https://orcid.org/0000-0003-4218-412X
Abstract
This paper assesses the challenges and the benefits of 2 methodological approaches for improving the study of transnational mobilities of older adults: mixed methods and inclusiveness. The first approach refers to a mixed-methods research design based on surveys and qualitative interviews. We share our experience of conducting a “fully mixed-concurrent-equal-status-design” research project, for which we collected data through quantitative surveys on transnational practices of people aged 55+. Furthermore, we conducted semistructured interviews with adults aged 64+ living or who have lived in Switzerland and spend part of the year in Spain. The second approach relates to the use of an inclusive sample diversified in terms of nationality and migration backgrounds. The inclusive design affects the formulation of questions asked to all participants (and the response options provided in the survey), as well as decisions related to language choices and translations. These elements have to be considered to ensure that older adults from different backgrounds feel included. In a research project on transnational mobility of older adults, conducting an inclusive plus mixed-methods research project pushes researchers to find strategies to balance research objectives with available resources. Maximizing the research team’s methodological background appears to be a suitable approach to address different population groups while working within a budget.
Research projects that aim to study the transnational mobility of older adults are confronted with the difficulty of reaching highly mobile individuals. These individuals might decide to move to another country or to live concurrently in two or more countries for different reasons (e.g., climate, health, family, culture). How adults organize and spend their years after retirement is thus not necessarily restricted to a single nation-state. This is supported by the literature on older adults’ transnational practices that has considerably expanded in recent years (Baykara-Krumme, 2013; Bender et al., 2018; Bocker & Gehring, 2015; Ciobanu et al., 2017). This research indicates that mobility increased in the last decades among older adults with and without a migration background due to globalization and affordable means of transport. Inclusive mixed-methods research designs are interesting in this regard, as they offer the possibility of maximizing sampling strategies and targeting different groups in a single project.
This paper contributes to methodological literature in aging research by presenting an inclusive, mixed-methods research design that investigates transnational practices and mobilities of older adults who have a close link to Switzerland.1 Switzerland can be regarded as an interesting case study for an inclusive research approach because of its global interconnectedness due to past and present immigration and emigration flows. It is commonly referred to as a land of immigration with one quarter of residents of foreign nationality and about 40% with a migration background (Swiss Federal Statistical Office [SFSO], 2020a). They generally came in the twentieth century mainly as young labor migrants in search of better working conditions or to escape a dictatorial regime. Foreign nationals now aged 55+ are mostly Italian (27.4%), German (14.6%), Portuguese (8.2%), French (5.7%), and Spanish (4.9%) (own computations with the Swiss census). However, Switzerland is also a land of emigration, with 11% of Swiss citizens living in another country. The main European destinations are France, Germany, Italy, the United Kingdom, and Spain (SFSO, 2020b). About 20% of Swiss citizens living abroad are aged 65 or older, indicating that transnational mobility is an important phenomenon within the older population of this country (Federal Department of Foreign Affairs [FDFA], 2020). An inclusive research design that analyzes the mobility and transnational practices of older adults with and without a former migration experience is therefore much needed for Switzerland and also for other northern countries (e.g., the United Kingdom, the Netherlands, Germany, the United States, and Canada). Research in these countries has, however, analyzed transnational mobility of older adults with and without a former migration experience separately. Research on older migrants concentrates on former labor migrants who either age in place (Buffel, 2017; Ciobanu et al., 2017; Ciobanu & Ludwig-Dehm, 2020) or return to their country of origin (Aguila & Vega, 2017; Baykara-Krumme, 2013; Bocker & Gehring, 2015). In contrast, research on older adults without a former migration experience focuses on adults from northern countries who live (for part of the year) in a southern country (Bender et al., 2018; Huber & O’Reilly, 2004; Rojas et al., 2014).
Based on our experience of using an inclusive, mixed-methods research design, this paper demonstrates how such a design can improve the study of older adults, and which challenges can arise in this process. In doing so, we will illustrate the construction and realization of a mixed-methods research design and outline some methodological approaches to construct inclusive samples, that is, samples that go beyond a specific nationality or migration background.
