青少年與青年學生在家庭互動、社會資本與心理健康之測量恆等性分析

Abstract

本研究目的旨在探究青少年與青年學生在家庭互動、社會資本與心理健康之測量恆等性分析，藉由不同年齡層的學生，將背景變項分為青少年與青年兩個群體，以檢驗兩個群體的測量恆等性。本研究資料來源為「臺灣貧窮兒少資料庫：弱勢兒少生活趨勢調查」。在研究對象上共計1,069人，青少年與青年學生分別575、494人，研究工具包含「家庭互動」、「社會資本」與「心理健康」量表，潛在變項之測量模型會先以驗證性因素分析確認效度，再透過跨群組恆等性檢驗，確認測量模型是否具備不同年齡層之測量不變性，最後採多群組結構方程模式進行分析，並探討各潛在變項在青少年與青年學生之差異情形。本研究結果呈現如下，首先，驗證性因素分析的結果顯示，家庭互動、社會資本與心理健康量表在青少年與青年學生上在效度與信度分別得到支持；其次，家庭互動、社會資本與心理健康之測量模型在青少年與青年皆具有測量恆等性；再者，家庭互動與社會資本對心理健康影響之多群組分析，在青少年與青年兩群體僅有形貌、結構係數與結構殘差具有測量恆等性，而在測量係數與測量殘差僅部分具有恆等性；最後，針對八個因素進行因素平均數的恆等性檢定時，呈現出個別因素在青少年與青年學生有差異；青少年在家庭凝聚、家庭資源、同儕互動與生活滿意度高於青年，但青少年在家庭支持與希望感的因素平均數低於青年。本研究茲就研究結果進行說明，並提出後續研究與應用的建議。

The purpose of the present study was to investigate assessing measurement invariance of family interaction, social capital and mental health between adolescent and youth students. The data in the research was from Taiwan Database of Children and Youth in Poverty (TDCYP). Participants were 1069 students, including 575 adolescent students and 494 youth students. All of the participants completed the family interaction scale, social capital scale, and mental health scale. First, confirmatory factor analysis of individual measurement models was used to assess the validity of the latent variables. Then, multi-group confirmatory factor analysis was employed to confirmwhether the measurement model had measurement invariance in adolescent and youth students. Finally, multi-group structural equation modeling was conducted to explore the differences of the impact of influencing factors on family interaction, social capital, and mental health between adolescent and youth students. The results are shown as follow. The results of confirmatory factor analysis showed the discriminant validity and scale reliability of family interaction scale, social capital scale, and mental health scale were supported for both adolescent and youth students. The measurement model of family interaction scale, social capital scale, and mental health scale had a good multi-group model fit for adolescent and youth students. The multi-group models of family interaction and social capital on adolescent and youth students supported the configural invariance, structural weights invariance, and structural residuals invariance but structural weights and measurement residuals were not. The test of the equivalence of factor means also represented adolescent and youth students difference in family interaction, social capital, and mental health factors individually. The Adolescent factor means of family cohesion, family resources, peer interaction, and life satisfaction were higher than youth but the youth factor mean of family support and teachers care was higher than adolescent. The implication of the finding for reaserch and practice are discussed.