官方六西格玛认证

六西格玛 DMAIC 流程 - 分析阶段 - 分析示例

单样本 t 检验示例:
定义: 单样本 t 检验将总体均值与指定参考值进行比较。
情境: 我们有 50 个周转时间观测值,希望了解总体均值是否大于竞争对手 4.5 分钟的标杆。

原假设: 平均 TAT <= 4.5 分钟
备择假设: 平均 TAT > 4.5 分钟

RowTAT (in Mins)TAT (in Mins)TAT (in Mins)TAT (in Mins)TAT (in Mins)
14.824.414.604.684.80
24.764.624.834.494.67
34.525.045.094.484.91
44.714.625.465.014.44
55.094.684.674.634.58
65.034.774.554.204.97
74.865.164.874.944.44
84.684.985.044.624.47
94.714.584.914.554.84
104.494.934.745.504.94
打开完整尺寸图表

周转时间(分钟)


One-Sample t-Test · Turnaround TimeOne-sided alternative: population mean > 4.5 minutes Variable | N | Mean | Std. Dev. | SE Mean | 95% Lower Bound | t | p TAT (minutes) | 50 | 4.7676 | 0.2586 | 0.0366 | 4.7063 | 7.3175 | 1.07031e-09 Measure | Value Reference mean | 4.5 minutes Degrees of freedom | 49 Confidence bound | One-sided 95% lower bound Values are rounded for display. The worked examples use a 5% significance level. Calculations use the displayed data. One-Sample t-Test · Turnaround TimeCalculated from the displayed observations.One-sided alternative: population mean > 4.5 minutesVariableNMeanStd.Dev.SE Mean95% LowerBoundtpTAT(minutes)504.76760.25860.03664.70637.31751.07031e-09MeasureValueReference mean4.5 minutesDegrees of freedom49Confidence boundOne-sided 95% lower boundValues are rounded for display. The worked examples use a 5% significance level.
打开完整尺寸图表

分析结果


结果: 单侧 p 值小于 0.05,因此我们拒绝原假设。数据提供证据表明,我们的平均周转时间超过 4.5 分钟,即我们的流程慢于竞争对手的标杆水平。


双样本 t 检验示例:
定义: 合并方差双样本 t 检验在等方差模型下比较两个总体均值。
情境: 我们有早班和夜班,希望了解两者的总体平均周转时间是否不同。

原假设: 早班与夜班的总体平均周转时间相等。
备择假设: 早班与夜班的总体平均周转时间不同。

RowMorning ShiftMorning ShiftMorning ShiftMorning ShiftMorning ShiftNight ShiftNight ShiftNight ShiftNight Shift
113.7915.5816.7013.9214.0015.5616.3721.6713.37
215.2914.6818.7114.119.1411.2817.6220.7820.95
313.4012.3720.128.8318.9715.0423.1518.3318.02
412.1012.7612.9917.0416.5322.2914.9219.1412.11
516.0017.4118.6712.3810.0216.7820.9617.7517.34
612.5317.0318.6815.2512.6420.1220.8314.6716.65
714.7518.7711.8315.4913.5819.2718.2319.399.89
821.1015.5610.6911.5115.8916.6516.8416.7617.81
914.1813.7917.2614.8415.5614.9718.1913.6515.07
1012.0615.5311.8510.709.8116.3819.3019.9016.72
1115.2515.7417.7711.4414.9215.8113.0915.5418.46
1213.4612.8514.8314.3317.4519.9615.9715.1721.92
1312.0615.4120.5819.7318.3417.3116.9019.4515.23
1414.8612.1613.0611.6714.0216.6622.2310.2519.07
1514.2516.7812.6815.0519.8314.3521.2816.6314.65
1613.978.369.8517.5921.9218.5821.7719.34
1714.7614.6511.5817.4111.8017.0616.9217.37
1813.8318.5820.1915.8211.2114.1814.4017.38
1915.6312.4212.0918.2410.9716.4514.9120.09
2015.3416.4615.9116.1915.4613.5913.6922.73
打开完整尺寸图表