Inclusiveness in Research Designs
A recurrent problem in research studying transnational practices is that the dominant category is “ethnicity.” According to Amelina and Faist (2012, p. 1710), this implies that “the starting point of this type of research is not geographical mobility, social networks or decision-making as such, but the particular ‘ethnic’ or ‘national’ group.” This problem is exacerbated by paying particular attention to migrant populations. This focus is supported by large quantitative surveys on transnationalism that target older migrants (see, e.g., the Transnational Module of the European Labour Force Survey). However, such research “runs the risk of supporting the view that migration-related difference is naturally given” (Dahinden, 2016, p. 2208). Two recent studies, Kahveci and colleagues (2020) in Turkey and research with data from the survey Vivre–Leben–Vivere (VLV) conducted in Switzerland (Ludwig et al., 2014), point to some similarities among people with and without a migration background. Kahveci and colleagues (2020) show that German and German-Turkish older adults develop comparable mobility patterns between the two countries. Furthermore, both groups are confronted with similar structural conditions such as low incomes and changing legal frameworks related to health care and insurance, and residence abroad. In addition, Bolzman and Vagni (2017) using the VLV illustrate that differences in the use of institutional care services between older migrants and older Swiss are mediated by health and marital status. In other words, poor health and a divorce or the death of a partner render the use of institutional care services more likely, irrespective of nationality or migration background. These studies demonstrate that an inclusive approach has the potential to reveal differences, similarities, and inequalities not just in terms of nationality or migration backgrounds, but also in terms of other factors such as social networks and health status.
The Mixed-Methods Research Design
Mixed method is regarded as the third methodological paradigm of social science, alongside quantitative and qualitative paradigms. It can be broadly defined as the use of quantitative and qualitative research designs in one study (Caruth, 2013; Happ, 2009). In doing so, quantitative research benefits from qualitative findings by adding narratives and meaning to numbers, while quantitative findings add precision to words and narratives (Caruth, 2013).
Leech and Onwuegbuzie (2009) portray a mixed-methods research design as a function of three dimensions: the level of mixing, time orientation, and the emphasis of quantitative and qualitative approaches in the project. The first dimension specifies where a project is located on the continuum from monomethod designs (e.g., quantitative or qualitative designs) to fully mixed-methods designs. While fully mixed-methods designs use quantitative and qualitative designs within several or all stages of the research, partially mixed-methods designs combine the research approaches only at one stage of the research (Schoonenboom & Johnson, 2017). The second dimension defines if the quantitative and qualitative parts of the project are conducted concurrently or sequentially. The major difference between these two time orientations is that the sequential use of quantitative and qualitative methods needs the results from the first part of the project to develop the second part. This is not necessary for the simultaneous time orientation (Johnson et al., 2007). The final dimension specifies if the project puts emphasis on quantitative or qualitative methods, and their respective epistemological considerations. The emphasis is, thus, crucial to the entire project as it defines the objectives, the design, and the kind of results produced by the research.
In general, there is a lack of use of mixed methods in gerontological research (Happ, 2009; Weil, 2017). Mixed-methods studies are, however, particularly useful for research with older adults because “mixed-methods research designs place high value on the stories behind the numbers—both in exploratory designs where the experiences and insights of the community under study inform the quantitative investigation, and in explanatory designs where they illuminate the quantitative data” (Gallo & Joo, 2018, p. 129).
The Transnational Aging Project
This paper is based on the project “Transnational Ageing: Post-Retirement Mobilities, Transnational Lifestyles and Care Configuration” that aims to investigate the life before retirement; the reasons for migration and/or im/mobility; transnational practices; social networks; and care needs and obligations of older adults who have a close link to Switzerland.2 The project is composed of two transnational aging surveys and a qualitative fieldwork focusing on the transnational mobility between Switzerland and Spain.
The Mixed-Methods Research Design
In our project, we combined quantitative and qualitative designs across several stages of the project and assigned our research project to the fully mixed-methods design. For example, we collaborated closely when defining the goals and the research questions of the overarching research project. We organized team meetings on a regular basis to discuss the main research axes of the mixed-methods research project and to receive feedback on the research instruments (e.g., quantitative surveys and the interview guide). Furthermore, we conducted the surveys and the qualitative fieldwork with the smallest possible time difference because the preestablished research funding time frame made simultaneous data collection and analysis unavoidable. The first quantitative survey took place in the first half of 2020 and the second survey was conducted in the first half of 2021. Qualitative data collection began in June 2020 and took about 1 year. The qualitative fieldwork has no longitudinal alignment. The concurrent phase of data collection had the advantage that we had to work together as a team from the beginning in order to present a coherent project (Figure 1).
Timeline for quantitative and qualitative data collection.
