本项目研究信息不完全下基于区间计量模型理论的我国家庭夫妻收入差异研究。家庭内部夫妻收入差异是社会平均性别收入差异的重要组成部分。本项目将开创一套系统性的针对信息不完全情况下数据特征和统计推断特点的区间计量建模技术与统计推断方法，可直接产生家庭夫妻收入的预测区间、测算社会经济因素等对家庭内部收入差异的影响、掌握家庭内部收入差异变化的规律和未来趋势，为相关部门提供决策依据。具体研究内容包括：（1）建立家庭内部夫妻收入差异的区间计量模型，解决传统方法按性别分组建模的信息损失问题（2）信息不完全下区间样本的小域估计（收入的区域差异估计）方法、所得小域估计量的极限理论和统计推断。（3）Generalized Cross Validation技术对区间模型设定偏误的稳健估计和检验方法（4）应用区间计量模型对我国的家庭夫妻收入差异的主要影响因素进行分析测算和预报，并进行国际比较。

To measure the intrahousehold wage inequality in China with incomplete information, we develop an interval-based methodology incorporating the small area estimation and generalized cross validation (GCV) methods. An income interval of a family unit is a direct measure of inequality. Instead of modeling the wage determination equations of male and female separately as in the traditional approaches, the proposed interval method maintains more information since it is accomplished by modeling the wage interval as an inseparable unity. To obtain reliable estimates of the average of the intrahousehold wage inequality under the incomplete information, we employ the small area estimation and the GCV methods for interval data. We develop the interval-based small area estimators, establish the asymptotic theory and the asymptotic efficiency of the proposed estimators, and explore the model specification of interval wage equation using GCV method. The new interval method will be applied to examine the determinants of the intrahousehold wage inequality (for regions) in China, and to help understand the international differences in intrahousehold inequality.