• 作者： 王橙霞; 洪志明
• 作者服務機構： 國立台灣師範大學化學系
• 中文摘要： 在決定標準參照精通評量工具的信度與效度時，傳統的難度指數、區別指數及相關係數的用途不大。本文設計了一個效化過程，首次將S-P表分析怯，應用於修訂及效化熔點測定技能標準參照精通評量工具。效化此工具的方法是用(1)前測與後測結果；(2)量表與紙筆測驗結果；(3)工作成果與測驗成績；及(4)在一套合理數目實驗活動中之成果之間的正相關來達成。本研究的結論為1.對於大二有機化學實驗課程中，一般固體之熔點測定的評量，本工具有滿意內容效度。2.在一套具有代理性合理數目實驗活動的成果，P曲線持續地向下移動，此趨勢顯示出試題的效度。所有試題的CPj值均小於0.20。以Pj對CPj作圖，所有的試題均落在象限A中，這些事實指明所有試題都是良好的。S曲線表示個別學生學習成果的一致性，S曲線持續地向右移動表示了試題有信度。3.此工具的效標關連效度可以由三個S-P表中，P曲線全部持續向下移動觀察出來·對於三個鑑別性屬性，其D*值在後測1為 0.2785，在後測2為0.1965。量表觀察結果與紙筆測驗成績有高的相關(90.9%)。紙筆測驗分數高的學生能準確地測出熔點（具體與確認階層均為86.4%，歸納階層為90.9%）。4.對於學習遷移之建構效度，可以從個別學生的(a)工作成果，(b)在一套合理數目活動中的測驗成績及(c)由大一至大二的學習進步情形來看出。5.此工具的信度，可以由(a)S曲線全部持續向右移動，(b)兩套對等形式的測驗有81.8%的一致性看出。(a)和(b)顯示出學習進步的一致性。6.本研究設計了一個包含S-P表分析法（一種無母數方法）的效化過程，能滿意地用於效化標準參照精通評量工具、此可由熔點測定評量工其的效化得到證明。
• 英文摘要： The traditional indexes of item difficulty, item discriminating power, and the correlationcoefficient are of little value in determining reliability and validity of criterion-referenced masterymeasurement (CRMM). A procedure has been designed in This Study to revise and validate the CRMMconstructed specifically for the capillary melting point determination technique (CMPDT), one of thechemistry manipulative techniques (CMT). The validation procedure includes a novel application ofS-P chart analysis used for the first time. The method of validating the tool uses positive correlations(1)between the results of pre-test/post-test 1 and post-test 1/post-test 2, (2)between the results ofcheck-list and paper-pencil test, (3)between task performance and test result, and (4)among the re-sults of a rational set of assignments. A computer program in BASIC has been written in This Study toquickly obtain necessary information on IBM PC. Conclusions: 1. Content validity of the tool is satis-factory for assessing whether students can perform CMPDT of ordinary cases in a college introductoryorganic chemistry laboratory courses. 2. Item validity is clearly observed in downward shifts of eachitem's portion of the P-curve, showing sensitivity of learning gain during a representative rational setof activities. CPj values of all items are less than 0.20. In the plots of Pj vs. CPj, all items are in Quad-rant A. These facts indicate that all items are of good quality. Item reliability is observed from arightward shift of that portion of the S-curve showing consistency of learning gain of individuals. 3.Criterion-related validity of the instrument is observed from the P-curve as a whole shiftingdownward in the three S-P charts. D* values are 0.2785 for post-test 1 and 0.1965 for post-test 2. Per-formance on the check-list is found to be highly related (90.9%) to performance on the paper-penciltest. Those students(86.4%) for both the concrete and the identity levels and 90.9% for theclassificatory level) with high scores in the paper-pencil test indeed obtained the accurate meltingpoints. 4. Construct-validity with respect to learning transfer is observed in the individal performanceprofile for (a) task performances and (b) test results in a rational set of activities. 5. Reliablity of theinstrument is observed in (a) the rightward shift of the S-curve as a whole and (b)81.8% agreement intwo equivalent halves of items in the test, both (a) and (b) showing consistency of learning gain. 6. Theprocedure designed in This Study involving S-P chart analysis (a non-parametric method) is satisfac-tory in validating assessment tools for classroom CRMM, as demonstrated by the application toCMPDT, a CMT.
• 中文關鍵字： reliability; validity; criterion-referenced mastery measurement; S-P chart analysis; chemistry manipulative technique; melting point determination; assessment; non-parametric method; performance test; item-response pattern
• 英文關鍵字： --