《中医药统计知识学知识题spss操纵步骤及答题格式.doc》由会员分享,可在线阅读,更多相关《中医药统计知识学知识题spss操纵步骤及答题格式.doc(41页珍藏版)》请在淘文阁 - 分享文档赚钱的网站上搜索。
1、,统计描述P.29 例:随机测得某地20名正常人血糖结果如下,编制频数分布表,绘制频数分布图。4934884834904544354124373344955195495255535855413954154514531. Analyze Description Statistics Frequencies Statistic Percentile,Quartiles 2. Analyze Description Statistics Explore Plot Normality plots with testsNValid20Missing0Mean 均数475.3500Std. Error o
2、f Mean 均数的标准误13.76748Median 中位数485.5000Mode 众数334.00aStd. Deviation 标准差61.57005Variance 方差3790.871Skewness 偏度系数-.321Std. Error of Skewness.512Kurtosis 峰度系数.054Std. Error of Kurtosis.992Range 极差,全距 251.00Minimum 最小值334.00Maximum 最大值585.0Sum 合计9507.00Percentiles25435.500050485.500075523.5000a. Multipl
3、e modes exist. The smallest value is shownTests of NormalityKolmogorov-Smirnova(大样本)Shapiro-Wilk(50,小样本)StatisticdfSig.StatisticdfSig.VAR0001.09920.200*.98220.958a. Lilliefors Significance Correction*. This is a lower bound of the true significance.一、统计量n=20, 均数 =475.35 , 中位数M=485.50 , 众数Mo= 334.00,
4、 P25= 435.50 , P75=523.50 , Xmax=585.00, Xmin=224.00, R=251.00, 方差S2=3790.87, 标准差S=61.57二、正态检验假设:H0:数据分布与正态分布无差异, H1:数据分布与正态分布有差异,=0.05,双侧检验。 P=0.958 0.05,拒绝H1 ,接受H0。 结论:数据符合正态分布。一、计量资料单样本t检验P.56 例4-7:已知人参中M物质的含量服从正态分布,u=63.5,今9次测得一批人工培植人参中M物质的含量为40.0、41.0、41.5、41.8、42.4、43.1、43.5、43.8、44.2,推断这批人工培植
5、人参中M物质的含量与野生人参是否相同。解题步骤:1.正态性检验,Analyze Description Statistics Explore Plot,Normality plots with tests;2.单样本t检验,得出P值。AnalyzeCompare Means One Sample T Test One-Sample StatisticsNMeanStd. DeviationStd. Error MeanM物质含量942.36671.40446.46815One-Sample TestTest Value = 0 tdfSig. (2-tailed)Mean Differece9
6、5% Confidence Interval of the DifferenceLowerUpperM物质含量90.4988.00042.3666741.287143.4462一、正态检验假设:H0:M物质含量与正态分布无差异, H1:M物质含量与正态分布有差异,=0.05,双侧检验P=0.8450.05,拒绝H1,接受H0结论:M物质含量符合正态分布,n=9, =42.3667, S=1.40446,远大于2S,可用t检验。二、单样本t检验假设:H0:1=2,即人工培植人参中M物质含量与野生人参相同; H1:12,即人工培植人参中M物质含量与野生人参不相同,=0.05,双侧检验。t= -45
7、.142,d=8,P=6.40410-110.05,拒绝H1,接受H0。结论:配对差值d符合正态分布。二、配对t检验假设:H0:配对差值d的总体均数d=0; H1:d0,=0.05,双侧检验。 对 =42.3667, S对 = 1.00907 S对= 0.31910实 =2.5000, S实 = 0.93095 S实=0.29439t=5.171 , d=9 , = 2.160 S= 1.32094 S=0.41772P= 0.001 0.05,拒绝H0,接受H1 结论:可认为三棱莪术液有抑瘤效果。正态分布资料两样本均数比较的t检验P.60例4-9:某医师研究转铁蛋白测定对病毒性肝炎诊断的临床
8、意义,测得12名正常人和15名病毒性肝炎患者血清转铁蛋白含量的结果如下:正常人:265.4、271.5、284.6、291.3、254.8、275.9、281.7、268.6、264.1、273.2、270.8、260.5患者:256.9、235.9、215.4、251.8、224.7、228.3、231.1、253.0、221.7、218.8、233.8、230.9、240.7、260.7、224.4本例为完全随机设计资料,推断转铁蛋白测定对病毒性肝炎诊断的意义。解题步骤:1. 检验正态性方差齐性,Analyze Description Statistics Explore Normalit
9、y Plots with test untransformedAnalyzeCompare Means Independent -Sample T TestTests of Normality观察对象Kolmogorov-SmirnovaShapiro-Wilk (小样本)StatisticdfSig.StatisticdfSig.转铁蛋白含量dimension1正常人.11612.200*.98612.998患者.14715.200*.92715.45a. Lilliefors Significance Correction*. This is a lower bound of the tr
