Abstract
The problem of determining the interval between waves of panel surveys, the sample size, and the study duration is formulated in this study as an optimization problem in which the sum of the cost of estimation error and the cost of the surveys is minimized. Numerical analyses using hypothetical parameter values indicate that the optimal sample size of a panel study can be much smaller, and the optimal inter-wave interval and the optimal study duration much shorter, than those values that have typically been used in previous panel studies.
