You are using an old browser with security vulnerabilities and can not use the features of this website.
We live in an aging society. The number of senior citizens has been increasing for years, which presents us with an unresolved problem: Who is going to take care of them? Trained health care workers are desperately sought after by retirement homes and in assisting with home-based care. The care once provided by daughters and daughters-in-law is now being provided by non-family members due to a higher retirement age as well as new work environments and family configurations in western democracies—especially when it comes to round-the-clock care. Germans have found an alternative: foreign caregivers. Prof. Dr. Ulrich Kohler and Prof. Dr. Lena Hipp are investigating such employment relationships with a focus on Polish caregivers in Berlin. In their research, they focus on employment situations that fall a bit outside the norm. To ensure robust results, the researchers and their team are using a technique not yet widely applied in Germany – respondent-driven sampling (RDS), an enhanced form of snowball sampling.
They are the “fairy godmother” in the house: When elderly people in Germany are no longer able manage living alone, their family members often prefer finding a live-in caregiver from another country. Without this assistance, many senior citizens would simply be unable to wash and dress themselves, comb their hair, or even eat or go to the bathroom. This is a challenging – and non-transparent – sector of the labor market. How do things work here? And are the employment conditions perhaps better than their reputation? Kohler and Hipp want to find out more about it. According to the Federal Statistical Office, 330,000 people are employed in mobile nursing. “The German Nurses’ Association estimates that between 100,000 and 800,000 caregivers live and work illegally in private households,” Hipp explains. “Nobody knows the actual number. We want to find out what the situation really is, particularly in Berlin.”
Reluctantly, Maria walks into the interview office. A friendly young man welcomes her. He asks her some questions, mainly about how many colleagues she knows who are doing the same job here in Berlin. Before long, she leaves the room, only to enter the next one. There she is asked to fill out a questionnaire on a computer. She is glad to find that it is written in Polish, so she is able to answer in her native language. She answers the approximately 40 questions, but the interview isn’t over yet. She returns to the hallway, where she opens another door, behind which a young woman is waiting. The student hands her 3 vouchers, invites her to pass them on to colleagues of hers, and gives her some money: 15€. An animated movie produced by students from the Berlin University of the Arts illustrates how this sum can be “multiplied”: For each colleague she recruits for the study, she will receive a certain amount of money from the interview office, which she will be able to collect later. For today, though, she is finished. Maria is glad to know that she is contributing to an important research project. And she is able to remain anonymous, because she was not required to give her name or address. This scenario is part and parcel of the plan designed by Hipp, Kohler, and their team. Given the delicacy of the topic, it first had to be approved by the ethics review board of the Berlin Social Science Center (WZB). “In theory, criminals could also be showing up,” Kohler explains the background. “If someone was working illegally on a large scale, we would be witnesses to a crime.”
The field phase will be starting in July. “We are interested in how the women cope with their task of providing round-the-clock care,” Kohler says. “How do they do it?” he wonders. His research tools include RDS as well as a standardized questionnaire, which covers many topics, including contractual situation, skill levels, job satisfaction, family situation, financial issues, and work-life balance. “Our first objective is to find out more about the situation and motivation of the women working in home-based care and nursing,” Kohler stresses. A second objective, which is particularly important to him, is to test RDS in Germany, where it has yet to be widely applied. Using this sampling technique, difficult-to-reach populations can also be recruited and statistically significant generalizations can be made about certain groups. “It allows us to address so many new issues that have never been examined in sociology – at least not in representative samples,” Kohler is thrilled. There are various snowball sampling techniques in sociology, but RDS is the only one that experts say fulfills the criteria for being able to make truly generalizable statements. There is another major difference: Respondents don’t have to give their names. “We want to make full use of the capital of trust between people,” Kohler explains. “This is essential for the technique to work.”
The project starts with about 10 “seeds”, or initial informants, which the researchers identified in extensive pre-studies. “If each of them recruits five respondents and these five recruit another five, the number could actually explode,” Kohler describes a potential aspect of uncertainty in this technique. If the group grows too quickly, the method will fail, since long recruitment chains are needed to make statistically significant generalizations. All in all, the team plans to interview around 800 respondents. “How the process evolves is not in our hands. This is a risk we have to take.” Kohler is fully aware of this but still wants to test the technique, even if it means offering higher incentives for recruitment, should all else fail.