早班与夜班数据


Two-Sample t-Test · Morning and Night ShiftsIndependent groups · pooled-variance t-test Shift | N | Mean | Std. Dev. | SE Mean Morning Shift | 100 | 14.7514 | 2.9019 | 0.2902 Night Shift | 75 | 17.2415 | 2.9349 | 0.3389 Measure | Value Difference: morning minus night | -2.4901 95% CI · lower | -3.3693 95% CI · upper | -1.6109 Pooled standard deviation | 2.9161 t statistic | -5.5902 Degrees of freedom | 173 Two-sided p-value | 8.69123e-08 Values are rounded for display. The worked examples use a 5% significance level. Calculations use the displayed data. Two-Sample t-Test · Morning and Night ShiftsCalculated from the displayed observations.Independent groups · pooled-variance t-testShiftNMeanStd. Dev.SE MeanMorning Shift10014.75142.90190.2902Night Shift7517.24152.93490.3389MeasureValueDifference: morning minusnight-2.490195% CI · lower-3.369395% CI · upper-1.6109Pooled standard deviation2.9161t statistic-5.5902Degrees of freedom173Two-sided p-value8.69123e-08Values are rounded for display. The worked examples use a 5% significance level.
打开完整尺寸图表

分析结果


结果: 双侧 p 值小于 0.05,因此我们拒绝总体均值相等的原假设。早班与夜班的平均周转时间不同,其中早班的样本均值更小。置信区间对应早班减去夜班的差值。


单因素 ANOVA 示例:
定义: 单因素 ANOVA 检验三个或更多总体均值是否全部相等。
情境: 我们比较员工1、员工2、员工3和员工4的平均周转时间。

原假设: 四个总体的平均周转时间均相等。
备择假设: 四个总体的平均周转时间并非全部相等。

RowStaff1Staff1Staff2Staff2Staff3Staff3Staff4Staff4
120.5613.4318.1816.1713.558.9319.4321.40
213.8414.4316.8614.616.559.8219.1514.67
314.3012.5115.2720.1215.8411.8815.5517.75
415.6612.4918.8019.6215.8313.1617.1814.29
515.2618.6415.4414.1514.0714.2018.8419.72
612.7713.8015.1913.7213.0812.8218.0318.98
78.3814.0816.8115.7115.9415.8419.0822.14
814.0116.4916.2118.988.5511.8319.3515.58
913.3415.2519.5414.4014.0012.8216.1016.30
1013.2814.0216.0118.639.9217.5318.5915.17
1114.9513.1219.1818.3415.6311.9619.0515.48
1218.6410.4316.8615.2916.0412.0821.6417.42
1314.2212.0814.8714.1714.2014.1320.2616.74
1415.3013.3317.3417.5313.2413.6120.9417.82
1515.7710.6316.9816.0616.2619.3220.2911.04
1617.2716.8415.5914.12
1719.2220.7311.7619.52
1817.2518.7711.0114.36
1912.9017.2812.4019.40
2015.8016.9810.1719.85
打开完整尺寸图表

员工 1、员工 2、员工 3 和员工 4 的数据


One-Way ANOVA · Four Staff GroupsANOVA · equal population means versus not all equal Source | DF | SS | MS | F | p Between groups | 3 | 475.4027 | 158.4676 | 26.9352 | 9.76896e-14 Within groups | 136 | 800.1284 | 5.8833 | | Total | 139 | 1275.5311 | | | Measure | Value Pooled standard deviation | 2.4256 R-squared | 37.27% Adjusted R-squared | 35.89% Staff summaries · confidence intervals use the pooled standard deviation Staff | N | Mean | Std. Dev. | 95% CI Lower | 95% CI Upper Staff1 | 35 | 14.4986 | 2.5232 | 13.6878 | 15.3094 Staff2 | 35 | 16.9040 | 1.8667 | 16.0932 | 17.7148 Staff3 | 35 | 13.2446 | 2.6995 | 12.4338 | 14.0554 Staff4 | 35 | 17.8637 | 2.5288 | 17.0529 | 18.6745 Individual 95% confidence intervals for the population means Individual 95% confidence intervals for the population means: Staff1 13.6878 to 15.3094, Staff2 16.0932 to 17.7148, Staff3 12.4338 to 14.0554, Staff4 17.0529 to 18.6745 Values are rounded for display. The worked examples use a 5% significance level. Calculations use the displayed data. Reject equal population locations at the 5% level; the overall test does not identify all differing pairs.One-Way ANOVA · Four Staff GroupsCalculated from the displayed observations.ANOVA · equal population means versus not all equalSourceDFSSMSFpBetween groups3475.4027158.467626.93529.76896e-14Within groups136800.12845.8833Total1391275.5311MeasureValuePooled standard deviation2.4256R-squared37.27%Adjusted R-squared35.89%Staff summaries · confidence intervals use the pooled standarddeviationStaffNMeanStd. Dev.95% CI Lower95% CI UpperStaff13514.49862.523213.687815.3094Staff23516.90401.866716.093217.7148Staff33513.24462.699512.433814.0554Staff43517.86372.528817.052918.6745Individual 95% confidence intervals for the population meansStaff1Staff2Staff3Staff411121314151617181920Values are rounded for display. The worked examples use a 5% significance level.
打开完整尺寸图表