10、ue significance.Test of Homogeneity of VarianceLevene Statisticdf1df2Sig.转铁蛋白含量Based on Mean1.877125.183Based on Median1.127125.299Based on Median and with adjusted df1.127122.304.300Based on trimmed mean1.798125.192Group Statistics观察对象NMeanStd. DeviationStd. Error Mean转铁蛋白含量dimension1正常人12271.86710
11、.39713.0014患者15235.20714.39073.7157Independent Samples Test转铁蛋白含量Levenes Test for Equality of Variancest-test for Equality of MeansFSig.tdfSig. (2-tailed)Mean DifferenceStd. Error Difference95% Confidence Interval of the DifferenceLowerUpperEqual variances assumed 方差齐1.877.137.40225.00036.66004.9528
12、26.459546.8605Equal variances not assumed 方差不齐7.67524.795.00036.66004.776426.818646.5014一、正态分布检验P正=0.998 0.05 , P患=0.245 0.05结论:血清转铁蛋白含量符合正态分布。二、方差齐性检验F=1.877,P= 0.183 0.05结论:可认为两总体方差齐性。三、两样本t检验 假设:H0:正=患,即正常人与病毒性肝炎患者的转铁蛋白含量相等; H1:正患,=0.05,双侧检验。 n正= 12 ,= 271.867 ,S正= 10.3971 , S 正= 3.0014 n患= 15 ,=
13、 235.207 , S患= 14.3907 , S 患= 3.7157 t=7.402P 0.05, 结论:E-SCF形成率服从正态分布。二、方差分析 假设: H0:各组E-SCF形成率相同; H1:各组E-SCF形成率不全相同,=0.05,双侧检验。SS组间=1978.944 ,d组间=3; SS组内= 296.800 ,d组内=35; MS组间= 659.648 ,MS组内=8.480; F= 77.789, P 0.05,拒绝1,接受H0结论:各组满足方差齐性。四、多重比较对照组: =12.33 ,S=2.291 , S=0.764 淫羊藿组: =32.40 ,S=4.115 , S=
14、1.301 党参组: =20.00 ,S=2.404 , S=0.760 黄芪组: =20.40 ,S=2.366 , S=0.748 对照组与淫羊藿组比较:P 0.05结论:对照组药物与淫羊藿对E-SCF的影响不同;对照组与党参组比较:P 0.05结论:对照组药物与对E-SCF的影响不同;对照组与黄芪组比较:P 0.05结论:对照组药物与黄芪对E-SCF的影响不同;淫羊藿组与党参组比较:P 0.05结论:淫羊藿与党参对E-SCF的影响不同;淫羊藿组与黄芪组比较:P 0.05结论:党参与黄芪对E-SCF的影响相同。积差相关系数的假设检验P.87 例6-1:测得某地10名三岁儿童的体重与体表面积
15、如下,进行相关分析。体重11.011.812.012.313.113.714.414.915.216.0面积5.2835.2995.3585.6025.2926.0145.8306.1026.0756.411解题步骤:1、 作散点图,直线相关;2、 对体重、体表面积进行正态性检验,符合正态性;3、 进行直线相关分析,得出P值。 AnalyzeCorrelate Bivariate Pearson ,Spearman Options Means and standard deviation 服从正态分布,积差相关Pearson;不服从正态分布等级相关SpearmanCorrelations体重体
16、表面积体重Pearson Correlation.918*Sg. (2-tailed).000N1010体表面积Pearson Correlation.918*1Sig. (2-tailed).000N1010*. Correlation is significant at the 0.01 level (2-tailed).一、正态检验体重P=0.790体表面积P=0.153结论:体重与体表面积的P值均0.05,符合正态性。二、积差相关r=0.918三、相关系数的检验假设:H0:相关系数与零无差异; H1:相关系数与零有差异;P0.05,接受H0结论:不能认为2-7岁急性白血病患儿的血小板数
17、与出血症状之间有直线相关关系。四格表的2检验P.123 例8-1:某中医院收治376例胃脘痛患者,随机分成两组,分别用新研制的中药胃金丹和西药治疗,结果如下,探讨两种药物疗效有无差别;组别有效例数无效例数合计有效率胃金丹组271527698.19%西药组742610074%合计3453137691.76%解题步骤:例数 加权 Dataweight case1、 进行X2检验,得出Pearson X2及P值。AnalyzeDescriptive Statistics Crosstabs组别Row,疗效Column,StatisticsChi-Square,Contingency coeffici
18、entCellobserved,expected组别 * 疗效 Crosstabulation疗效Total有效无效组别胃金丹组Count2715276Expected Count253.222.8276.0西药组Count7426100Expected Count91.88.2100.0TotalCount34531376Expcted Coun345.031.0376.0Chi-Square TestsValuedfAsymp. Sig. (2-sided)Exact Sig. (2-sided)Exact Sig. (1-sided)Pearson Chi-Square56.772a1.
