背　景:生死教育乃透過協助學習者面對死亡，探索生命意義，進而建立健康人生觀。這種不斷地反思死亡、生命是需要學習者以批判思考能力來協助達成。因此，在評值生死教育課程時，需要有批判思考技巧融入設計的生死議題來追蹤學生的學習。過去發展22題的「生死批判思考量表」(TCTS-LD）未呈現效度資料，且信度係數為 .51，有必要修訂量表並測試之。目　的:修訂TCTS-LD以評量生死學有關課程的批判思考技能。方　法:採用橫斷性描述研究設計，以兩年時間招募南台灣715位大專學生同意為研究對象。參與學生於課堂中共需填完三種量表（批判思考意向〔CTDS〕，成人版批判思考測驗〔TCTS-A〕，修正版TCTS-LD）。結　果TCTS-LD 15個題目之內在一致性信度KR-20係數為 .54，折半信度之斯皮爾曼-布朗係數為 .47、組內相關係數為 .40，再測信度（n = 22）之皮爾森相關係數為.58（p =.004）、組內相關係數為.56（p = .003）。題目採用以下三種效度檢核方法：.與成人版的批判思考測驗呈顯著相關（TCTS-A; r = .27, p < .001），.國立大學教育系與私立五專四年級護理科兩群學生的平均分數呈顯著差異（t = 2.71, p < .01），.驗證性因素分析得出：假設、評鑑、歸納等3個因素（χ2 = 81.800, p = .158; Normed chi-square χ2/df = 1.169; CFI = 0.976; TLI = .984; RMSEA = 0.015），總共可解釋31.19%的生死批判思考變異量。結　論:此生死批判量表修訂雖有某種程度進展，為提升生死教育效能，增加醫護學生處理生死議題的困境，未來仍須以此15題量表持續進行信度、效度上的測試，以利大專生生死教育相關課程之評量，以及教學策略之研發。

Background:Death education involves acquiring knowledge,changing behavior, and developing proper views of life in both the affective and the value domains. Critical thinking that is honed through reflecting on life-and-death issues represents a way to reach these goals. Designing assessments able to measure college student content and critical thinking skills related to life-and-death issues is thus important. The Test of
Critical Thinking Skills for Life-And-Death content (TCTS-LD)
instrument requires the administration of additional tests to assess reliability and validity for future use in the assessment of
perceptions on life and death.
Purpose:The purpose of this study was to refine the TCTS-LD.
Methods:A cross-sectional, descriptive design was used to recruit 715 college students in southern Taiwan. Three structured scales were administered in class to the participants. Data were collected in 2004 and 2006. Confirmatory factor analysis was applied to validate the structure of scales.
Results:Examination of the reliability of the three-factor and 15-item scale revealed a KuderYRichardson coefficient of internal consistency of .54. The split-half reliability coefficients were .47 in the SpearmanYBrown correlation and .40 in the intraclass correlation
coefficient (ICC). The testYretest reliability coefficients (n = 22)
were .58 in Pearson correlation and .56 in ICC. In addition to content
validity verification by experts and face validity by students,the validity of this test was assessed using three methods,including (a) a comparable validity rating between this test and the TCTS-A (r = .34, p G .001; (b) a contrast-group technique with different responses to the instrument between those in education and nursing majors (t = 2.71, p < .01), with scores of 10.98 (SD =2.42) and 9.82 (SD = 2.25), respectively; and (c) a confirmatory factor analysis confirming that TCTS-LD is related to the three dimensions of assumption, evaluation, and induction (#2 = 81.800,p = .158, normed chi-square #2/df = 1.169, comparative fit index [CFI] = .976, TuckerYLewis index = .984, root mean square error of approximation [RMSEA] = 0.015). Three factors explained
31.19% of total variance for the revised TCTS-LD.
Conclusions:The revised TCTS-LD scale improved performance and effectiveness to a certain degree. However, reliability and construct validity must be further tested to permit its use as an evaluation tool.