分析结果


结果: p 值小于 0.05,提供证据表明四个总体均值并非全部相等。我们拒绝均值相等的原假设。仅靠 ANOVA 无法确定哪些员工配对存在差异;要得出该结论,还需要进行两两比较。


Mood 中位数检验示例:
定义: Mood 中位数检验比较两个或更多组的总体中位数。
情境: 我们比较员工1、员工2、员工3和员工4的周转时间中位数,使用与 ANOVA 示例相同的、每位员工展示的 35 个观测值。

原假设: 四个总体的周转时间中位数均相等。
备择假设: 四个总体的周转时间中位数并非全部相等。

RowStaff1Staff1Staff2Staff2Staff3Staff3Staff4Staff4
120.5613.4318.1816.1713.558.9319.4321.40
213.8414.4316.8614.616.559.8219.1514.67
314.3012.5115.2720.1215.8411.8815.5517.75
415.6612.4918.8019.6215.8313.1617.1814.29
515.2618.6415.4414.1514.0714.2018.8419.72
612.7713.8015.1913.7213.0812.8218.0318.98
78.3814.0816.8115.7115.9415.8419.0822.14
814.0116.4916.2118.988.5511.8319.3515.58
913.3415.2519.5414.4014.0012.8216.1016.30
1013.2814.0216.0118.639.9217.5318.5915.17
1114.9513.1219.1818.3415.6311.9619.0515.48
1218.6410.4316.8615.2916.0412.0821.6417.42
1314.2212.0814.8714.1714.2014.1320.2616.74
1415.3013.3317.3417.5313.2413.6120.9417.82
1515.7710.6316.9816.0616.2619.3220.2911.04
1617.2716.8415.5914.12
1719.2220.7311.7619.52
1817.2518.7711.0114.36
1912.9017.2812.4019.40
2015.8016.9810.1719.85
打开完整尺寸图表

员工 1、员工 2、员工 3 和员工 4 的数据


Mood's Median Test · Four Staff GroupsMood's median test · four groups of 35 observations Measure | Value Overall sample median | 15.61 Chi-square statistic | 29.3714 Degrees of freedom | 3 p-value | 1.871e-06 Lower means at or below the overall median; Higher means above. IQR is Q3 minus Q1. Staff | N | Lower | Higher | Median | IQR | CI Low | CI High Staff1 | 35 | 25 | 10 | 14.08 | 2.65 | 13.34 | 15.26 Staff2 | 35 | 10 | 25 | 16.86 | 3.34 | 16.01 | 17.53 Staff3 | 35 | 26 | 9 | 13.24 | 3.80 | 12.08 | 14.20 Staff4 | 35 | 9 | 26 | 18.59 | 3.94 | 16.74 | 19.35 Individual median confidence intervals · at least 95% coverage Individual median confidence intervals · at least 95% coverage: Staff1 13.3400 to 15.2600, Staff2 16.0100 to 17.5300, Staff3 12.0800 to 14.2000, Staff4 16.7400 to 19.3500 Central order-statistic intervals; attained coverage 95.9%. Quartiles use the (n + 1)p rank rule. Values equal to the overall median join the lower group. Values are rounded for display. The worked examples use a 5% significance level. Calculations use the displayed data. Reject equal population locations at the 5% level; the overall test does not identify all differing pairs.Mood's Median Test · Four Staff GroupsCalculated from the displayed observations.Mood's median test · four groups of 35 observationsMeasureValueOverall sample median15.61Chi-square statistic29.3714Degrees of freedom3p-value1.871e-06Lower means at or below the overall median; Higher means above. IQR is Q3 minus Q1.StaffNLowerHigherMedianIQRCI LowCI HighStaff135251014.082.6513.3415.26Staff235102516.863.3416.0117.53Staff33526913.243.8012.0814.20Staff43592618.593.9416.7419.35Individual median confidence intervals · at least 95% coverageStaff1Staff2Staff3Staff411121314151617181920Central order-statistic intervals; attained coverage 95.9%.Quartiles use the (n + 1)p rank rule. Values equal to the overall median join thelower group.Values are rounded for display. The worked examples use a 5% significance level.
打开完整尺寸图表

