THE CHALLENGE OF LOGISTICS

The unit of analysis of my original research question was the union, and I planned to explore the question of differing responses by labor unions to immigration through data collection across three countries (Denmark, Germany, and the United Kingdom (UK)) and two sectors (elder care homes in the public sector and food manufacturing firms in the private sector). I conceptualized unions as having two categories of responses, internal and external. Internal responses included organizational discourse within unions referring to immigrants and representation of immigrants as a fraction of the membership and elected positions. External responses covered lobbying by the union in society on behalf of immigrant concerns. For my initial data collection plans, I completed a pilot research trip in 2009 and planned to spend 13 additional months abroad from 2010 to 2011 conducting in-depth interviews with trade unionists, activists, employers, politicians, and journalists across these three country cases and two sectors. I also planned to do surveys of elected union representatives in each country, compile observational data by visiting six workplaces, one in each sector per country, as well as visit archives.

Logistical challenges presented themselves with an early set of interviews in November 2010. I had received introductions to five officials responsible for immigration issues, each affiliated with a different sectoral union in Germany. My informants’ locations ranged from two- to five-hour train rides from my research base in Berlin, Germany. I faced the logistic challenge of scheduling all the interviews within a two-week period, physically getting to the places and staying long enough to follow up with any referrals to additional informants or worksites. After scheduling these interviews, I realized that it had taken me two weeks with those five informants to coordinate and cluster the interview appointments, while the interviews would take place over the following three-week period. Scheduling a handful of interviews within one of my country cases took more than twice as long as I expected, therefore I discovered it would probably be quite challenging to maintain the scope of the project across multiple countries, sectors, and types of data.

To move forward, I had to decide whether to scale back my ambitions or to press ahead. Because I thought that it was too early to change course and drop a country case, sector, or type of data, and I was unsure if a windfall of data was just around the corner, I settled on a strategy involving regular evaluations of the data-gathering process. I committed myself to sticking with my research project as long as I met reasonable substantive and temporal thresholds. For example, the substantive goal I set for those five interviews with informants was to secure leads to worksites where I could collect observational data and gain permission for and aid with the distribution of a survey of union representatives. For temporal parameters, I allowed myself a month after each interview to set up worksite visits and the surveys, assuming my informants were willing and able to help me. I also shared these thresholds with local academics and members of my dissertation committee to get advice about the feasibility of my expectations and to create accountability.

RECOGNIZING THE IMPRACTICABILITY OF THE SURVEY

During early trips to field sites in the UK and Denmark in 2010 and 2011, I talked to informants and local academics about possible worksite locations where I could collect survey data. Some informants quickly expressed their inability to help me field surveys in a straightforward fashion. One British informant at the executive level of an industrial union said that the union could not afford to let me loose to do my own survey because of what I might find. Union officials would only allow outside researchers to add questions if the union had already committed to conducting their own survey. Another British informant at a public sector union told me she did not have the authority to give me permission to do the survey and that she could not direct me to anyone with that authority. I faced similar obstacles in Denmark, however, union officials stated that I would also need the permission of employers to field a survey, and their permission would be difficult to obtain without an affiliation with a Danish university and Danish academic partners.

A more difficult situation arose when informants conveyed enthusiasm about aiding me with my research, but weeks and months later did not respond to e-mails or phone calls about the specifics. For example, in March 2011, I had a particularly informative and congenial interview with a midlevel union official in a public sector union in Denmark. At the end of the interview, I broached the topic of locating a worksite for observation and fielding a survey. He reeled off several potential places that he thought could be feasible based on his professional connections to union representatives and managers and told me to follow up with him by e-mail and phone so that we could arrange my visit during the next two months. Although I sent several e-mails and left a number of voicemail messages, I was never able to contact him.