Finally, the emphasis of our research design depends equally on the quantitative and qualitative parts of the project. On the one hand, we wanted to generate generalizable results on a wide variety of transnational practices (e.g., social networks, mobility, political activism/participation, feelings of belonging) of Swiss residents and Swiss citizens living abroad, both aged 55+. To achieve this, we conducted quantitative surveys with large and randomly selected samples. On the other hand, we wanted to make analytic generalizations to nourish our newly developed theoretical framework retirement mobilities (Onwuegbuzie & Corrigan, 2014; Tomás et al., 2020). For this, the qualitative case study of Spain was essential. It was therefore crucial to regard both parts with the same importance. The equal dependence of the mixed-methods project on quantitative and qualitative data becomes also evident in the composition of the team, which is made up equally of quantitative and qualitative scholars. However, and here lies the difficulty of such a design, researchers with different epistemological views had to work closely together. While the quantitative researchers held a positivist worldview, the qualitative researchers adopted a constructivist philosophy to their research (Weil, 2017). The willingness to learn from each other and to engage with other (sometimes contrary) worldviews is crucial for the success of a mixed-methods research project.
In sum, we attribute the research design of our project to the “fully mixed-concurrent-equal-status-design” (Leech & Onwuegbuzie, 2009, pp. 270–271). Although the project was conceived as a mixed-methods research, the choice of the specific design was not predetermined. Rather, it evolved in its empirical implementation phase and was influenced by the realities of fieldwork and the members of the research team. However, the choice of the project’s mixed-methods design was crucial as it enabled us to maximize the target groups, including migrants and nonmigrants, and mobile and immobile older adults.
The Transnational Aging Surveys
The sample of the first survey (“Survey on the life and travel experiences of people aged 55+ in Switzerland”) was constructed in collaboration with the SFSO. It contained a total of 14,860 addresses encapsulating the three main linguistic regions of Switzerland and the main nationalities among the population 55+, namely Italian, German, Portuguese, French, and Spanish. Furthermore, we included three heterogeneous groups for nationals from the Balkans (accounting for 16% of older migrants according to the Swiss census including Albania, Kosovo, Bosnia-Herzegovina, Macedonia, Montenegro, Serbia, and Turkey), remaining Europeans and non-Europeans. Foreign nationals were oversampled in anticipation of lower response rates than from Swiss respondents due to less stable addresses, longer absences, and different cultural attitudes toward population surveys (Steiner & Landös, 2019; see Table A1> and A2 of the Supplementary Material for details about response rates). The sample was further stratified by gender for all and by retirement age for Swiss nationals. The second survey (“Survey of Swiss men and women aged 55 and over living outside Switzerland”) included 10,000 addresses of people living in another country than Switzerland (mainly, but not only, France, Spain, and Germany, in Europe more in general, and in the United States). These addresses were provided by the FDFA.
Swiss surveys are usually translated in three official languages (German, French, and Italian). To make the survey as inclusive as possible for the large proportion of migrants aged 55+, our team translated the survey also into Spanish, English, and Portuguese using reliable validated scales following examples of Survey of Health, Ageing and Retirement in Europe (SHARE) and the European Social Survey, which are already translated in several languages. For the first survey, the language for correspondence was chosen according to the language spoken in the municipality of residence (either German, French, or Italian). In addition, Italians, Germans, Portuguese, French, and Spaniards received letters in their national language, if this was different from that of the municipality. Furthermore, an English version could be chosen online or requested on demand by participants. We assumed that participants understood at least a part of the information given in the invitation letter. This assumption was made because we did not target recently arrived migrants, but rather nationals and older migrants who had lived in Switzerland for several years. These language choices meant that we were unfortunately not able to solicit older migrants from the Balkan region or from other countries in their national language. Even though our project was committed to inclusivity, we had to take some restrictive decisions due to budget and time constraints. Because respondents from the Balkan region are particularly representative of more recent migratory movements to Switzerland, we decided to focus on the six languages spoken by the majority of our participants. For the second survey, we used the registered language of correspondence of the FDFA. Most respondents abroad used more than one language, and, in some cases, they were no longer fluent in the language registered by the FDFA. In those cases, we provided a translated version in Spanish, Portuguese, and English. Important language issues emerged mainly among respondents living in China.
