Presentation is loading. Please wait.

Presentation is loading. Please wait.

Statistics 04-1 Testing for Differences. 平均数的显著性检验(样本与总体) 总体正态分布 总体方差已知 总体方差未知 总体非正态分布 平均数差异的显著性检验(总体与总体) 两组样本独立 两个总体方差 σ 1 2 、 σ 2 2 未知 两个总体方差不等 两组样本相关.

Similar presentations


Presentation on theme: "Statistics 04-1 Testing for Differences. 平均数的显著性检验(样本与总体) 总体正态分布 总体方差已知 总体方差未知 总体非正态分布 平均数差异的显著性检验(总体与总体) 两组样本独立 两个总体方差 σ 1 2 、 σ 2 2 未知 两个总体方差不等 两组样本相关."— Presentation transcript:

1 Statistics 04-1 Testing for Differences

2 平均数的显著性检验(样本与总体) 总体正态分布 总体方差已知 总体方差未知 总体非正态分布 平均数差异的显著性检验(总体与总体) 两组样本独立 两个总体方差 σ 1 2 、 σ 2 2 未知 两个总体方差不等 两组样本相关 相关系数未知 相关系数已知 方差的显著性检验 样本方差与总体方差差异 两个样本方差

3 平均数的显著性检验(样本与总体) 总体正态分布、总体方差已知 Z Test |Z| = | ( Xˉ - μ 0 ) /(σ 0 /SQRT(n))| where Xˉ:mean of the sample μ 0 : mean of the population σ 0 : standard deviation of the population n: size of the sample Note: Z value is estimated on the basis of standard error of the population

4 平均数的显著性检验(样本与总体) 总体正态分布、总体方差未知 t test |t|=| ( X ˉ - μ 0 ) /(s/SQRT(n-1)) | where X ˉ:mean of the sample μ 0 : mean of the population s: standard deviation of the population n: size of the sample Note: t value is estimated on the basis of standard error of the population Z test is also called large sample test. t test is also called small sample test.

5 平均数的显著性检验(样本与总体) 总体非正态分布 When n ≦ 30, nonparametric test is used.(Chapter 10) When n ≧ 30, Z’ test is applied. |Z| = | ( X ˉ - μ 0 ) /(s/SQRT(n)) | where X ˉ:mean of the sample μ 0 : mean of the population s: standard deviation of the sample n: size of the sample

6 平均数差异的显著性检验(总体与总体): 两组样本独立 两个总体方差 σ 1 2 、 σ 2 2 已知 已知: σ 1 2 、 σ 2 2 : 两个总体方差 X 1 ¯ 、 X 2 ¯ :来自两个总体的两个样本平均数 求证: μ 1 和 μ 2 是否存在显著性差异 因此: Z =( ( X 1 ¯ - X 2 ¯ ) -( μ 1 - μ 2 )) / SQRT ( σ 1 2 /n 1 + σ 2 2 /n 2 )

7 Case Two samples of CET-4 scores from two schools of the north and the south respectively. Sample 1 (from the north): Size: n=150 Sample mean: X 1 ¯ =67.5 Standard deviation of Population 1 (from the north): σ 1 =5.6 Sample 2 (from the south): Size: n=460 Sample mean: X 2 ¯ =64.5 Standard deviation of Population 1 (from the north): σ 2 =5.14 平均数差异的显著性检验(总体与总体): 两组样本独立

8 H 0 : μ 1 =μ 2 (There is no significant difference between the means of the two populations) H 1 : μ 1 ≠μ 2 Z =( ( X 1 ¯ - X 2 ¯ ) -( μ 1 - μ 2 )) / SQRT ( σ 1 2 /n 1 + σ 2 2 /n 2 ) = ((67.5-64.5)-0)/SQRT(5.6 2 /1505.14 2 /460) = 5.81 Look up in Norma Distribution Table for the critical Z value: Z (0.01/2) =2.58 Z> Z (0.01/2) H 0 rejected There is significant difference between the means of the two populations. 平均数差异的显著性检验(总体与总体) :两组样本独立

9 两个总体方差 σ 1 2 、 σ 2 2 未知 Given : X ˉ 1 : mean of Sample 1 s 1 2 : variance of Sample 1 n 1 : size of sample 1 X ˉ 2 : mean of Sample 2 s 2 2 : variance of Sample 2 n 2 : size of Sample 2 Calculation : t test Degree of freedom: n 1 + n 2 - 2 T = (X ˉ 1 - X ˉ 2 )/SQRT(((n 1 - 1) s 1 2 +( n 2 - 1 ) s 2 2 )/( n 1 + n 2 - 2))*(1/ n 1 + 1/ n 2 )) 平均数差异的显著性检验(总体与总体) :两组样本独立

10 两个总体方差不等 Given : X ˉ 1 : mean of Sample 1 s 1 2 : variance of Sample 1 n 1 : size of sample 1 X ˉ 2 : mean of Sample 2 s 2 2 : variance of Sample 2 n 2 : size of Sample 2 平均数差异的显著性检验(总体与总体) :两组样本独立

11 平均数差异的显著性检验(总体与总 体):两组样本相关 同一组学生做两次测试 相关系数未知 相关系数已知

12 相关系数未知 Given : X 1i : i raw scores from Test 1 X 2 i : I raw scores from Test 2 S 1 2 : Variance of Test 1 S 2 2 : Variance of Test 2 平均数差异的显著性检验(总体与总 体):两组样本相关

13 相关系数未知 Steps: 1. Calculate d: differences between scores from the two tests d=X 1 - X 2 2. Sum up the differences: dˉ = Σd i 3. Calculate the variance of di : s d 2 = Σ(d - dˉ) 2 /n 4. Calculate t value : t = (Xˉ 1 - Xˉ 2 )/Sqrt(s d 2 /(n - 1)) Critical t value: from t value table: df=n-1 平均数差异的显著性检验(总体与总 体):两组样本相关

14 相关系数已知 Given : X 1i : i raw scores from Test 1 X 2 i : I raw scores from Test 2 S 1 2 : Variance of Test 1 S 2 2 : Variance of Test 2 X ˉ 1 : mean of Test 1 X ˉ 2 : mean of Test 2 r: correlative coefficient t = (X ˉ 1 - X ˉ 2 )/Sqrt((S 1 2 + S 2 2 - 2*r* S 1 * S 2 )/(n - 1)) Critical t value: from t value table: df=n-1 平均数差异的显著性检验(总体与总 体):两组样本相关

15 Case Researches 1. vocabulary size and reading comprehension 2. background knowledge and listening ability 3. nominal structure and productive ability 4. effect of new teaching methodology 5. diet and language ability

16 两个总体方差不等 Cochran-Cox Test t’ = (X ˉ 1 - X ˉ 2 )/SQRT(s 1 2 /n 1 + s 2 2 / n 2 ) critical value: t’ α /2 = ((s 1 2 /n 1 )* t α /2 *( n 1 - 1) + (s 2 2 / n 2 )* t α /2 *( n 2 - 1))/( s 1 2 /n 1 + s 2 2 / n 2 ) 平均数差异的显著性检验(总体与总体) :两组样本独立


Download ppt "Statistics 04-1 Testing for Differences. 平均数的显著性检验(样本与总体) 总体正态分布 总体方差已知 总体方差未知 总体非正态分布 平均数差异的显著性检验(总体与总体) 两组样本独立 两个总体方差 σ 1 2 、 σ 2 2 未知 两个总体方差不等 两组样本相关."

Similar presentations


Ads by Google