Setting substantive and temporal thresholds proved useful in my decision to abandon the survey. As an initial visit to the UK in November 2010 involved contact with local academics only, I allowed myself a month to follow up with referrals directly connected to potential worksites as union representatives or employers and an additional month to set up interviews and worksite visits. When none of the British informants I met in January 2011 could commit to helping me field a survey, I decided to abandon the survey in the UK. By January 2011 I was sure the worksite surveys would no longer be part of my dissertation, but decided to try in Denmark. If I succeeded in fielding a survey in even a single country case, I could still use the data for an article or book project later. However, as described earlier, although some Danish informants expressed interest in helping me in March 2011, they were nonresponsive during April and May, which led me to abandon the survey portion of my data collection by the end of May.

CHANGING SECTORS TO ACCESS COMPARABLE DATA

In addition to managing geographic distances when accessing data across country cases and deciding to drop one type of data—surveys— questions of scope can also crop up for scholars when attempting to acquire the remaining data types in equivalence across cases. Setting the parameters of this project to include multiple cases, sectors, and types of data seemed feasible owing largely to the success of a pilot foray into the field. I spent two months in the summer of 2009 locating and contacting 40 potential interviewees in Denmark and was able to interview nearly 30 of them during that short visit. Yet, extrapolating from the relative ease with which I located contacts in my 2009 pilot trip proved to be a mistake. In the process of getting feedback from local academics on substantive and temporal thresholds for setting up interviews, I realized that I had to change my sectors of focus a few months after arriving in the field in 2010.

In Denmark, I found contacts from 2009 willing to facilitate workplace visits to a food manufacturing plant and elder care home. Yet, before I had organized parallel workplace visits in Germany and the UK, local academics urged me to change my sectoral focus because of the difficulty I would have obtaining comparable data. Although these sectors had union presences in Germany and the UK, unions had a more visible presence in major manufacturing firms and public services such as metal manufacturing/autos and public hospitals, respectively. Most important, immigrants in Germany and the UK were more likely to come in contact with unions in these sectors. The chance that immigrants might have little contact with unions in my two original sectors made them potentially poor sites to gather evidence on my outcome of differences in union responses to immigration across country cases.

TURNING TO A NEW OUTCOME

Researchers may at times exit from fieldwork abroad with a new dependent variable. Factors such as difficulties obtaining comparable data across cases, locating informants, or the discovery of local knowledge contradicting the existing literature (LaPorte, this symposium), can force scholars to change their outcome. At the end of several months of fieldwork, rather than seek to explain differences in union responses to immigration, my project shifted to instead explain when native and immigrant union members were able to cooperate to take part in workplace protests such as strikes.^{Footnote 1} My arrival at a new outcome unfolded as I conducted fieldwork for my original question and highlights the iterative and often unpredictable nature of research in comparative politics. Just as scholars cannot extrapolate from past experiences in the field and need to prepare for developments that may be hard to foresee, scholars can also stumble on important caches of data or “opportunities for observation” and insights into innovative analysis (Bissaillon and Rankin Reference Bisaillon and Rankin2013).

The first inkling of this impending change came after a conversation in July 2010 with a prominent German scholar of immigration with whom I had coffee at the beginning of fieldwork. I told him about my interest in understanding why trade unions had such different responses to immigration, and he said that one reason might be because German unions, more than their British and Danish counterparts, have a greater reliance on the strike capacity of members.

After this scholar’s suggestion, including questions about strikes or any other workplace protests became part of my interview scripts. Yet, until I changed my sectors of focus from food manufacturing and elder care to metal manufacturing/autos and public hospitals by the end of 2010, the strike/protest questions did not result in particularly useful responses from informants. At my first worksite visit to a public hospital in March 2011, I was able to interview several union members, some of them union reps and works councilors, many of them with immigrant backgrounds. One of the things I asked them was how they decided to join the union, and whether they thought the union was doing anything for immigrant integration. I received a variety of responses on the integration question, however some of the platitudes about how great the hospital and union were at integration contrasted with other complaints my interviewees had about obstacles to promotion they faced at work.^{Footnote 2} One interviewee got involved in the union because he wanted higher wages and thought that approaching the union as an individual would be the best way to do it. Yet what he discovered when talking to the union rep was that it was not something he could do alone. The condition of union help was getting his colleagues on board with the demand for higher wages. Certainly, in those interviews, I realized that union representatives encouraged collective action as a problem solving strategy, yet I still did not think about collective action as an outcome for my research. Not until the very last interview I conducted a few weeks before returning to the United States, for a different hospital work site, did I suddenly grasp the potential of protests as a new outcome.