分析结果


结果: p 值小于 0.05,因此我们拒绝总体中位数相等的原假设。总体检验并不能确立每一对员工之间都存在差异。这些计数、中位数和置信区间使用每位员工所展示的 35 个观测值。


卡方检验示例:
定义: 卡方检验将分类计数与原假设下的期望计数进行比较。这里,我们比较两个独立组的赞同比例。
情境: 我们希望了解男性组与女性组的赞同和不赞同比例是否不同。

原假设: 两组的赞同比例相同。
备择假设: 两组的赞同比例不同。

MenWomen
Agree1520
Disagree3035
打开完整尺寸图表

卡方检验数据


Chi-Square Test · Agreement CountsAgreement proportions · Pearson chi-square without continuity correction Response | Men | Women | Total Agree | Observed: 15 Expected: 15.75 Contribution: 0.035714 | Observed: 20 Expected: 19.25 Contribution: 0.029221 | 35 Disagree | Observed: 30 Expected: 29.25 Contribution: 0.019231 | Observed: 35 Expected: 35.75 Contribution: 0.015734 | 65 Total | 45 | 55 | 100 Measure | Value Chi-square statistic | 0.099900 Degrees of freedom | 1 p-value | 0.75195 Values are rounded for display. The worked examples use a 5% significance level. Calculations use the displayed data. Chi-Square Test · Agreement CountsCalculated from the displayed observations.Agreement proportions · Pearson chi-square without continuitycorrectionResponseMenWomenTotalAgreeObserved: 15Expected: 15.75Contribution: 0.035714Observed: 20Expected: 19.25Contribution: 0.02922135DisagreeObserved: 30Expected: 29.25Contribution: 0.019231Observed: 35Expected: 35.75Contribution: 0.01573465Total4555100MeasureValueChi-square statistic0.099900Degrees of freedom1p-value0.75195Values are rounded for display. The worked examples use a 5% significance level.
打开完整尺寸图表

分析结果


结果: p 值大于 0.05,因此我们未能拒绝赞同比例相等的原假设。这些数据没有提供组别与回答之间存在关联的证据。此结果并不能证明总体比例完全相同。



您的六西格玛培训

目录

报名参加认证

我们保证,您的免费在线培训将使您通过 Six Sigma 认证考试!

  1. 01

    在线报名

    选择您希望获得的认可认证,完成报名。

  2. 02

    接收访问码

    您的专属考试访问码将立即发送到您的电子邮箱。

  3. 03

    准备并参加考试

    学习免费培训材料,在线回答 50 道选择题。

  4. 04

    获得终身有效的认证

    领取全球有效的证书和可验证的数字徽章。

Certified Six Sigma Master Black Belt™ (CSSMBB™) Certification 证书示例
报名参加认证

您的免费优质在线培训

您的 Six Sigma Revealed 第二版已准备就绪。

我们携手由各主要行业知名企业和团队领导者、Six Sigma 教练、导师、专家及权威人士组成的联盟,为您开发了培训。所有访问 International Six Sigma Institute™ 的人士均可学习。请通过以下栏目查阅材料: 您的免费六西格玛图书 以及 您的免费优质六西格玛培训 获取培训资料。这两个栏目都在顶部菜单中。

这些材料属于 International Six Sigma Institute™ 的知识产权,受版权保护。我们免费向所有人分享这些材料,因为传授专业知识能够帮助 Six Sigma 专业人士,并推动这一领域的发展。

您的认证考试采用选择题。学习在线培训内容,掌握通过考试和获得 Six Sigma 认证所需的知识。

我们保证,您的免费优质在线培训将使您通过 Six Sigma 认证考试!