本研究應用自編之「對科學的態度量表」(AttitudesToward Science Scale, ATSS)、「自然科成就測驗」(Science Achievement Test, SAT)、「國小學生學習情況調查問卷」、「國小自然科教學情況調查問卷」等四項工具，探討臺北地區1371位國小高年級學生之對科學的態度，以及其與科學學習成就和相關變項之關係。
「對科學的態度量表」(ATSS)與「自然科成就測驗」(SAT)兩項工具的內部均質性信度，以Cronbach α係數表示時，分別為0.96和0.80，構念效化考驗的結果顯示，這兩項工具之內容效度、同時效度、構念效度均稱滿意，顯示這兩項工具為可信、可靠之科學教育屬性評測工具。
研究結果顯示，全體樣本均具正向的「對科學的態度」，對科學的態度在性別之間並無差異，惟六年級學生之對科學的態度較五年級稍低。全體樣本之科學學習成就狀況則不盡理想，六年級學生之表現較五年級要佳，男女學生間之科學學習成就並不因性別不同而異。
對科學的態度與科學學習成就及自然科學業成績之間呈低度相關。當將樣本依ATSS得分之高低，分成高、中、低三群來分析時，結果顯示對科學的態度較正向的學生，科學學習成就和自然科學業成績也較好。
本研究同時亦探討了對科學的態度與學生背景和家庭變項、學校變項的關係。在學生背景和家庭變項中，家中是否飼養小動物或栽種植物、對自然科和上自然課的興趣、學習自然科的意願、參與課外自然科學活動的意願、對自然科成績的自我看法、每天花在做功課和看電視的時間、閱讀自然科學課外書籍的頻率等，與對科學的態度間有密切關係存在，而學生是否參加數理科課外補習、數理科才藝班與對科學的態度之間，則無顯著相關。
學校變項中，包含了自然科相關教學資源及教學狀況與自然科教師背景因素兩部份。在自然科相關教學資源及教學狀況部份，學生參與學校舉辦的課外自然科學活動情形、到自然科教室上課的頻率、教師使用校內自然科相關設施的情形、教師在課堂上鼓勵學生發問、以及做示範實驗的頻率、學生在課堂上親自動手做實驗的頻率等，均與對科學的態度間有密切關係存在。至於在自然科教師背景因素部份，教學年資、數理科知識背景、最喜歡任教的科目、和同事交換教學心得等，與對科學的態度間有顯著相關。
當以「對科學的態度」為依變項，進行逐步迴歸分析時，結果顯示有11個獨立變項，包括對上自然課的興趣、閱讀自然科學課外書籍的頻率、花更多時間學習自然科的意願、參與課外自然科學活動的意願、在課堂上親自動手做實驗的頻率、對自然科成績的自我看法、自然科教師的教學年資、自然科教師花額外時間準備自然科教學的頻率、最喜歡的學科、是否參加過學校舉辦的課外自然科學活動、自然科教師的數理科知識背景等，具有顯著預測力，總共解釋了53.6%的變異量。當另以「科學學習成就」(SAT)為依變項，進行逐步迴歸分析時，結果則得到13個具顯著預測力的獨立變項。
本研究的結果不但可供改進國小自然科教學以及科學態度和科學學習成就評量之參酌，而且亦可作為未來改進國小師資教育方案之參考。

In the present study, four instruments, namely, Attitudes Toward Science Scale (ATSS), Science Achievement Test (SAT), Learning Environment Questionnaire for Elementary Students(LEQ), Science Teaching Environment Questionnaire in Elementary School(STEQ), were developed and used to investigate the relationships between attitudes toward science and science achievement, academic achievements, student's background variables, and school variables of elementary students in Taipei area. The subjects included in this study consisted of 1371 fifth and sixth grade students.
The results of validation study showed that the internal-consistency reliability (Cronbach α coefficient) of the ATSS and the SAT were 0.96 and 0.80 respectively. The results also indicated that the content validity, concurrent validity, and construct validity of the two instruments were satisfactory.
The results of the study revealed that the attitudes toward science of the subjects were positive. Attitudes toward science of fifth grade students were significantly better than those of sixth grade students. On the other hand, science achievement of sixth grade students was significantly better than those of fifth grade students. However, no significant differences between male and female subjects on the ATSS and SAT scores were found.
The correlations between attitudes toward science and science achievement and science academic achievement scores were low but significant. When the subjects were categorised into high, medium, and low groups according to ATSS scores, data analysis showed that students scored higher in ATSS performed significantly better on SAT and science academic achievement than those who scored lower.
The relationships between attitudes toward science and student's background variables and school variables were also explored. The student's background variables investigated were whether raise small animals and plants at home or not, the most and least favorite subjects, perception of science class, willingness of learning science, willingness of participating extracurricular activities related to science, self perception of their own science achievement, time spent daily on homework and television-viewing, and frequency of reading extracurricular materials related to science. Data analysis showed that those variables were all closely related to students' attitudes toward science. However no significant correlations were found between amounts of time spent weekly on out-of-school science enrichment programs and students' attitudes toward science.
Two school variables, namely, supply of resources related to science teaching and situation of science teaching, and science teacher's background correlated positively and significantly with students' attitudes toward science. Significant correlations were also found between students' attitudes toward science and several school variables, such as participation of related science activities in school, using laboratory and equipments, frequency of student-initiated questioning, frequency of doing laboratory demonstrations, and opportunity for students to do hands-on laboratory activities. In addition, several science teacher's background variables, such as years of teaching, subject matter knowledge, most favorite teaching subject, and interactions with colleagues were also found to correlate significantly with students' attitudes toward science.
When using attitudes toward science as dependent variable, the results of stepwise multiple regression showed that eleven independent variables, including perception of science class, frequency of reading extracurricular materials related to science, willingness of learning science, willingness of participating extracurricular activities related to science, opportunity for students to do laboratory activities, self perception of their own science achievement, years of teaching, frequency of spent out-work time on preparation of science teaching, most favorite subject, whether or not participated extracurricular activities in school, and teachers' subject matter knowledge were significant predictors. Total variances explained were 53.6%. When science achievement(SAT) was used as dependent variable, the same regression analysis showed that thirteen independent variables were significant predictors.
The results of this study not only will make significant contribution to the improvement of science teaching and assessment of attitudes toward science and science achievement in elementary schools, but also will provide valuable implications for the reforms of elementary science teacher education.