Within the chosen languages, we adapted the semantics of the questions to widen inclusiveness for both Swiss, binational, and non-Swiss nationals. As a result of feedback received during 10 pretests, we avoided the words “foreign,” “country of origin,” and “abroad,” substituting them with “in another country/a different country than Switzerland.” We could reformulate almost all questions, but we had to keep “country of origin” in some response options due to the lack of alternatives as “country of birth” or “country where people spent most of their life” were not judged appropriate by other migration colleagues. In addition, we paired questions about voting behavior and political activism in another country, which concerned mostly non-Swiss respondents, with questions about activism in global causes, which could potentially concern all respondents. With this strategy, we obtained low item nonresponses for attitudinal questions (between 0.8% and 4.2%). The disadvantage was that questions multiplied, increasing the burden on respondents (in the first survey, we had a total of 110 questions, of which 47 had multiple response options; seven Likert scales for satisfaction and health questions and 39 questions that could be skipped depending on the filters; we reduced the number of questions to 86 for the second survey). The topic of our questionnaires were: (a) past and current international mobilities; (b) transnational retirement (plans); (c) other links with Switzerland and with other countries (e.g., nationalities, real estate, voting behaviors, feelings of belonging); (d) family in Switzerland and in other countries; and (e) sociodemographics.
In our questionnaires, we formulated mainly multiple-choice questions with an open category for 37 questions. This option was particularly well suited to older respondents as it was adapted to their multiple life-course patterns and mobility trajectories. Moreover, some old-old respondents answered the questions as if the past reflected their current situation (e.g., they answered they took care of grandchildren now even though they did so only in the past). This “other” option enabled them to clarify their choices and to indicate whether they had “updated” their responses. We could also recode the answers given in the “other” category including additional positive or negative connotations linked to the classification (e.g., a sick or a healthy person, an in/dependent person, a selfish/caring person).
The Qualitative Fieldwork
The qualitative fieldwork was conducted in Switzerland and Spain with two target groups: first, with older adults residing in Switzerland who spend at least 3 months per year in Spain, and second, with older adults who spent part of their working life in Switzerland and moved to Spain in retirement age.
In view of constructing an inclusive sample for the qualitative project, we advertised our project through various channels. We contacted organizations for older adults, and got in touch with religious or national associations and asked them to distribute our flyers (available in German, French, Italian, Spanish, and English) during events, to invite us to present the project at formal or informal gatherings and to publish information about it in their newsletters. We also searched for participants in our personal and professional networks and included a paragraph on our qualitative study in the invitation letter for the survey. Besides, we wrote an article on our research project for the Swiss Review, a magazine for Swiss abroad published six times a year. All people registered with a Swiss representation are entitled to receive this magazine. With the help of this magazine, we were able to contact a large number of older adults; however, they were all of Swiss nationality. In order to reach a wider variety of older adults, we collaborated in spring 2021 with the Central Compensation Office (CCO) of Switzerland. This institution is responsible for paying the old-age and survivors’ insurance to any person who has worked in Switzerland regardless of nationality. These benefits are not just paid to older adults living in Switzerland, but also to older adults who left Switzerland to live in another country, yet who have once worked in Switzerland. Between 2001 and 2019, for instance, the number of pensions paid to older adults living abroad has doubled (Office fédéral des assurances sociales, 2020). Spain on its own accounts for approximately 145,000 pensions paid by the CCO. In total, older adults of 63 different nationalities were entitled to receive pensions in Spain in December 2018.
Although we aimed to include for consistency and comparison purposes the same age group (55+) in the quantitative and qualitative research, we decided to interview adults who have reached retirement age. We perceive this adjustment as acceptable because of the opportunities that the collaboration with the CCO offered. Furthermore, the sampling strategies chosen for the two quantitative surveys and the qualitative fieldwork overlap in two respects. While the first target group of the qualitative fieldwork enables us to establish a connection to the mobile population residing in Switzerland questioned in the Transnational Ageing Survey I (TAS I), the second target group joins with the Swiss living abroad interrogated in TAS II. In sum, we collect quantitative and qualitative data for two specific subgroups: mobile older adults residing in Switzerland and Swiss citizens living in Spain. We can, therefore, conclude that our mixed-methods research project follows a multilevel sampling design (Onwuegbuzie & Corrigan, 2014; Weil, 2017). While there are overlaps in the quantitative and qualitative samples (particularly between TAS II and the qualitative fieldwork), we can also find differences in terms of age and nationality.
The qualitative project consists of semistructured interviews conducted in German, French, and Spanish. The questions were formulated in an open and narrative-generating manner, and the interview guide offered enough room to respond to each interviewee’s narrative. The interview guide was organized according to the following subjects: (a) introductory questions (sociodemographic questions, life before retirement); (b) life after retirement and transnational mobility during retirement (reasons/motivations, mobility patterns); (c) other transnational practices (political participation/activism, feelings of belonging); (d) health, health system, and satisfaction with it; (e) social networks; and (f) future mobility plans. Due to the narrative-generating formulation of the questions, not all questions were asked systematically and in the same order. Depending on how the interview evolved, it was possible to speak about the health system or the social network earlier during the interview than planned in the interview guide.
