Research on strategies for survey design with planned missing data

点击翻译按钮获取中文摘要

Strategies for split questionnaire survey design, where some portion of the items are intentionally set to missing in specific forms, thus enabling the estimation of correlations among more combinations of questionnaire items, were investigated in this study. We assumed two settings on group of items, the one is the case of general items and the other is the case of factor-analytic items, and proposed a evaluation of design strategies for these cases, by considering a utility function or precision of statistical estimation.