实验的设计SCM32.pptVIP

Section 3.2Designing Experiments 實驗設計 實驗設計 對被觀察個體，人、動物或物件，施以處理後，量測有興趣的變數的反應。 實驗設計常用名詞： 被觀察個體稱為實驗單位(experimental unit)，人則稱為被實驗者(subject)； 施行於實驗單位之特定的實驗條件稱為處理(treatment)； 實驗設計中的解釋變數稱為因子(factor)； 每一特定的實驗條件稱為水準(level)。 實驗設計實例 (例3.9)藥的吸收情況： 對25位被實驗者(subjects)注射一種藥劑(treatments)，30分鐘後檢查血液中的藥含量(response)。 一個因子：藥劑 一個水準：劑量 對25位被實驗者(subjects)分別注射三種不同的藥劑，30分鐘後檢查血液中的藥含量。 一個因子：藥劑 三個水準：三種劑量 Another Example (例3.10) Effects of TV advertising Experiment Subjects: undergraduate students Each viewed a 40-minute TV program. Some includes a 30-second commercial on a 35mm camera; others, a 90-second commercial. The same commercial was repeated either 1, 3, or 5 times during the program. Response variables: recall of the ad, their attitude toward the camera, and their intention to purchase it. Two factors: length of commercial (2 levels: 30-second and 90-second) and repetitions (3 levels: 1,3, 5) 6 treatments Advantages of Designed Experiments over Observational Studies Experimentation allows us to study the effects of the specific treatments we are interested in. We can try to control all lurking variables and so see only the effect of the treatments on the response. Experiments can give good evidence for causation. Experimentation allows us to study the combined effects (interaction) of several factors simultaneously. What’s News Randomized comparative experiments provide the best evidence for medical advances. Do newspapers care? Researchers looked at 1192 articles in medical journals, of which 7% were turned into stories by the two newspapers examined. observational studies vs. randomized experiments 37% vs. 25% (journal articles) 58% vs. 6% (newspaper stories) Conclusion: the newspapers want exciting stories, especially bad-news stories, whether or not the evidence is good. 實驗進行程序 進行程序為： 實驗單位 處理 反應 控制實驗環境只允許有不同處理水準，以避免混合效應的情況。 (例3.11)冷凍法對胃潰瘍減輕疼痛的效應： 胃潰瘍病人，施以冷凍法(一般相信可降低胃酸的產生)，觀察疼痛症狀減輕的情況。 冷凍法確實有症狀減輕的效應。 比較性實驗Comparative Experiments 實驗進行程序為： 處理 實驗單位 反應