187 PROBLEMS OF MEASUREMENT IN CROSS-CULTURAL RESEARCH USING SECONDARY DATA Carol R. Ember, Marc Howard Ross, Michael L. Burton, and Candice Bradley* Behavior Science Research Volume 25, Nos. 1-4 1991 Measurement, the process of linking abstract concepts to empirical indicants (Blalock 1968:12; Zeller and Carmines 1980: 2), is critical in the testing of theory. No matter how sophisticated the derivation of hypotheses, sampling design, and statis- tical analysis, theory-testing may fail if measurement is inadequate. After briefly examining basic measurement concepts, we review a number of problems of measurement in the use of secondary data in cross-cultural research. We order our review by stages of the research process, from the design of measures to the data analysis, and we suggest some possible solutions to a variety of measurement prob- lems. Since researchers face problems that vary with the research question, the type of theory, and the resources available, we hope that our discussion builds aware- ness of the strengths and weaknesses of various solutions to the different problems of measurement usmg secondary data and, in the process, improves the quality of cross-cultural research. By &dquo;research using secondary data&dquo; we are referring here to studies that code data from ethnographic reports and then use these codes to develop quantitative generalizations across a sample of cases within a region or throughout the whole world. Although we focus on measurement here, one of our central premises is that measurement concerns cannot be separated from questions of theory. Theory specifies which variables are of interest and models how variables may be inter- related. Ultimately, we argue, theory is critical not only in the selection of concepts of interest but also in our evaluation of the adequacy of the measures that are employed and in the interpretation of results. Finally, although it is not a point we develop here, cross-cultural studies using secondary data represent only one way of evaluating propositions about human behavior; propositions also need to be *Carol R. Ember is Professor of Anthropology at Hunter College of the City University of New York and Research Associate at the Human Relations Area Files. Marc Howard Ross is Professor of Political Science at Bryn Mawr College, Bryn Mawr, Pennsylvania. Michael L. Burton is Professor of Anthropology at Umversity of California, Irvine, Califorma. Candice Bradley is Assistant Professor of Anthropology at Lawrence University, Appleton, Wisconsin. c~ 1991 HRAF at UNIV CALIFORNIA IRVINE on January 12, 2015 ccr.sagepub.com Downloaded from