本研究採用主計處「家庭收支調查」1991-2000 年資料，探討台灣地區「貧窮女性化」的現象及其特性。研究發現，貧戶中女性戶長家戶比例十年間大致呈現上升，且女性戶長家戶貧窮率每年也都比男性戶長家戶為高；但經由「貧窮比」分析，發現貧窮性別差異十年間頗為穩定。因此，可以說台灣地區「貧窮女性化」現象，只能算是方興未艾！在比較不同家戶組成的分析中，發現「單親戶」貧窮率不是最高的，其趨勢不僅沒有增加反而在1993 年之後有好轉的現象；值得注 意的是，「有依賴子女」的女性戶長家戶具有穩定偏高的貧窮率，可是性別貧窮比也未呈現任何上升的趨勢。最後，根據多變項分析結果，在戶長年齡、戶長教育、依賴子女數等變項的控制下，女性戶長家戶相對於男性戶長家戶具有較高的機會落入貧窮，而「是否為單親」對貧窮的影響則不顯著，這種情形在「有依賴子女」的家戶當中，更為明顯。可以說，「成為單親」並非是台灣女性戶長家戶落入貧窮的主因，而是「有依賴子女」才具有更穩定的解釋力。

The phenomenon of feminization of poverty has been widely discussed in US, mainly due to the increasing proportion of poor who were females, during 1960-1980. Along with the trend of urbanization and modernization, since 1990, Taiwan has begun to experience the increase in divorce rates and concomitantly the increase in number of one-parent families. Using data from the Survey of Family Income and Expenditure (SFIE), 1991-2000 waves, conducted by DGBAS annually, the author intended to examine the hypothesis of feminization of poverty in terms of changes in proportion of poverty rate as well as poverty ratio. Instead of using individuals as analytical units, the author focused on comparing various types of household units and found the trend of increasing proportion of female headed households that were poor. However, no tangible trends were found either between families of male and female householders or among different family types, when poverty ratios were used for testing the feminization hypothesis. When the traits of householders were under controlled, unlike those found in the US and some Western countries, mother-only families in Taiwan were not significantly related to the likelihood of poverty, but the number of cohabiting children had positive effects on being poverty of households during 1990s.