, each case explores how urban areas have changed throughout history to adapt with migration.
A group of case studies may be analyzed collectively to observe broader trends. RIT analyzes nested cases—looking at multiple neighborhoods within the same urban area—and parallel cases—looking at multiple urban areas all around the globe.
There are several strengths of case studies that make this approach effective for the RIT project.
First, case studies are valuable for exploring complex processes that may have multiple perspectives, and for yielding findings that are rooted “in real world settings.” Refugee urban integration is influenced by a long list of complex factors and may be described accurately from many different perspectives between and within migrant and host populations. There might also be very stark differences between the way integration is described in law or in theory versus how it is experienced “in real world settings.”
Second, case studies are complementary to other types of studies because they give context, open up the opportunity to discover new theory, and allow for testing existing hypotheses with real world observations. In the study of refugee integration, there is a wealth of statistical method research conducted by large humanitarian and development organizations, governments, and consultants. The RIT project uses a series of case reports to complement this existing data, building a grounded understanding of refugee integration (more on grounded theory here), adding context, testing current theories, and discovering “new and unexpected results.”
Finally, case studies are “especially valuable in practice-oriented fields,” making this approach useful for the RIT project’s goal of producing findings that will be useful for practitioners and policymakers, not just academics.
The Feinstein International Center’s literature review on localization in humanitarian action defines localization based on de Geoffroy & Grunewald’s work (2018) as the “collective process involving different stakeholders that aims to return local actors, whether civil society organizations or local public institutions, to the center of the humanitarian system with a greater role in humanitarian response.” Local actors are defined by Gingerich and Cohen (2015) as those geographically, socially, ethnically, religiously, or nationally proximate to the area of concern. Localization contrasts with the majority of research on forced migration and integration that tends to rely on external, international experts who are not present in the area being studied.
Localized methods have both advantages and disadvantages. Advantages of localized research include improved access to the population (especially to illegible areas or hidden populations), contextual knowledge, native language skills, accountability to the population being studied, and lower cost and resource demands relative to imported international researchers. Localization has particular advantage for collecting data on subjective measures and indicators as well as qualitative data that is highly culturally specific.
Disadvantages of localized research includes the partiality of local actors and the effect on power balances caused by hiring some locals over others, and the fact that localized research will inevitably amplify some of the population’s voices over others.
The RIT project chose to use localized methods to bring a depth of cultural, historical, and linguistic contextual knowledge to findings. We also have chosen to use localized methods to follow the humanitarian localization agenda founded in the United Nations Sustainable Development Goals (SDGs):
“Localizing development is a process to empower all local stakeholders, aimed at making sustainable development more responsive and...relevant to local needs and aspirations. Development goals can be reached only if local actors fully participate.”
The Site-Based Approach
The site-based approach uses local, context-specific recruitment of research participants. In qualitative studies, the use of a site-based approach is beneficial for selecting and recruiting participants, especially when dealing with forced migrant populations that are not documented and are not already described in accessible datasets.
Participatory Action Research
Localization overlaps with participatory action research, whereby research is conducted “with those people whose life-world and meaningful actions are under study.” This approach comes with several advantages. For one, participatory action research helps to bridge power imbalances between those conducting research and the populations under study. Second, the approach links local, grassroots action with national or international government, humanitarian, and development actors. Third, participatory action research helps to identify the research priorities of the population of interest, rather than imposing priorities based on external assumptions. Despite these advantages, this approach also comes with limitations and challenges, such as the question of who is included in participation, and the need to have a safe space for groups to communicate openly in order for participatory action to be conducted effectively and ethically.
RIT uses participatory action to encourage findings that are relevant to the populations being studied, and to develop findings that might serve as the foundation for programming that will be facilitative of integration.