19、0001T5Continuity Correctionb53.6191.000Likelihood Ratio49.4681.000N40 或 T1Fishers Eact Test.000.000Linear-by-Linear Association56.6211.000N of Valid Cases376a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 8.24.b. Computed only for a 2x2 tableSymmetric MeasuresValueApp
20、rox. Sig.Nominal by NominalContingency Coefficient.362.000N of Valid Cases376有效观测数40,所有理论频数5,采用Pearson X2结果;若1理论频数5,则采用Continuity Correction X2假设:H0:胃金丹组与西药组疗效无差异; H1:胃金丹组与西药组疗效有差异,=0.05,双侧检验T11= 253.2 T12= 22.8 n= 376 T21= 91.8 T22= 8.2 pearson2 检验 2=56.772 P 0.05,拒绝H0,接受H1结论:认为西药治愈率不同,pearson列联系数C
21、=0.362四格表确切概率法P126 例8-3:研究中药制剂红花散改善周围血管闭塞性病变患者的皮肤微循环状况,以安慰剂作对照,将37个病例随机分到2组,结果如表,分析红花散的疗效。组别改善无效合计改善率红花散组1552075.0安慰剂组3141717.6合计18193748.6解题步骤:例数 加权 Dataweight case2、 进行X2检验,得出Pearson X2及P值。AnalyzeDescriptive Statistics Crosstabs组别Row,疗效Column,StatisticsChi-Square,Contingency coefficientCellobserve
22、d,expected 处理 * 疗效 Crosstabulation疗效Total改善无效处理用药组Count15520Expected Count9.710.320.0安慰剂组Count31417Expected Count8.38717.0TotalCount18937Expected Count18.019.037.0Chi-Square TestsValuedfAsymp. Sig. (2-sided)Exact Sig. (2-sided)Exact Sig. (1-sided)Pearson Chi-Square12.099a1.001Continuity Correctionb9
23、.9131.002Likelihood Ratio12.9281.000Fishers Exact Test.001.001Linear-by-Linear Association11.7721.001N of Valid Cases37a. 0 cells (.0%) have expected count less than 5. The minimum expected count is 8.27.b. Computed only for a 2x2 table假设:H0:红花散无效; H1:红花散有效,=0.05,双侧检验T11= 9.7 T12= 10.3 n= 37 40 T21=
24、 8.3 T22= 8.7 Fishers Exact 检验 P =0.01 0.05,拒绝H0,接受H1结论:认为红花散能够改善周围血管闭塞性病变患者的皮肤微循环状况。三、非参数检验、Ridit分析Wilcoxon符号秩和检验(配对及单样本比较用)P139 例9-1:对12份血清分别用原法(检测时间20分钟)和新法(检测时间10分钟)检测谷-丙转氨酶,结果见下表,检验两种检测结果是否相同?编号123456789101112原法6023695802422121902522038142195新法802001008224024320538220441522431.配对差值Transform Com
25、pute d=实验-对照2.d值正态检验3. Analyze Nonparametric Tests Legacy Dialogs 2 Related SamplesTests of NormalityKolmogorov-SmirnovaShapiro-WilkStatisticdfSig.StatisticdfSig.d.20412.181.93112.392a. Lilliefors Significance CorrectionRanksNMean RankSum of Ranks新法 - 原法Negative Ranks2a5.7511.50Positive Ranks9b6.065
26、4.50Ties1cTotal12a. 新法 原法c. 新法 = 原法Test Statisticsb新法 - 原法Z-1.913aAsymp. Sig. (2-tailed).056一、对差值进行正态检验P=0.392 0.05结论:差值d服从正态分布。/差值不d服从正态分布。二、配对t检验/配对秩和检验假设:H0:原法与新法的效果无差异;H1:原法与新法的效果有差异,=0.05,双侧检验。T(+)=11.50 , T(-)=54.50Z=-1.913P=0.056 0.05 , 拒绝H1,接受H0.结论:尚不能认为两法检测谷-丙转氨酶的结果不同。Wilcoxon检验和Mann-Whitne
27、yU检验(两组比较用)P142 例9-3:为了考察中药葛根对心脏的影响,使用3g/100mL和5g/100mL葛根的剂量,测定大鼠用药后一分钟心肌收缩的抑制率(用药后心肌的收缩量/用药前心肌的收缩量),比较这两种葛根剂量对心肌收缩作用是否相同。3g/100mL抑制率94.4554691.2481.8085.1986.9291.325g/100mL抑制率90.7198.2579.9290.6868.6491.90解题步骤:1 Normality plots with tests2 Analyze Nonparametric Tests Legacy Dialogs 2 Independent Samples test variable list:抑制率 Define Range 1 2 Tests of Normality葛根剂量Kolmogorov-SmirnovaShapiro-WilkStatisticdfSig.StatisticdfSig.抑制率3g/100 ml组.3447.012.7287.0075g/100 ml组.3136.067.8896.314a. Lilliefors Significance CorrectionRanks葛根剂量NMean RankSum of R
限制150内