本研究係根據現行國中自然學科之課程目標，所進行一系列學科過程技能評鑑工具發展之一。 　　基於前研究所建立之設計實驗技能組織因子之定性結構化模式，於本研究中再延伸為定量結構，並設計以五個層次作為命題之行為目標及評分準則，佐以現行國中自然科教學目標為內容及難度參考依據。另一方面，根據CRT理論，發展並效化設計實驗技能學習層次分析測驗工具，進而實證且修訂前研究之科學過程技能組織因子半經驗模式。 　　研究結果發現國一與國二學生間及國一與國三學生間在設計實驗技能上，不論選擇題型或申論題型的學習層次，皆有顯著差異。但是國二、國三學生間沒有顯著差異，顯示學力調查時，不僅必須注意題型設計之意義且作不同題型應作妥切的數據解釋。另一方面，國三學生設計實驗的技能未見成長，頗值得注意。在此研究中，亦同時發現以「變因」及「定性」所形成的學習層次的量化結構，以作為雙向評量準則的設計，確實可行。 　　在性別的考驗中，於紙筆測驗上，選擇題型未有區別，作答題型則女生優於男生，此發現與英國APU研究結果一致。但是，APU另行再發展實作測驗 ，我國亦宜進行此項研究，以確定那種方式最能公平、確實地反應設計實驗技能的心智能力。

The instrument was developed in order to evaluate junior high school students' technical skills during the process of their taking courses in science. According to the five levels of hierarchy organizors in; designing an evaluation as established in the former study, the objectives and scoring criteria of the instrument were constructed. Also, the contents of the current science text-book were used as a basis for content validity and difficulty reference. The findings showed that there was significant difference in students' performance in the identifying and generating types of evaluation between grade 7 and grades 8,9. One should notice the different degrees of assessment for different types of item designing. Furthermore, the findings supported that the assessing criteria could be constructed by considering the cross interaction designing between the roles of the paramters of variables and qualitative criteria.