Methodological Issues in the Sampling Design of the Third Wave of TEPS

This paper identifies the basic methodological issues involved in the sampling design of the third wave of Taiwan Educational Panel Survey (TEPS). Central to the design are three design questions (1) sampling of schools, (2) the sampling of panel members from the second wave of junior high students for two more waves of follow-up, and (3) sampling of classes within school. A systematic discussion of successive hypothetical scenarios contributes to resolving potential confusions and concerns with the sampling design. The motivation for a partially fixed school sample design is discussed and justified. The proposed solutions to the three design issues are tightly integr ated and can simultaneously address the need for intercohort tests of institutional and policy impacts, efficient estimation of school effects, avoidance of following up an overly dispersed panel, and maximizing the potential for obtaining class contexts not just for the fresh K11 sample but also for the continuing panel.