雙樣本檢定
- 檢定兩個樣本間是否為相同母體
- 獨立資料和成對資料處理方式不同(變異數算法)
不變的道理
- 算合併標準差(pooled standard deviation)
- 寫出拒絕域
- 寫出 p 值提出拒絕 證據

兩獨立樣本
兩組樣本,無法一一對應。例如:
- 採用方法 A 和方法 B 工法的流水線產品瑕疵率
- 國產與進口車平均維修價格差異
X₁ and X₂ represent the annual maintenance costs of domestically produced cars and imported cars, respectively. Random samples of 10 domestically produced cars and 10 imported cars are taken, and their annual maintenance costs are recorded:
| X₁ | 6800 | 5900 | 6300 | 7800 | 8900 | 7500 | 6100 | 10500 | 5400 | 6900 |
|---|---|---|---|---|---|---|---|---|---|---|
| X₂ | 8900 | 9800 | 12350 | 10670 | 9500 | 8700 | 13400 | 9700 | 8600 | 8800 |
Assume that X₁ and X₂ are normally distributed with unknown but equal variances. At a significance level of α = 0.025, test whether the annual maintenance cost of domestically produced cars is lower than that of imported cars.
σ₁², σ₂² 已知 且σ₁²≠ σ₂²
- 兩母體變異數不同
- 兩參數已知
如果有給兩樣本變異數 SE 就會變成:
σ₁², σ₂² 未知 且σ₁²= σ₂²
- 兩母體變異數相同
- 兩參數未知
步驟:
-
寫出假設:
- : (: )
- :
-
計算兩組樣本的平均值與變異數:
- : 兩組樣本的變異數
-
合併變異數(Pooled variance):