常用统计方法的SPSS过程_第1页
常用统计方法的SPSS过程_第2页
常用统计方法的SPSS过程_第3页
常用统计方法的SPSS过程_第4页
常用统计方法的SPSS过程_第5页
已阅读5页,还剩12页未读 继续免费阅读

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

1、PAGE PAGE 16常用统计方法的SPSS过程 TOC o 1-3 h z u HYPERLINK l _Toc256430441 1 统计描述 PAGEREF _Toc256430441 h 1 HYPERLINK l _Toc256430442 (1)描述统计量 PAGEREF _Toc256430442 h 1 HYPERLINK l _Toc256430443 (2)几何均数 PAGEREF _Toc256430443 h 1 HYPERLINK l _Toc256430444 (3)统计图 PAGEREF _Toc256430444 h 2 HYPERLINK l _Toc256

2、430445 a.条图 PAGEREF _Toc256430445 h 2 HYPERLINK l _Toc256430446 b. 圆图与构成比条图 PAGEREF _Toc256430446 h 2 HYPERLINK l _Toc256430447 c. 线图 PAGEREF _Toc256430447 h 2 HYPERLINK l _Toc256430448 d. 直方图 PAGEREF _Toc256430448 h 3 HYPERLINK l _Toc256430449 e. 箱图 PAGEREF _Toc256430449 h 3 HYPERLINK l _Toc2564304

3、50 f. 散点图 PAGEREF _Toc256430450 h 3 HYPERLINK l _Toc256430451 2计量资料的统计分析 PAGEREF _Toc256430451 h 3 HYPERLINK l _Toc256430452 (1)两样本均数比较的t检验 PAGEREF _Toc256430452 h 3 HYPERLINK l _Toc256430453 (2)配对t检验 PAGEREF _Toc256430453 h 3 HYPERLINK l _Toc256430454 (3)样本均数与总体均数 PAGEREF _Toc256430454 h 3 HYPERLIN

4、K l _Toc256430455 (4)完全随机设计资料的方差分析 PAGEREF _Toc256430455 h 4 HYPERLINK l _Toc256430456 (5)随机单位组设计资料的方差分析 PAGEREF _Toc256430456 h 4 HYPERLINK l _Toc256430457 (6)重复测量资料的方差分析 PAGEREF _Toc256430457 h 4 HYPERLINK l _Toc256430458 (7)析因分析 PAGEREF _Toc256430458 h 5 HYPERLINK l _Toc256430459 (8)二阶段交叉设计资料的方差分

5、析 PAGEREF _Toc256430459 h 5 HYPERLINK l _Toc256430460 3计数资料的统计分析 PAGEREF _Toc256430460 h 6 HYPERLINK l _Toc256430461 (1)样本率与总体率比较 PAGEREF _Toc256430461 h 6 HYPERLINK l _Toc256430462 (2)两样本率的比较 PAGEREF _Toc256430462 h 6 HYPERLINK l _Toc256430463 (4)多个样本率比较的2检验 PAGEREF _Toc256430463 h 6 HYPERLINK l _T

6、oc256430464 4非参数统计分析 PAGEREF _Toc256430464 h 7 HYPERLINK l _Toc256430465 (1)配对计量资料比较的秩和检验 PAGEREF _Toc256430465 h 7 HYPERLINK l _Toc256430466 (2)两独立样本比较的秩和检验 PAGEREF _Toc256430466 h 7 HYPERLINK l _Toc256430467 (3)两组等级资料比较的秩和检验 PAGEREF _Toc256430467 h 7 HYPERLINK l _Toc256430468 (4)多个独立样本比较的秩和检验 PAGE

7、REF _Toc256430468 h 7 HYPERLINK l _Toc256430469 (5)多组等级资料比较的秩和检验 PAGEREF _Toc256430469 h 8 HYPERLINK l _Toc256430470 (6)随机单位组设计资料的秩和检验 PAGEREF _Toc256430470 h 8 HYPERLINK l _Toc256430471 5. 多元统计分析 PAGEREF _Toc256430471 h 8 HYPERLINK l _Toc256430472 (1)直线回归分析 PAGEREF _Toc256430472 h 8 HYPERLINK l _To

