雖然教育部建立一套補救教學科技化評量系統，但如何從那套系統篩選出適 合各校的補救教學對象一直受到爭議，本研究擬以校本評量資料，針對補救教學 篩選決策歷程，試圖找出適合學校系統的運作模式。本研究以一所新生智力表現 接近理論人口分布之國中為研究對象，透過科技化評量系統篩選測驗、新生智力 測驗、段考學期平均等資料，將學習低成就學生分為五群，並以百分比及描述統 計分析，探究各群組學生學習表現，最後以基本閱讀能力測驗結果作為效標，檢 核學習低成就學生篩選區分效度。研究結果如下：（1）學習低成就學生中約有 3%－6%具基本閱讀能力困難；這些學生可用科技化評量 ≤ 35分找出其中六至八 成。（2）PRIORI-tbt 篩選測驗結果，對於高階基本閱讀能力（閱讀理解）困難高 風險的偵測有良好效度，對於基礎閱讀能力（識字解碼）困難高風險學生的區分 效度則相對低弱許多，故適合以多元篩選指標（智力、段考成績）篩選出識字解 碼困難高風險學生；內在能力與實際學業表現差距愈大的學生，愈可能具顯著識 字解碼困難，宜提供特定補救教學。根據結果提出校本篩選決策的建議。

In order to find an appropriate screen mechanism to make decisions for remediation in junior schools, the study aimed to investigate the empirical data from a school, including Project for Implementation of Remedial Instruction-technologybased testing (PRIORI-tbt), the intelligence test collected in freshman orientation, and formative achievement tests of Chinese as well. Then, we examined the classification accuracy of screen based on Chinese PRIORI-tbt with the basic reading skills tests. All data were analyzed through percentage and descriptive statistical analysis. The major findings were following: (a) There were about 3%-6% low achievers that were at-risk for basic reading difficulties, in which PRIORI-tbt scores ≤ 35 could rule in 40%-60% those who were at-risk for reading problems. (b) PRIORI-tbt shows higher prediction accuracy for reading comprehensive skills; however, its predictive power was lower for decoding skills. Students with greater discrepancy between intelligence and achievement could be more likely to encounter decoding problems; they might need to be supported by specific decoding remediation. The suggestions for schoolbased screen decision-making are made from the result.