In total, we conducted 35 interviews in Switzerland and Spain between June 2020 and May 2021 with 43 older adults. Our interviewees were between 67 and 85 years old and the majority resides today in Spain. Most of our interviewees are Swiss citizens (34), of which six have a migration background. Two have Spanish nationality and the remaining seven interviewees have neither Swiss nor Spanish nationality (Argentinian, German, French [two], British, Italian-Argentinian). Thus, the cooperation with the CCO has supported us in diversifying our sample, which becomes particularly apparent in the seven interviewees without a Swiss or a Spanish nationality.
Discussion and Implications
In this paper, we presented the benefits and challenges of two methodological approaches for improving the study of transnational mobilities and practices of older adults. First, an inclusive sampling approach and second a mixed-methods research design. We argue that an inclusive, mixed-methods approach is particularly valuable for northern countries, and for Switzerland, where academic work on migration and transnational mobility in old age focuses either on former labor migrants or on international retirement migrants.
Such an inclusive approach also has its difficulties. First, it is not an easy task to find a sampling strategy that goes beyond a specific nationality or a migration background. While we found a straightforward solution for implementing the quantitative surveys thanks to the SFSO, a more diversified sample was harder to create for the qualitative fieldwork. As a diversification strategy, we collaborated with the CCO of Switzerland and the FDFA. The support of national institutions was therefore crucial to our project. Second, inclusiveness creates the need to adapt language, the formulation of questions and, in the case of the quantitative questionnaire, response options. To make sure that all older adults felt addressed by the questionnaire and the interviews, we avoided the words “foreign,” “country of origin,” and “abroad,” substituting them with “in another/a different country than Switzerland.” For the same reason, we included additional questions on voting behaviors and political activism by not only linking them to one or several countries but also to global issues. These adjustments can be time-consuming for both the research team and the participants, as they increase the number of questions. Finally, the decision to only translate the questionnaire into German, French, Italian, Spanish, English, and Portuguese had clear consequences on the participants from the Balkan region: this heterogeneous group had the lowest response rate (7%). As this last group is very diversified in terms of languages, it is necessary to forecast a high translation budget if the aim is to adequately include them into the research project.
The mixed-methods design of the project increased the possibilities of reaching immobile and highly mobile older adults. The latter are particularly difficult to contact, because of their back-and-forth mobility between two countries. Maximizing the variety of research tools and instruments seemed to us a suitable strategy to take into account the different profiles of respondents. Due to the project’s limited time span of 4 years, we decided to conduct the data collection of the quantitative and qualitative parts simultaneously. Because there is only limited space for interaction between the two parts during the simultaneous phase of data collection, we perceived it as important to define the goals, questions, hypotheses, and research instruments collaboratively to gain the greatest possible benefit from the mixed-methods design. Our team consisted of quantitative and qualitative researchers who had in-depth skills in their respective fields and were willing to learn from each other. Open-mindedness and methodological flexibility were crucial to common discussion on how to develop our overall research design and the research instruments. Thus, all researchers involved in a mixed-method project must be prepared and open for negotiation, discussion, and compromise (Happ, 2009). On the one hand, these negotiations and compromises are useful tools to prepare the fieldwork and to maximize the benefits of a mixed-methods research design. On the other hand, mixed-methods research designs are more time-consuming than purely quantitative or qualitative research designs.
The choice of the most appropriate sampling strategies and research methods is dependent on the project’s budgetary and time constraints. Although there might be adjustments to be made depending on the institutional context in which research is conducted, our experience provides useful hints about the dos and don’ts for future research on the transnational mobility of older adults.
Funding
This work was supported by the Swiss National Science Foundation (51NF40-182897) through the nccr—on the move (IP33 “Transnational Ageing”).
Conflict of Interest
None declared.
Footnotes
Here, we refer to older adults who (a) have the Swiss nationality but do not necessarily live in Switzerland, and (b) live or once lived in Switzerland but do not necessarily have the Swiss nationality.
For more information, visit the following website: https://nccr-onthemove.ch/projects/transnational-ageing-post-retirement-mobilities-transnational-lifestyles-and-care-configurations/.
Acknowledgments
We would like to thank the reviewers and the journal editor for their very helpful comments on earlier paper drafts. We would also like to express our gratitude to Dr. Başak Bilecen and Dr. Tineke Fokkema for their support, guidance, and useful critique in the elaboration of this paper.