8、c256430473 (2)直线相关分析 PAGEREF _Toc256430473 h 8 HYPERLINK l _Toc256430474 (4)多重回归分析 PAGEREF _Toc256430474 h 9 HYPERLINK l _Toc256430475 (5)协方差分析 PAGEREF _Toc256430475 h 9 HYPERLINK l _Toc256430476 (6)判别分析 PAGEREF _Toc256430476 h 10 HYPERLINK l _Toc256430477 (7)聚类分析 PAGEREF _Toc256430477 h 10 HYPERLIN

9、K l _Toc256430478 a. 样品聚类 PAGEREF _Toc256430478 h 10 HYPERLINK l _Toc256430479 b.指标聚类 PAGEREF _Toc256430479 h 11 HYPERLINK l _Toc256430480 (8)主成分分析 PAGEREF _Toc256430480 h 11 HYPERLINK l _Toc256430481 (9)因子分析 PAGEREF _Toc256430481 h 11 HYPERLINK l _Toc256430482 (10)Logistic回归分析 PAGEREF _Toc256430482

10、 h 12 HYPERLINK l _Toc256430483 a. 反应变量为频数变量 PAGEREF _Toc256430483 h 12 HYPERLINK l _Toc256430484 b. 反应变量非频数变量 PAGEREF _Toc256430484 h 12 HYPERLINK l _Toc256430485 (11)生存分析 PAGEREF _Toc256430485 h 13 HYPERLINK l _Toc256430486 a寿命表法 PAGEREF _Toc256430486 h 13 HYPERLINK l _Toc256430487 bKaplan-Meier法

11、PAGEREF _Toc256430487 h 13 HYPERLINK l _Toc256430488 c. Cox回归模型 PAGEREF _Toc256430488 h 14 HYPERLINK l _Toc256430489 (12) 多项反应Logit模型(multinomial Logistic regression) PAGEREF _Toc256430489 h 14 HYPERLINK l _Toc256430490 a. 一个解释变量 PAGEREF _Toc256430490 h 14 HYPERLINK l _Toc256430491 b. 二个解释变量 PAGEREF

12、 _Toc256430491 h 15 HYPERLINK l _Toc256430492 (13)有序反应变量的累积Logit模型 PAGEREF _Toc256430492 h 161 统计描述(1)描述统计量(用于计量资料)求描述统计量,如均数,中位数,标准差,标准误,最大值,最小值,第2.5、25、50、75、97.5百分位数;求偏度和峰度系数及其标准误;绘制直方图。数据格式:1个反应变量,变量名为“y”。Analyze Descriptive Statistics FrequenciesVariable(s):yStatistics(描述统计量)Percentile Values(百

13、分位数) Percentiles: 2.5 / 25 / 50 / 75 / 97.5(5个百分位点)Central Tendency(集中趋势) Mean Median(均数和中位数) Dispersion离散趋势指标 Std. deviation S.E. mean(标准差和标准误) Distribution(分布形态) Skewness(偏度系数及其标准误) Kurtosis(峰度系数及其标准误) Charts(绘制统计图)Chart Type Histogram(直方图) With Normal Curve(正态曲线)(2)几何均数 数据格式: 1个反应变量,变量名为“x”Transf

14、orm ComputeTarget variable: lgxNumeric Expression: LG10(x) (常用对数变换)Analyze Descriptive Statistics DescriptivesVariable(s):lgx Save standardized values as variablesOption Mean Std. deviation S.E. mean Skewness Kurtosis将求得的均数再求反常用对数,即几何均数。另一种方法:Analyze Reports Case SummariesVariable(s):lgxStatistics G

15、eometric Mean(3)统计图a.条图数据格式: 1个分类变量“死因”,1个反应变量 “死亡率”Graphs Bar Simple & Values of individual casesDefineBars represent: 死亡率Category Labels Variable: 死因 HYPERLINK /statdtedm/SAS_SPSS.htm l 总目录#总目录 b. 圆图与构成比条图 (a)圆图 数据格式: 1个分类变量“死因”,1个反应变量 “构成比”GraphsPie Summaries for groups of casesDefine Other summe

