居民收入调查在搜集数据的过程中,经常会遇到一些离群值的出现,如职工买断工龄的收入、居民购买彩票的“大奖”收入、年底“分红”收入或技术失误造成的奇异数据等,从而对收入的统计量产生一定影响。因此必须采用适当的稳健统计方法进行处置,保证提供的数据和指标更加准确、清晰。本文结合收入指标对稳健统计方法进行探讨、解读,其中涉及到对离群值的识别和处理技术,以及如何恰当选择稳健的统计量。 In the process of collecting data, residents ’income survey often encounter some outliers such as the income of workers’ buying seniority, the “grand prize” income of residents purchasing lottery tickets, the singular data caused by the “bonus” income or technical mistakes at the end of the year Etc., which will have a certain impact on the statistics of income. Therefore, we must adopt appropriate and stable statistical methods for disposal, to ensure that the data and indicators provided more accurate and clear. In this paper, we discuss and interpret the robust statistical methods in terms of income indicators, which involve the identification and processing of outliers and the proper selection of robust statistics.