Reflexivity & Positionality
Positionality is “the notion that personal values, views, and location in time and space influence how one understands the world.” In social science research—especially when dealing with highly subjective topics like migrant integration—it is helpful for researchers to work to determine their own position in relation to their research questions, their methods, and the populations being studied. This work is achieved the researcher’s “ongoing critique and critical reflection of his or her own biases,” referred to as reflexivity. This may be achieved through group discussion, keeping a reflexive journal, or reporting on positionality when presenting research findings. Reflexive discussions may take place through formal analytical exercise, or informally through “kitchen table” chats about researchers’ positionalities before, during, and after data collection. Ultimately, the goal of reflexivity is not to eliminate bias, but for the researcher to become aware of his or her biases and incorporate this awareness into the research process.
Reflexivity has value for all social science research—quantitative or qualitative—particularly during the analysis phase when researchers’ biases may easily be introduced into the conclusions and the outputs produced from the data. However, reflexivity is also time consuming, and may not be of interest to some consumers of research findings.
“Critical urban theorists” and forced migration scholars—particularly anthropologists—employ the reflexive approach during research. However, the methods sections in forced migration research do not consistently discuss positionality and rarely discuss researchers’ positions relative to their process or findings. RIT encourages our case report authors to use the guided, everyday talk of "Kitchen Table Reflexivity," to help think about their relationships to neighborhoods, other migrants, and host populations. This everyday talk may take place with people they are close with, intellectual mentors, or virtually with other RIT report authors.
All of the above types of research may use a variety of different methods for collecting both qualitative and quantitative data. Here, we outline a few of these methods.
There are many different types of interviews which fall on a spectrum from structured—with a pre-defined list of questions—to unstructured. Each end of the spectrum brings advantages and disadvantages. All interviews are founded on trust created between the interviewer and the respondent. Two primary types of interviews are Key Informant (KI) and in-depth interviews. In a KI, the interviewee does not speak about him/herself, but rather about a community, institution, or phenomenon as an expert. In an in-depth interview, the interviewee speaks specifically about his or her personal experiences. As an example, in the RIT project KI interviews include resettlement agency representatives, while in-depth interviews include long-time residents of neighborhoods where refugees have settled. These interviewees also might provide referrals to others who would be willing to speak with the researcher.
The RIT project, and other research projects that involve interviewing refugees and other potentially vulnerable populations, faces a unique set of methodological challenges as well as ethical and legal considerations related to protecting sources and personal information that must be addressed before beginning interviews.
Depending on a number of factors in the research process—including number of enumerators, time, accessibility to populations, and financial constraints—interview samples may be statistically representative or non-representative. Quantitative studies will require statistical techniques to develop a representative and statistically powerful sample with high confidence levels (see quantitative data section below). However, qualitative research can still achieve strong validity without statistical significance in part by using a well-designed sampling strategy.
A sampling strategy is used before beginning interviews to determine which demographics and relevant groups may need to be interviewed so as to have representation in the sample. For example, when selecting which refugees to speak with in a RIT case, we often “purposively” sample refugees from a range of ages, genders, economic and education levels, or lengths of time spent in the urban area.
In the RIT project’s case reports, interviews typically include a wide range of people to accurately represent the diversity of interests and groups in the town or neighborhood the case is exploring. We might also target specific populations for interviewing based on unique characteristics that could be affecting their integration, such as refugees belonging to the LGBTQI+ population.
Ethnography is a process of observing and recording social interactions, rules, and behaviors. This may be formal and structured and follow specific techniques, or may be less formal and unstructured, more akin to "hanging out with forced migrants" while taking notes on the interactions between people and the setting they are in.
Because RIT’s researchers are already living in the neighborhoods that their case reports are studying and are often forced migrants themselves, they typically use informal, unstructured ethnographic observations from their daily lives to inform their reports.
Participant Observation is a method of data collection in which the researchers establish rapport with the population being studied. In RIT’s case, this involves building connections with both refugee and host populations so that the researcher can move freely within and between them. In participant observation, the researcher gathers data through unobtrusive methods including observation, natural conversations, and interviews. Researchers using participant observation should be open and nonjudgmental, and be interested in learning about the different experiences and attitudes of a wide range of people. They must also be careful observers and good listeners. In RIT’s cases, the researchers can be members of either the refugee or host population but seek to become familiar with both groups. Typically, by the time RIT report authors join the project, they already have established a broad and deep rapport in their town from years of living amongst refugee and host populations.