16、ry function (Mean) Variable: 构成比 Define Slices:死因(分类变量) HYPERLINK /statdtedm/SAS_SPSS.htm l 总目录#总目录 (b)构成比条图 数据格式:2个分类变量,“肿瘤”和“year”;1个反应变量 “构成比”。GraphsBarStacked & Summaries for groups of casesDefine Other summary function(Mean)Variable: 构成比Category Axis:yearDefine Clusters by:肿瘤 HYPERLINK /statdte

17、dm/SAS_SPSS.htm l 总目录#总目录 c. 线图数据格式: 1个时间变量(分类变量)“year”,1个反应变量“rate”。GraphsLineSimple & Summaries for groups of casesDefine Other summary function(Mean)Variable: rateCategory Axis:yeard. 直方图 HYPERLINK /statdtedm/SAS_SPSS.htm l 总目录#总目录 数据格式: 1个频数变量 “cases”,1个反应变量“age”Data Weight Cases Weight cases by

18、Frequency Variable: casesGraphsHistogram Variable:age Display normal curve e. 箱图 HYPERLINK /statdtedm/SAS_SPSS.htm l 总目录#总目录 数据格式: 1个分类变量 “state”,1个标识变量“city”,1个反应变量“popu”。GraphsBoxplotSimple & Summaries for groups of casesDefineVariable:popu Category:stateLabel Cases:city f. 散点图 HYPERLINK /statdted

19、m/SAS_SPSS.htm l 总目录#总目录 数据格式: 2个源变量“so2”和“mort”,1个残差变量“zre_1”GraphsScatterSimpleDefineY Axis:zre_1X Axis:so22计量资料的统计分析(1)两样本均数比较的t检验AnalyzeCompare Means Independent-Samples T Test Test Variable x Grouping Variable(s) group(2)配对t检验AnalyzeCompare Means Paired-Samples T TestPaired Variables x1 x2(3) 样

20、本均数与总体均数AnalyzeCompare Means One-Samples T Test Test Variable x Test Value ?(4)完全随机设计资料的方差分析AnalyzeGeneral Linear ModelsUnivariate(1)Dependent:因变量。本例选“沉降率”;Fixed Factors:固定因素。可选多个因素,本例只有一个因素“抗凝剂”。(2) Model Specify ModelFull factorial:包括所有因素的主效应及所有因素不同水平各种组合的交互效应分析。系统默认。Statistics:自定义模型。用户根据需要确定交互作用项

21、。本例选此项,并将“抗凝剂”选入Model框。 Build Terms:分析效应选项。Interaction:考虑因素不同水平各种组合的交互效应分析。Main effects:只考虑主效应,不考虑交互效应。All 2-way:考虑所有2个因素的交互效应。以下同理。(3) Post Hoc:多重比较。多重比较方法以是否满足方差齐性要求(Equal Variances Assumed/not Assumed)分为两大类。(5)随机单位组设计资料的方差分析AnalyzeGeneral linear Models Univariate;Dependent Variable x 选Model,选Cust

22、om,选“批次”、“测量条件”入Model,点Build Terms并选Main effects; 选Continue,然后点击Ok。(6)重复测量资料的方差分析a.单因素 数据格式: 1个标识变量“no”,4个重复测量变量“t0”,“t45”,“t90”,“t145”Analyze General Linear Models Repeated MeasuresNumber of Levels: 4 Add Define Within-Subjects Variables(factor1): t0 / t45 / t90 / t145Options Estimated Marginal Mea

23、ns Display Means for: group / factor1 / group*factor1 Compare main effects Confidence interval adjustment: LSD (none)b多因素 数据格式:1个分组变量“group”,5个重复测量变量“t0”,“t1”,“t2”,“t3”,“t4”Analyze General Linear Models Repeated Measures Number of Levels: 5 Add Define Within-Subjects Variables(factor1): t0 / t1 / t2

