多元数据分析
2011-6
机械工业出版社
Joseph F. Hair,William C. Black,Barry J. Babin,Rolph E. Anderson
800
无
《多元数据分析（英文版）（第7版）》是一本面向应用的经典多元数据分析教材，自1979年出版第1版至今，深受读者好评。《多元数据分析（英文版）（第7版）》循序渐进地介绍了各种多元统计分析方法，并通过丰富的实例演示了这些方法的应用。书中不仅涵盖多元数据分析的基本方法，而且还介绍了一些新方法，如结构方程建模和偏最小二乘法等。
作者：（美国）海尔（Joseph F.Hair.Jr.） （美国）William C.Black （美国）Barry J.Babin 等海尔（Joseph F Hair，Jr.），于1971获得佛罗里达大学市场营销博士学位.现为肯尼索州立大学市场营销系教授。他出版了四十多本书，包括《Marketing》、《Marketing Essentials》等。他是美国市场营销协会、市场营销科学学会、西南市场营销协会和南方市场营销学会委员。2004年他被美国市场营销科学学会授予杰出教育奖，2007年被市场管理协会授予创新性市场营销人才。William C.Black，于1980年获得德州大学奥斯汀分校博士学位，现为路易斯安那州立大学工商管理学院市场营销系教授。他的研究兴趣包括多元统计、应用信息技术，以及与电子商务相关的市场原理的进展。他是《Journal of BusinessResearch》编审委员会成员。Barry J.Babin于1991年获得路易斯安那州立大学工商管理学博士学位，现为路易斯安那理工大学市场营销与定量分析学教授、商学院Max P.Watson教授。他主要研究零售的各个方面和服务管理。他还曾被美国市场营销科学研究院和市场营销学会评为杰出研究员。Rolph E.Anderson，拥有佛罗里达大学博士学位，现为Drexei大学工商管理学院R0yal H.Gibson Sr教授。他曾两次获得Drexel大学优秀教师奖，并获得过《Journal of Personal Selling＆Sales Management》杰出评论奖、Drexel大学商学院科研成就奖等。
preface iii
about the authors v
chapter 1 introduction: methods and model building
what is multivariate analysis?
multivariate analysis in statistical terms
some basic concepts of multivariate analysis
the variate
measurement scales
measurement error and multivariate measurement
statistical significance versus statistical power
types of statistical error and statistical power
impacts on statistical power
using power with multivariate techniques
a classification of multivariate techniques
dependence techniques
interdependence techniques
types of multivariate techniques
principal components and common factor analysis
multiple regression
multiple discriminant analysis and logistic regression
canonical correlation
multivariate analysis of variance and covariance
conjoint analysis
cluster analysis
perceptual mapping
correspondence analysis
structural equation modeling and confirmatory factor
analysis
guidelines for multivariate analyses and interpretation
establish practical significance as well as statistical
significance
recognize that sample size affects all results
know your data
strive for model parsimony
look at your errors
validate your results
a structured approach to multivariate model building
stage 1: define the research problem, objectives,
and multivariate technique to be used
stage 2: develop the analysis plan
stage 3: evaluate the assumptions underlying the multivariate
technique
stage 4: estimate the multivariate model and assess overall model
fit
stage 5: interpret the variate(s)
stage 6: validate the multivariate model
a decision flowchart
databases
primary database
other databases
organization of the remaining chapters
section i: understanding and preparing for multivariate
analysis
section ii: analysis using dependence techniques
section iii: interdependence techniques
section iv: structural equations modeling
summary 28 . questions 30 . suggested readings
references
section i understanding and preparing for multivariate
analysis
chapter 2 cleaning and transforming data
introduction
graphical examination of the data
univariate profiling: examining the shape of the
distribution
bivariate profiling: examining the relationship between
variables
bivariate profiling: examining group differences
multivariate profiles
missing data
the impact of missing data
a simple example of a missing data analysis
a four-step process for identifying missing data and applying
remedies
an illustration of missing data diagnosis with the four-step
process
outliers
detecting and handling outliers
an illustrative example of analyzing outliers
testing the assumptions of multivariate analysis
assessing individual variables versus the variate
four important statistical assumptions
data transformations
an illustration of testing the assumptions underlying
multivariate analysis
incorporating nonmetric data with dummy variables
summary 88 . questions 89 . suggested readings
references
chapter 3 factor analysis
what is factor analysis?