When the peak phase of the projects starts in July, however, the researchers will not be starting from scratch. Qualitative pre-studies have been conducted to determine how many “seeds” will be needed to start the snowball. Because the researchers did not want to exhaust the Berlin market, they did their pre-studies and tested out their questionnaire in Nuremberg. “You cannot use standardized questions without testing them first,” Kohler explains the procedure. The testing went well and gave them valuable insight into ways to improve certain aspects.
The researchers heard very personal stories. Whether these are isolated cases or representative patterns remains to be seen. “What surprised me,” Kohler says, “is the often obviously good relationship between nursing services and caregivers. In this area, few difficulties were reported. It was also understood that the – mostly female – caregivers had not taken the job out of necessity but because they saw it as a good employment opportunity, one that fit well with their personal circumstances. Some of them even gave up something in their home country for it. “So the initial vague picture seems to be more positive than expected. But it reflects the situation on the surface, as we know. We hope to be able to dig deeper into the matter.” With regard to their technique, the sociologists are also optimistic. The network density required for the study seems to be given. There were also hints that restricting themselves to a comparatively “small” region like Berlin was a good decision. Only then are you really able to make reliable statements. “And we want to have a representative sample,” Kohler adds. “This is the ultimate claim of the method.”
Prof. Dr. Ulrich Kohler studied sociology, history, and law at the University of Constance as well as sociology, economic and social history/modern history, and public law at the University of Mannheim. Kohler has been Professor for Methods of Empirical Social Research at the University of Potsdam since 2012.
Prof. Dr. Lena Hipp studied political sciences, Romance studies, and history. In October 2017, she was appointed Professor of Social Structure Analysis, esp. Labor and Organization, a joint appointment with the Berlin Social Science Center (WZB), where she heads the Research Group “Work and Care”.
Sandra Leumann studied journalism and communication science, sociology, and business administration at Freie Universität Berlin and the University of Potsdam. Since 2015, she has been working at the Berlin Social Science Center (WZB). Currently, she is a research assistant on this project.
Who‘s going to be taking care of grandma? An empirical trial of respondent-driven sampling in the (informal) home-based elderly care sector
Funding: German Ministry of Labor and Social Affairs, Interdisciplinary Social Sciences Research Network
Funding amount: EUR 150,000
Participants: University of Potsdam and Berlin Social Science Center (WZB)
Gabriela Kapfer and Nadia Zeissig, students at the University of the Arts (UdK) Berlin, are accompanying the project for a year as part of the “Visual Society Program”, a collaboration between WZB and UdK. The objective of the collaboration is to challenge disciplinary boundaries in order to open up new avenues of access to socially relevant topics and to present the results of social science research visually and based on analytics. To this end, they attended several survey conferences and documented them in graphic recordings. This enabled them to learn about the working methods of social scientists and offered the researchers new perspectives on their work. The students also provided the graphics for some of the testing material, including the animated movie about how to use the vouchers.
Respondent-driven sampling (RDS) is a sampling technique in sociology research not yet well established in Germany. It provides an alternative to the traditional register-based sampling. In RDS, recruitment follows the snowball principle: Specifically selected initial informants, or “seeds”, are interviewed and given a number of vouchers to be used to recruit new respondents – for which the “seeds” receive additional financial rewards. The new respondents also get recruitment vouchers. Who recruited whom can be traced back using a serial number on the voucher.
For the method to work, the size of the respondents’ network has to be established as well. When RDS succeeds, statistically significant samples of hidden populations can be identified.
A mathematical model of the recruiting process weights the sample to compensate for non-random recruitment patterns. This model is based on a synthesis and extension of two areas of mathematics: Markov chain theory and biased network theory.
Text: Petra Görlich
Translation: Monika Wilke
Published online: Alina Grünky
Contact to the online editorial office: firstname.lastname@example.org