24、 / t3 / t4 Between-Subjects Factor(s):groupOptions Estimated Marginal Means Display Means for: group / factor1 / group*factor1 Compare main effects Confidence interval adjustment: LSD (none)(7)析因分析数据格式:2分组变量“druga”和“drugb”,1个反应变量“y”Analyze General linear Models Univariate Dependent: y Fixed Factor(s

25、): druga / drugb Model Full Factorial Include intercept in modelOptions Display Descriptive statisticsPost Hoc Post Hoc Tests for: druga / drugb Equal Variances Assumed LSDPlots Horizontal Axis: druga Separate Lines: drugb(8)二阶段交叉设计资料的方差分析数据格式:3分组变量“treat”,“stage”和“block”,1个反应变量“x”Analyze General Li

26、near Models Univariate Dependent Variable(s): x Fixed Factor(s): treat / stage / block Model Custom Model: treat / stage / block3计数资料的统计分析(1)样本率与总体率比较数据格式:1个分组变量“受孕”,1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyze Nonparametric Test BinomialTest Variable List: 受孕Test Proportion: 0.55(2)两样

27、本率的比较数据格式:2个分类变量,“肿瘤类型”和“淋巴转移”;1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyze Descriptive Statistics CrosstabsRow(s): 肿瘤类型Column(s): 淋巴转移Statistics Chi-squareCells Row配对计数资料比较(McNemar检验)数据格式:2个分类变量,“免疫荧光”和“乳胶凝集”;1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyze Descriptive Stati

28、stics CrosstabsRow(s): 免疫荧光Column(s): 乳胶凝集Statistics McNemar(4)多个样本率比较的2检验数据格式:2个分类变量,“检测方法”和“检测结果”;1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyze Descriptive Statistics CrosstabsRow(s): 检测方法Column(s): 检测结果Statistics Chi-squareCells Row4非参数统计分析(1)配对计量资料比较的秩和检验数据格式:2个反应变量分别为“原法”和“新法”Analy

29、ze Nonparametric Tests 2 Related SamplesTest Pair(s) List: 原法新法 Test Type: Wilcoxon (2)两独立样本比较的秩和检验数据格式:1个分组变量“group”,1个反应变量 “r1值”Analyze Nonparametric Tests 2 Independent SamplesTest Variable List: r1值Grouping Variable: group Test Type: Mann-Whitney U (3)两组等级资料比较的秩和检验数据格式:1个分组变量“group”,1个反应变量 “含量”,

30、1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyze Nonparametric Tests 2 Independent SamplesTest Variable List: 含量Grouping Variable: group Test Type: Mann-Whitney U (4)多个独立样本比较的秩和检验数据格式:1个分组变量“药物”,1个反应变量 “死亡率”Analyze Nonparametric Tests K Independent SamplesTest Variable List: 死亡率Grouping Va

31、riable: 药物 Test Type: Kruskal Wallis H (5)多组等级资料比较的秩和检验数据格式: 1个分组变量“疾病”,1个反应变量 “白细胞”,1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyze Nonparametric Tests 2 Independent SamplesTest Variable List: 白细胞Grouping Variable: 疾病 Test Type: Kruskal Wallis H(6)随机单位组设计资料的秩和检验数据格式: 4个反应变量分别为“频率a”,“频率b”

32、,“频率c”和“频率d”Analyze Nonparametric Tests K Related SamplesTest Variables: 频率a 频率b / 频率c / 频率d Test Type: Friedman5. 多元统计分析(1)直线回归分析数据格式: 1个分组变量“g”,1个自变量“x”,个因变量“y”(将两组拆分,分别做回归分析)DataSplit Files Organize output by groupsGroups Based on: g)Analyze Regression Linear Dependent: y Independent(s): x(2) 直线相

33、关分析数据格式:2个列变量分别为“x”和“y”Analyze Correlate BivariateVariables: x / yCorrelation Coefficients Pearson Spearman(3)曲线拟合先变换(以常用对数变换为例),然后按(一)直线回归分析处理Transform Compute Target Variable: x Numeric Expression:LG10(x) Type & Label Use expression as labelAnalyze Regression Linear Dependent: y Independent(s): x(

