The SSRC Eurasia Program was pleased to hold the first installment in the new Webinar Series on Issues in Quantitative Methods in Eurasia Studies. Following on from the Eurasia Program’s two summer 2012 Workshops in Quantitative Methods, the SSRC will offer 4 online, interactive webinars aimed at increasing the quantitative skills of researchers of Eurasia.
Jane Zavisca is an associate professor of sociology at the University of Arizona. In addition to a PhD in sociology, she has an MA and postdoctoral training in statistics. She has designed two original surveys in Russia, as well as worked with secondary surveys such as RLMS and GGS.
This webinar introduces the principles of and rationale for complex survey designs, provides an illustration of the steps involved in designing such a sample, and offers practical advice on how to identify and adjust for complex designs when performing statistical analysis. Examples and advice given are tailored for the Eurasian context. Participants need only have a basic background in statistics (i.e. an introductory graduate-level course that covers simple linear regression).
Most large-scale surveys employ complex sample designs, including multistage samples with clustering and/or stratification, and oversampling of subpopulations. Yet researchers routinely fail to correctly adjust for sample design when performing statistical analyses. Most graduate curricula in quantitative methods present statistical techniques that assume a simple random sample, as does statistical software by default. Incorrect assumptions can lead to incorrect estimates of both sample statistics (e.g. means, regression coefficients) and their standard errors.
A lack of reliable population lists, quality control issues, and cost concerns all present particular challenges for drawing high quality samples in Eurasian countries. The issues at hand, and strategies for solving them, are presented with reference to the Russian Longitudinal Monitoring Survey, and methods for adjusting for the resultant complex design are illustrated using Stata. Lessons learned from this example will enable participants to identify and adjust for analogous issues in their own research using other datasets.?
This webinar series is made possible through generous funding provided by the Department of State, Bureau of Intelligence and Research, Office of Outreach Title VIII Program for Research and Training on Eastern Europe and Eurasia (Independent States of the Former Soviet Union).