本文以「華人家庭動態資料庫」為資料來源，探究以追蹤資料分析勞動參與行為時，樣本流失偏誤是否存在。此外，本文區分流失樣本、未流失樣本，探討其勞動參與行為是否具有顯著差異。文中主要參照Zabel（1998）建構計量模型，同時估計「工作與否」、「樣本流失與否」兩條迴歸式。文中的分析結果顯示，流失樣本、未流失樣本在勞動參與比例上，未見顯著差異。而在勞動參與迴歸式中，樣本流失偏誤亦不顯著。顯示在以「華人家庭動態資料庫」的追蹤資料分析勞動參與行為時，樣本流失偏誤問題並不嚴重。

The purpose of this paper is to analyze the attrition bias problem in the Panel Study of Family Dynamics (PSFD) conducted in Taiwan. In particular, the focus of this paper is on the relationship between sample attrition and labor force participation. Following the framework of Zabel (1998), the author constructed an econometric model to estimate the labor force participation equation and the sample attrition equation jointly. The results show that labor force participation rates of attriters and nonattriters are similar in size. Furthermore, the attrition bias in the labor force participation equation is not significant.

華人家庭動態資料庫；樣本流失；勞動參與

Panel Study of Family Dynamics；sample attrition；labor force participation