本研究旨在探討偏鄉地區國中生家長參與、科學學習動機與一年後科學學習成就三者間的關係。研究對象為425 位八年級國中生，包括221 位男生和204 位女生，平均年齡為14.36 歲，標準差0.48 歲。研究工具為阿拉巴馬教養量表兒童版的正向親職、家長參與分量表，以及國中生科學學習動機量表，以套裝軟體IBM SPSS 18 和Amos 20 進行資料分析，使用結構方程模式進行考驗，並採取兩階段的分析策略：第一階段考驗研究的測量模式；第二階段考驗研究的結構模式，再以多群組分析以考驗假設模式的跨性別不變性，以拔靴法（bootstrapping）考驗科學學習動機的中介效果。研究結果支持家長參與、科學學習動機、科學學習成就之間的理論關係，顯示假設模式具有跨性別的不變性，且科學學習動機在家長參與和科學學習成就間扮演部份中介變項的角色。綜上，研究結果顯示偏鄉地區家長參與越高，子女一年後的科學學習成就越高，當家長的陪伴能適當引發和鼓勵子女的科學學習動機時，更能有助提昇學生的科學學習成就。研究結果與Coleman的社會資本理論一致，並支持Grolnick 與Slowiaczek 的所提出的動機模式。最後依據研究結果提出促進偏鄉地區學生科學動機之相關建議。

This study explored the relationships among parental involvement, science motivation and one-year later science achievements of the junior high school students in rural towns in Taiwan. The research subjects consisted of 425 8th grade junior high school students, including 221 males and 204 females. The mean age of the students was 14.36 years, with a standard deviation of 0.48. The parental involvement and positive parenting subscales of the Alabama Parenting Questionnaire Child Form (APQ-Child Form) as well as the Junior High School Student Science Motivation Scale were used in the study. The software IBM SPSS 18 and AMOS 20 were used for analyzing the data. The structural equation modeling (SEM) techniques were used to evaluate the hypothesized theatrical model, and a two-stage analysis strategy was applied. The fit of the measurement model was estimated in the first phase, and the fit of the structural equation model was estimated in the second stage. Moreover, multiple group analysis was applied in order to evaluate gender differences, and bootstrapping was applied in order to examine the mediation effects of science motivation. The study results offered support for the theoretical relationships among parental involvement, science motivation and science achievements. The multiple group analysis results provided support for the gender invariance of the hypothesized theatrical model, and the bootstrap results confirmed the partially mediating role of science motivation in the relationship between parental involvement and science achievements. In summary, the results demonstrated that parental involvement can impact their children’s science motivation, which in turn can improve the school’s success in science education one year later for the junior high school students living in rural towns. The findings were consistent with Coleman’s theory of social capital and provided evidence that supports Grolnicks and Slowiaczek’s motivational model. Based on the above results, some suggestions were proposed for improving students’ science motivation in rural towns.