Political scientist and ACOR-CAORC fellow Rana B. Khoury was in Jordan during fall 2016 researching networks of Syrian activists. She writes below about her research methodology.
As researchers, we ask many questions related to the characteristics of populations. How many voters plan to go to the polls on election day? How satisfied are citizens with a country’s economic performance? We could survey every member of a population to answer these questions, but conducting such a census is immensely expensive and impractical. It’s also unnecessary because we can make good estimates of a population based on just a sample. Randomly selecting individuals from a sampling frame—such as a voter registry or a telephone directory—allows us to estimate population characteristics with a certain amount of confidence. Collect a sample, understand a population—statistics can be beautiful.
Beautiful, unless a sampling frame for our target population does not exist. How can we study populations that are not listed anywhere? Or those who engage in informal behaviors? Hard-to-reach or “hidden” populations might include groups such as undocumented migrants, the self-employed, or jazz musicians. We can use a variety of non-probability sampling methods to reach these people, for example by “snowballing” through their connections. But in these cases we can only draw conclusions about the people in our sample—we cannot infer to the population.
Rather than giving up on the systematic study of hidden populations, researchers have developed an innovative sampling method that approximates probability sampling: respondent-driven sampling (RDS). The driving force of RDS has been the epidemiological study of populations at high-risk of HIV infection. Lately RDS has been adopted for sociological studies, especially of migrant populations. I was excited to encounter one of these studies in a course on “human rights statistics” at Northwestern University, where I am studying for a doctorate in political science. RDS struck me at that moment as a promising way to systematize my study of activists among Syrian refugees.
SUPPORT THE ACOR FELLOWSHIP PROGRAM – Donate to the ACOR Annual Fund
ACOR is proud of the innovative scholarship that the ACOR fellowship program supports. Help ensure that future scholars will be able to enjoy the support of ACOR in Jordan by donating to the ACOR Annual Fund today.
Activism among Syrian refugees is nonconventional activity undertaken on behalf of the Syrian cause. Broadly understood, activism among refugees can include engagement in humanitarian relief, development projects, media and advocacy efforts, political party activities, and so forth. Such behavior is not formally accounted for as a population characteristic—that is, there is no sampling frame of activist Syrians. My prior qualitative research in Jordan did, however, indicate that activists are networked with each other, making them well suited for a network referral method like RDS. In the course of a survey, RDS researchers assess respondents’ social network size to estimate their inclusion probability and weight them into the analysis, allowing for population estimates to be made, thereby approximating a probability sample.
RDS begins with the selection of “seeds,” members of the survey population who we select to be the initial participants of the study. Seeds recruit a limited number of their peers to participate, and the process repeats with each new wave of recruits. The more waves we achieve, the more independent we become of the non-randomly selected seeds.
Maintaining the confidentiality of respondents, each survey is conducted as a 30-minute computer-assisted telephone interview. Respondents answer questions about the nature of their activism, their demographic characteristics and political values. I hope the data collected will help inform us about the development of activism under seemingly adverse circumstances and about the potential impact that activists can have in their communities and on the conflict. While studies of displacement increasingly recognize refugees as political agents, I hope to further our understanding with a systematic study of how that agency is enacted.
Today Syrians constitute the largest refugee population in the world. But this topic extends beyond Syria. The number of people displaced by conflict has dramatically increased in the decades since the Second World War. Rather than disconnected victims, many of these refugees are actively involved in the events around them. Respondent-driven sampling provides a means to better understand how regular people are transforming and transformed by conflict.
Written by Rana B. Khoury
Rana B. Khoury is currently a Ph.D. candidate in the Department of Political Science at Northwestern University. She has an M.A. in Arab Studies (2012) from Georgetown University and B.A. in Political Science (2008) from American University. Her recent book, As Ohio Goes: Life in the Post-Recession Nation, explores the impact of the Great Recession and income inequality on regular Americans, placing their experiences in a political-economic context. To learn more about her work, visit https://ranakhoury.com/