本研究目的有二：（1)發展「家務性別角色量表（Domestic Gender Role Scale, DGRS )」與「家 務性別角色内隱聯結測驗（Domestic Gender Role Implicit Association Test, DGR IAT)」，測量台 灣人家務分工的外顯與内隱性別刻板印象；及（2 )檢驗DGRS與DGR IAT信、效度。本研究先 招募92位Y世代台灣人參與預備性研究，完成DGRS初步發展與DGR IAT家務工作的詞彙選 擇；再招募99位Y世代與50位嬰兒潮世代台灣人，共149位成為正式研究參與者以檢驗這兩個 測驗的信、效度。删題後之驗證性因素分析結果證實DGRS的因素結構與預期相符，包含男、女、 中性家事三面向；因素結構與實際觀察資料適配度良好。除了嬰兒潮世代男性家事分量表信度係 數略低於.70之外，DGRS三分量表的alpha值與組合信度皆高於.70,具有良好的信度。DGRIAT 方面，僅在嬰兒潮世代先進行和諧版施測的信度係數高於.70，仍有部分改善空間；但其IAT效 果已達顯著，證實DGR IAT適用於台灣人家務分工的内隱性別刻板印象測量。透過相關分析證 實DGR IAT與DGRS之間，呈現低度相關，顯示兩者所測量的構念不盡相同，具有區別效度， 且DGRS三個因素之間，呈現中度相關，證實DGRS之聚斂效度。以新發展測驗而言，研究結 果初步證實DGRS與DGR IAT具有構念效度、DGRS的信度良好，但DGR IAT的信度仍可進一 步改善；後續研究可招募不同教育程度、婚姻狀況的台灣人參與研究，驗證其跨樣本的穩定性及 應用性。

The purpose of this study was to develop and examine the psychometric properties of two new instruments-the Domestic Gender Role Scale (DGRS) and the Domestic Gender Role Implicit Association Test (DGR IAT)-in assessing Taiwanese people’s explicit and implicit gender stereotypes regarding domestic labor. In a pilot study, 92 Taiwanese adults were recruited to obtain a sample of genderized household chores for development of the initial DGRS and DGR IAT versions. In the main study, another 149 Taiwanese people were recruited on a voluntary basis to examine the psychometric properties of the DGRS and DGR IAT. Confirmatory factor analysis of this sample indicated that the DGRS was a three-factor model. Except for the DGRS subscale for male baby boomers, the Cronbach’s alpha values and composite reliability coefficients of the three DGRS subscale scores were all above .7, indicating satisfactory reliability. The significant IAT effects of the DGR IAT indicated that it was a valid instrument for assessing implicit gender stereotypes regarding domestic labor. Its reliability coefficient was over .7 only when assessing Taiwanese baby boomers using the congruent association-first version of the DGR IAT. Moreover, the weak association between its IAT effects and the three DGRS subscale scores revealed the divergent validity of these two instruments, whereas the moderate correlation among the three DGRS subscale scores indicated its convergent validity. As newly developed instruments, the DGRS demonstrated suitable reliability and validity, whereas the DGR IAT had suitable validity but required further modifications to improve its reliability. This study provided only an initial validation of these two instruments, and further research should validate their psychometric properties using samples with diverse education levels and marital statuses.