Minorities at Risk: Addressing Selection Bias Issues and Group Inclusion Criteria for Ethno-Political Research

点击翻译按钮获取中文摘要

Protracted conflicts over the status and demands of ethnic and religious groups have caused more instability and loss of human life than any other type of local, regional, and international conflict since the end of WWII. In recent years, particularly since the end of the Cold War in 1990, mobilized and repressed ethnic groups have been among the key sources of terrorist recruitment, and have also contributed to international instability in the form of state failure. The Minorities at Risk (MAR) project is thus far the only sustained effort to collect systematic and replicable data on politically-active communal groups and their political actions. Originally designed in the late 1980s, the dataset and associated activities became a public good, used by academics, journalists, governments, and non-governmental organizations to answer a variety of descriptive and analytic questions unanticipated at the project's inception.Intensified attention to MAR revealed limitations and problems with the dataset, and raised questions about selection bias, measurement error and overall usefulness. It seems clear now that MAR's inclusion criteria (discrimination and/or political mobilization) were not systematically applied across countries, and that groups seem selected based on violence or expectation of violence. Concerns with selection bias are sufficient to warrant substantial changes to the current MAR dataset. Furthermore, the dataset in its present form is unwieldy and in need of restructuring as it currently includes over 450 variables, only about one ninth of which are identified as being of significant interest to the community of users as measured by usage in research papers, articles, and books.The project will take specific steps designed to address some of the deficiencies in MAR that have been identified by researchers. We will carefully define and enlarge the universe of ethno-political groups in the dataset to include both politically mobilized and non-mobilized groups, while streamlining the data to include only a subset of current variables. To this end, we propose a three-pronged concrete effort, portions of which are already underway. These efforts center on distinguishing between attributes of extant and proposed MAR groups and disseminating this information to users of the data in its current form, determining the appropriate universe of cases to ameliorate the selection bias and adding to the data and coding of new groups, and streamlining the number of variables. Most importantly, this proposal requests NSF support for this work at a critical moment in MAR's development and at a potential turning point in empirical research on ethno-political conflict, as both the academic and policy communities seek to better understand the dynamics of ethnic conflict.