本研究旨在探討臺灣青少年性別、城鄉差異、人際關係及學校環境對於學生從國中到高中（職）憂鬱情緒的發展變化。研究樣本取自臺灣教育長期追蹤資料庫（Taiwan Education Panel Survey, TEPS）之2001年、2003年、2005年及2007年追蹤樣本學生的資料，共計3,613人。使用潛在成長曲線模型（Latent GrowthCurve Models）進行長期追蹤資料分析。研究結果發現：一、七至九年級呈現憂鬱情緒成長現象，之後至高中三年級呈現持平狀態沒有明顯成長或下降的變化；二、性別是青少年憂鬱情緒的重要預測因素，女生憂鬱情緒初始狀態高於男生，且憂鬱情緒成長速率也比男生來得高；三、城鄉差異是青少年憂鬱情緒的重要預測因素，鄉村學生憂鬱情緒初始狀態高於城市學生，但兩者在憂鬱情緒成長速率則是沒有差異；四、人際關係對憂鬱情緒具有顯著的負向影響，學生的人際關係愈好，則憂鬱情緒愈低，人際關係成長率愈高的學生，會降低憂鬱情緒成長率；五、學校環境對憂鬱情緒具有顯著負向影響，對學校環境正面評價愈高，將會減少學生的憂鬱情緒，學生對學校環境正面評價的成長率愈高，會降低憂鬱情緒成長速率。文末，根據研究結果進行討論與建議。

This longitudinal study examined the effects of gender, urban-rural differences, interpersonal relationships, and school environments on adolescent depression, as well as the developmental changes in 7th to 12th grade students by using repeated assessments of Taiwanese adolescents. The participants were 3,613 junior and senior high school students selected from the Taiwan Education Panel Survey (TEPS) in 2001, 2003, 2005, and 2007. Latent growth curve models (LGCMs) were used to analyze the longitudinal data. This study revealed the following results: (1) The rate of depressive moods was high among junior middle school students in Grades 7-9; since then, it either remained flat or exhibited no significant change when they entered their final year of high school (i.e., Grade 12). (2) Gender was a significant predictor of adolescent depression. Furthermore, the initial stage of depression was observed more frequently in girls than in boys, and the growth rate in girls was considerably higher than in boys. (3) Urban-rural differences were a significant predictor of adolescent depression, and the initial stage of depression was observed more frequently in rural students than in urban students; however, no difference in the growth rate of depression was observed between the rural and urban students. (4) Interpersonal relationships exerted a significantly negative impact on adolescent depression, wherein the higher-quality interpersonal relationships led to lower levels of depression. Moreover, students with a higher growth rate of interpersonal relationships exhibited relatively lower growth rates in depression. Finally, (5) school environments exerted significantly negative effects on student depression; those receiving a positive assessment from students reduced the level of adolescent depression. Furthermore, higher growth rates resulting from a positive assessment of a school environment lowered the growth rate of depression. This study provides a discussion and recommendations according to the results.