衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn) 秩和檢驗(yàn)-Npar1way過(guò)程課件

秩和檢驗(yàn)（非參數(shù)檢驗(yàn)）實(shí)習(xí)十二

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程目的要求掌握利用univariate過(guò)程完成配對(duì)設(shè)計(jì)資料的非參數(shù)檢驗(yàn)；掌握利用npar1way過(guò)程及選擇項(xiàng)Wilcoxon的使用完成完全隨機(jī)設(shè)計(jì)兩組、多組資料的秩和檢驗(yàn)；熟悉利用freq過(guò)程完成隨機(jī)區(qū)組設(shè)計(jì)資料的秩和檢驗(yàn)且能對(duì)結(jié)果作分析。衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程非參數(shù)統(tǒng)計(jì)的適用情況（1）等級(jí)資料；（2）偏態(tài)分布；（3）分布不明；（4）個(gè)別數(shù)據(jù)偏離過(guò)大；（5）各組方差明顯不齊。衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程非參數(shù)統(tǒng)計(jì)的優(yōu)缺點(diǎn)優(yōu)點(diǎn)：不受總體分布的限制，適用范圍廣。缺點(diǎn)：沒(méi)有充分利用資料提供的信息。檢驗(yàn)效能低于參數(shù)檢驗(yàn)。適用參數(shù)檢驗(yàn)條件的資料，應(yīng)首選參數(shù)檢驗(yàn)。衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程

例1

某研究者欲研究保健食品對(duì)小鼠抗疲勞作用，將同種屬的小鼠按性別和年齡相同、體重相近配成對(duì)子，共10對(duì)，并將每對(duì)中的兩只小鼠隨機(jī)分到保健食品兩個(gè)不同的劑量組，過(guò)一定時(shí)期將小鼠殺死，測(cè)得其肝糖原含量（mg/100g），結(jié)果見下表，問(wèn)不同劑量組的小鼠肝糖原含量有無(wú)差別？

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程配對(duì)資料的秩和檢驗(yàn)dataSAS1;inputx1x2@@;d=x1-x2;Cards;620.16958.47866.50838.42641.22788.90812.91815.20738.96783.17899.38910.92760.78758.49694.95870.80749.92862.26793.94805.48;衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程proc

univariate

normal;vard;run;

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程TestsforLocation:Mu0=0Test-Statistic-pValueStudent'stt-2.25916Pr>|t|0.0502SignM-3Pr>=|M|0.1094SignedRankS-21Pr>=|S|0.0313TestsforNormalityTest--StatisticpValueShapiro-WilkW0.843269Pr<W0.0483Kolmogorov-SmirnovD0.230079Pr>D0.1362Cramer-vonMisesW-Sq0.106441Pr>W-Sq0.0824Anderson-DarlingA-Sq0.657567Pr>A-Sq0.0625符號(hào)秩和衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程單樣本秩和檢驗(yàn)dataa;inputnf@@;cards;2.152.102.202.122.422.522.622.722.993.193.374.57;proc

univariate

normal

mu0=2.15;varnf;run;

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程TestsforLocation:Mu0=2.15Test-Statistic-pValueStudent'stt2.898449Pr>|t|0.0145SignM3.5Pr>=|M|0.0654SignedRankS29.5Pr>=|S|0.0059TestsforNormalityTest--StatisticpValueShapiro-WilkW0.837966Pr<W0.0262Kolmogorov-SmirnovD0.182275Pr>D>0.1500Cramer-vonMisesW-Sq0.092826Pr>W-Sq0.1281Anderson-DarlingA-Sq0.651796Pr>A-Sq0.0704衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程

例2為比較兩種刺激物的對(duì)人的影響力的持續(xù)時(shí)間，現(xiàn)隨機(jī)抽取13個(gè)試驗(yàn)對(duì)象接受第一種刺激物，6個(gè)試驗(yàn)對(duì)象接受第二種刺激物。持續(xù)時(shí)間如下表，問(wèn)二種刺激物的持續(xù)時(shí)間是否有差異？?jī)瑟?dú)立樣本秩和檢驗(yàn)衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程

持續(xù)時(shí)間表(minutes)刺激物1組1.941.942.922.922.922.923.273.273.273.273.703.703.74刺激物2組3.273.273.273.703.703.74衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程PROCNPAR1WAY過(guò)程格式

PROCNPAR1WAY

Wilcoxon

；

CLASS

變量；

VAR

變量表；

RUN；衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程datasas1；doa=1to2；inputn；doj=1ton；inputx@@；output；end；end；cards；衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程

131.941.942.922.922.922.923.273.273.273.273.703.703.7463.273.273.273.703.703.74；

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程procsort；排序過(guò)程bya；run；procunivariatenormal；考慮各組的正態(tài)性varx；bya；run；procnpar1waywilcoxon；classa；varx；exact；輸出精確的概率run；

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程結(jié)果衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程TheNPAR1WAYProcedureWilcoxonScores(RankSums)forVariablexClassifiedbyVariableaSumofExpectedStdDevMeanaNScoresUnderH0UnderH0Score113110.50130.011.0047848.5002679.5060.011.00478413.250

每組得分和H0成立時(shí)的得分和H0成立時(shí)的標(biāo)準(zhǔn)差平均得分衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程Averagescoreswereusedforties.WilcoxonTwo-SampleTestStatistic(S)79.5000

NormalApproximationZ1.7265One-SidedPr>Z0.0421Two-SidedPr>|Z|0.0843tApproximationOne-SidedPr>Z0.0507Two-SidedPr>|Z|0.1014T統(tǒng)計(jì)量正態(tài)近似法結(jié)果（超出表范圍時(shí)使用）近似t檢驗(yàn)結(jié)果衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程

ExactTestOne-SidedPr>=S0.0527

Two-SidedPr>=|S-Mean|0.1054小樣本時(shí)可通過(guò)加exact選項(xiàng)要求輸出精確的概率。大樣本時(shí)可直接采用正態(tài)近似的結(jié)果。衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程例3多組獨(dú)立樣本資料的秩和檢驗(yàn)衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程datasasbook；doa=1to4；doj=1to10；inputx@@；output；end；end；cards；衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程1.231.141.631.001.352.011.641.131.011.702.590.963.960.231.612.941.963.680.542.592.503.923.330.232.073.902.842.423.723.000.400.512.490.742.592.363.003.313.213.12；procsort；bya；run；

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程procunivariatenormal；bya；varx；run；procnpar1waywilcoxon；classa；varx；run；

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程WilcoxonScores(RankSums)forVariablexClassifiedbyVariableaSumofExpectedStdDevMeanaNScoresUnderH0UnderH0Score110130.00205.032.00661013.000210202.50205.032.00661020.250310275.00205.032.00661027.500410212.50205.032.00661021.250

衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程Averagescoreswereusedforties.Kruskal-WallisTestChi-Square7.7513DF3Pr>Chi-Square0.0514k>3,且每組例數(shù)均大于5，則此時(shí)不能查表，且H值近似服從自由度為k-1的卡方分布。衛(wèi)生統(tǒng)計(jì)學(xué)實(shí)驗(yàn)秩和檢驗(yàn)-Npar1way過(guò)程例4：隨機(jī)區(qū)組設(shè)計(jì)的秩和檢驗(yàn)dataex107;dob=1

to

8;doa=1

to

4;inputx@@;output;end;end;cards;8.49.69.811.711