Geographic Information Systems (GIS)
GIS is the “science of where.” GIS data boils down to a list of information coded with geolocations. These data are developed into maps that help tell a story about the influence of location on a phenomenon.
For the RIT project, GIS helps tell a story about how place, “place-making,” and the differing uses of places by migrant and host populations influence integration. Specifically, RIT’s mapping helps answer two questions:
1) Are refugees/migrants clustering in certain areas, and why?
2) How do the locations of private and public services relative to refugee/migrant residential locations influence access to these services?
Large amounts of geodata are already freely available to the public through resources like Google Maps and Humanitarian Data Exchange. However, in data-poor environments—like many developing cities where the majority of refugees reside—researchers may need to collect original geolocation data on their own. Researchers can easily do this with a range of methods, such asking residents to identify places of interest on a digital map, or taking a walk through the neighborhood with a smartphone loaded with a map application like Google Maps and dropping geolocated “pins” at key places. Researchers can then take these data and build maps with free, open-source software such as Google Maps, Google Earth, OpenStreetMap (OSM), or QGIS, each of which has numerous how-to videos that are widely available online.
Researchers might also make subjective maps to explore how refugee and host populations perceive certain areas differently in terms of the accessibility or purpose of the space.
Map making in areas with forcibly displaced persons presents some unique ethical challenges. It is critical to first consider the safety and protection of personally identifiable information before beginning mapping. Researchers might also consider the power dynamics of maps: in refugee research, formal institutions of host populations tend to appear on maps like Google while informal businesses, community centers, or other migrants institutions are not as frequently mapped (often deliberately to protect populations from the host population, other migrants, or the state), meaning maps might perpetuate or disrupt these imbalances.
As a rule of thumb, information that cannot be easily obtained publicly (e.g. by walking down the street or using an internet search) or that puts anyone’s individual location (e.g. refugee residential or workplace addresses) should not be mapped publicly. Individual residential location data might be generalized at the neighborhood or census tract level to avoid introducing risk to refugees. In some countries, any mapping at a more granular level than the city—such as the neighborhood or street level—introduces unwarranted risk for forced migrants who rely on obscurity for protection.
In forced migration research, quantitative datasets typically are derived from randomized or representative large-n statistically powerful surveys. Survey data for many countries are already publicly available online in the form of census records or humanitarian databases and simply need to be collected, organized, and analyzed by researchers. However, surveys conducted by an independent researcher or that contain sensitive information—such as UNHCR registration data—can be more difficult to access. Often researchers must first establish trust with the holders of such data and request permission access these data. Building such relationships between data holders and researchers can take time, but these collaborations are usually worth the effort for all involved as researchers can bring new expertise to analyzing datasets that an institution may not already have in house.
Quantitative data and qualitative data ideally should be used together through triangulation. Triangulation involves designing a study’s instruments, methods, measures, and indicators so that findings from one method can be tested against findings from another, thereby increasing the study’s overall validity strength. In addition, triangulation can be designed into the composition of a research team through reflexivity.
Triangulation is especially critical when data are scarce—as is the case in many developing cities—or when studying hidden or difficult to observe populations—like unregistered forcibly displaced persons in cities.
One example of triangulation in forced migration research is the ethnosurvey which triangulates quantitative survey data with qualitative ethnography to reach more valid population estimates than would be possible with a single measure or instrument. RIT aims to triangulate findings between GIS maps, original qualitative data, quantitative data from aid agency and census surveys, and archival information from municipal sources.
After collecting data, researchers need to make sense of these data through analysis. Organizing data and analyzing it can be just as difficult and time consuming as collecting it. There are a wide range of analytical techniques for qualitative research. At the RIT project, we predominantly apply grounded theory. Regardless of which technique is used for analysis, reflexivity remains key, especially when dealing with forced migrant populations and subjective measures and indicators.
You may also be interested in our lists of key academic articles, practitioner reports, policy documents, books, databases, and journals related to refugee urban integration. These sources are also summarized in our Literature Review.