34、4)多重回归分析a所有变量放入放入模型 数据格式: 3个列变量Analyze Regression Linear Dependent: diam Independent(s): temp / time Method: EnterStatistics Estimates Confidence interval Model fit Descriptivesb逐步回归 数据格式: 3个列变量Analyze Regression Linear Dependent: y Independent(s): x1 / x2 / x3 / x4 / x5 / x6 Method: Stepwise / Back

35、ward(前进/后退法)Statistics Estimates Confidence interval Model fit Descriptives(5)协方差分析数据格式:1个分组变量“treat”,1个反应变量 “post”,1个协变量“pre”Analyze General Linear Models Univariate Dependent Variable(s): post Fixed Factor(s): treat Covariate(s): preOptions Descriptive statistics(6)判别分析数据格式: 1个分类变量 “group”,3个自变量An

36、alyze Classify Discriminant Grouping Variable: group(1 / 2) Independents: x1 / x2 / x3 Enter independents togetherStatistics Fishers UnstandardizedClassifyPrior Probabilities Compute from group sizeDisplay Casewise results Summary tableSave Predicted group membership Discriminant scores (7)聚类分析a. 样品

37、聚类数据格式:10列数据,其中1个标识变量 “no”Analyze Classify Hierarchical Cluster Variable(s): x1x9Cluster CasesDisplay Statistics / PlotsStatistics Agglomeration schedulePlots Dendrogram All clusters VerticalMethod Cluster: Between-groups linkage Measure: Interval: Squared Euclidean distanceSaveCluster membership Si

38、ngle solution: 2 Clustersb.指标聚类将9个变量聚类Analyze Classify Hierarchical Cluster Variable(s): x1x9Cluster VariablesDisplay Statistics / PlotsPlots Dendrogram All clusters VerticalMethod Cluster: Between-groups linkage Measure: Interval: Squared Euclidean distance(8)主成分分析数据格式:3个列变量Analyze Data Reduction F

39、actorVariables: sgpt / f / znt / afpDescriptives KMO and Bartletts test of sphericity(9)因子分析数据格式: 10个列变量,其中1个为标识变量“t”Analyze Data Reduction FactorVariables: x1x9Descriptives Initial solution KMO and Bartletts test of sphericity ExtractionMethod Principal axis factoring(可尝试其他选择) Correlation matrix Un

40、rotated factor solution Scree plot Number of factors 2-5 RotationMethod Equamax(可尝试其他选择) Rotated solution Maximum Iterations for 25 Scores Save as variables Method Regression Options Suppress absolute values less 0.30 (10)Logistic回归分析a. 反应变量为频数变量数据格式: 3个分类变量,“case_ctr”,“drinking”和“smoking”,其中,“case_

41、ctr”为反应变量;1个频数变量“freq”Data Weight Cases Weight cases by: freqAnalyzeRegressionBinary LogisticDependent: case_ctrCovariates: drinking / smoking Method: EnterOptions CI for exp 95% Last stepb. 反应变量非频数变量数据格式: 10个列变量,其中,1个标识变量“no”,1个反应变量“y”,其余8个自变量AnalyzeRegressionBinary LogisticDependent: yCovariates:

42、x1 x8 Method: Enter / Forward LRCategorical Categorical Covariates: x1 / x7SavePredicted Values Probabilities Group membershipOptionsStatistics and Plots Classification plots CI for exp 95%Display Last step(11)生存分析数据格式:“group”和“status”,1个生存时间变量“time”a寿命表法Analyze Survival Life Tables Time: timeDispla

43、y Time Intervals 0 through 60 by 1 Status: status Define Event: Single value: 1 Factor: groupOption Life table(s) Plot SurvivalCompare Levels of First Factor PairwisebKaplan-Meier法Analyze Survival Kaplan-MeierTime: timeStatus: status Define Event: Single value: 1 Factor: groupCompare Factor Log rank / Breslow / Tarone-WareOptions Statistics Survival table(s) / Mean and median survival

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

评论

0/150

提交评论