Talking to bystanders and participants in May Day protests in Berlin, I learned about a hospital strike nearby. Halfway through an interview with one of the strike organizers, he shared a surprising piece of information. Work in the hospital was bifurcated not only by skill but also, largely by ethnicity. Most immigrants worked in low paid, insecure jobs for a private contractor in cleaning and catering while Germans held highly skilled, secure public sector jobs in medical care at the hospital. German trade unionists directly employed by the public hospital decided to combine their protest for higher wages, with that of the immigrants in the hospital working for the private contractor. What surprised me was that German trade unionists would not gain materially from this alliance, yet they had gone out of their way to set it up. I asked my informant why they were doing this. He gave me several reasons: solidarity, ideological opposition to privatization, poor work conditions, and a sense that immigrants as a group lacked a lobby. I wondered if their immigrant partners would give the same reasons for cooperation. When I asked my informant if the alliance had succeeded, he said that it had been very difficult to maintain. When the German skilled workers had won concessions, most of them returned to work, although the immigrant workers had not yet achieved their goals. A small hard core of German trade union activists remained committed to helping the immigrants.^{Footnote 3} Immediately, I wanted to know why, and I wanted to know more about the immigrant partners. What would they have to say about the partnership? Why did these groups cooperate at all?

The following spring of 2012, I returned for follow-up interviews with other activists at the hospital and researched newspaper coverage of the protests to put together a fuller picture of partnership between immigrant and German workers. The fact that I uncovered this fascinating case after changing my sectoral focus, caused me to actively rethink what I had assumed and what I had learned during fieldwork. I considered the likelihood that parallel dynamics in the export-oriented manufacturing sectors and sheltered public sectors probably influenced the way natives and immigrants cooperated in protest at work. When I switched sectors from food manufacturing and elder care to metal manufacturing/autos and public hospitals, I had only thought of manufacturing as the sector where globalization influenced relations between native and immigrant workers. Employers regularly threatened manufacturing unions with offshoring jobs in Eastern Europe and further afield (Harrison Reference Harrison1997; Jacoby Reference Jacoby1995), and sometimes, native workers saw immigrant fellow workers as symbolic scapegoats of foreign competition. In public hospitals, local authorities face limited budgets partly because of globalization and the reduced ability of the state to collect tax revenue from corporations. One of the ways public hospitals can continue to provide services and balance budgets is by contracting out nonclinical services, often to global or European firms (Bach and Givan Reference Bach and Givan2010; Boehlke and Schulten Reference Boehlke and Schulten2008; Broadbent, Gil, and Laughlin Reference Broadbent, Gil and Laughlin2003; Mosebach Reference Mosebach2007). Although immigrants bear the brunt of new low-wage employment created by contracting, this type of privatization is unpopular and may also lead to scapegoating of immigrants.

CONCLUSION AND LESSONS FOR OTHERS

The central lesson of this article is that scholars have to be aware of the likely possibility that one’s research question may change. Even with extensive preparation before fieldwork, it can be hard to know in advance what part of the project will actually be doable and most importantly, what aspect will be the most stimulating. Being open to inspiration and stimulation in the field is valuable because it can propel one through the hard work of writing and rewriting after returning from the field. For scholars of interest groups and organizations in particular, fieldwork showed how even heavily bureaucratized organizations such as unions can be fairly dynamic on the inside. It was fascinating to get a sense of splinter groups and fragmentation in specifically corporatist unions, which had been characterized by the literature as organizations where change happened very slowly. Along this vein, another lesson is to be aware that sometimes the way that the literature describes the phenomena under study can be outdated (Chambers-Ju, this symposium). Although this article discusses questions of scope and setting substantive and temporal thresholds as strategies allowing one to decide when to pare back the parameters of a project, this should not be confused with an admonition to embark only on narrowly defined projects.