a hypothetical example of factor analysis
factor analysis decision process
stage 1: objectives of factor analysis
specifying the unit of analysis
achieving data summarization versus data reduction
variable selection
using factor analysis with other multivariate techniques
stage 2: designing a factor analysis
correlations among variables or respondents
variable selection and measurement issues
sample size
summary
stage 3: assumptions in factor analysis
conceptual issues
statistical issues
summary
stage 4: deriving factors and assessing overall fit
selecting the factor extraction method
criteria for the number of factors to extract
stage 5: interpreting the factors
the three processes of factor interpretation
rotation of factors
judging the significance of factor loadings
interpreting a factor matrix
stage 6: validation of factor analysis
use of a confirmatory perspective
assessing factor structure stability
detecting influential observations
stage 7: additional uses of factor analysis results
selecting surrogate variables for subsequent analysis
creating summated scales
computing factor scores
selecting among the three methods
an illustrative example
stage 1: objectives of factor analysis
stage 2: designing a factor analysis
stage 3: assumptions in factor analysis
component factor analysis: stages 4 through 7
common factor analysis: stages 4 and 5
a managerial overview of the results
summary 148 . questions 150 . suggested readings
references
section ii analysis using dependence techniques
chapter 4 simple and multiple regression
what is multiple regression analysis?
an example of simple and multiple regression
prediction using a single independent variable:
simple regression
prediction using several independent variables:
multiple regression
summary
a decision process for multiple regression analysis
stage 1: objectives of multiple regression
research problems appropriate for multiple regression
specifying a statistical relationship
selection of dependent and independent variables
stage 2: research design of a multiple regression analysis
sample size
creating additional variables
stage 3: assumptions in multiple regression analysis
assessing individual variables versus the variate
methods of diagnosis
linearity of the phenomenon
constant variance of the error term
independence of the error terms
normality of the error term distribution
summary
stage 4: estimating the regression model and assessing overall
model fit
selecting an estimation technique
testing the regression variate for meeting the regression
assumptions
examining the statistical significance of our model
identifying influential observations
stage 5: interpreting the regression variate
using the regression coefficients
assessing multicollinearity
stage 6: validation of the results
additional or split samples
calculating the press statistic
comparing regression models
forecasting with the model
illustration of a regression analysis
stage 1: objectives of multiple regression
stage 2: research design of a multiple regression analysis
stage 3: assumptions in multiple regression analysis
stage 4: estimating the regression model and assessing overall
model fit
stage 5: interpreting the regression variate
stage 6: validating the results
evaluating alternative regression models
a managerial overview of the results
summary 231 . questions 234 . suggested readings
references
chapter 5 canonical correlation
what is canonical correlation?
hypothetical example of canonical correlation
developing a variate of dependent variables
estimating the first canonical function
estimating a second canonical function
relationships of canonical correlation analysis to other
multivariate techniques
stage 1: objectives of canonical correlation analysis
selection of variable sets
evaluating research objectives
stage 2: designing a canonical correlation analysis
sample size
variables and their conceptual linkage
missing data and outliers
stage 3: assumptions in canonical correlation
linearity
normality
homoscedasticity and multicollinearity
stage 4: deriving the canonical functions and assessing overall
fit
deriving canonical functions
which canonical functions should be interpreted?
stage 5: interpreting the canonical variate
canonical weights
canonical loadings
canonical cross-loadings
which interpretation approach to use
stage 6: validation and diagnosis
an illustrative example
stage 1: objectives of canonical correlation analysis
stages 2 and 3: designing a canonical correlation analysis and
testing the assumptions
stage 4: deriving the canonical functions and assessing overall
fit
stage 5: interpreting the canonical variates
stage 6: validation and diagnosis
a managerial overview of the results
summary 258 . questions 259 . references
chapter 6 conjoint analysis
what is conjoint analysis?
hypothetical example of conjoint analysis
specifying utility, factors, levels, and profiles
gathering preferences from respondents
estimating part-worths
determining attribute importance
assessing predictive accuracy
the managerial uses of conjoint analysis
comparing conjoint analysis with other multivariate methods
compositional versus decompositional techniques
specifying the conjoint variate
separate models for each individual
flexibility in types of relationships
designing a conjoint analysis experiment
stage 1: the objectives of conjoint analysis
defining the total utility of the object
specifying the determinant factors
stage 2: the design of a conjoint analysis
selecting a conjoint analysis methodology
designing profiles: selecting and defining factors and
levels
specifying the basic model form
data collection
stage 3: assumptions of conjoint analysis
stage 4: estimating the conjoint model and assessing overall
fit
selecting an estimation technique
estimated part-worths
evaluating model goodness-of-fit
stage 5: interpreting the results
examining the estimated part-worths
assessing the relative importance of attributes
stage 6: validation of the conjoint results
managerial applications of conjoint analysis
segmentation
profitability analysis
conjoint simulators
alternative conjoint methodologies
adaptive/self-explicated conjoint: conjoint with
a large number of factors
choice-based conjoint: adding another touch of realism
overview of the three conjoint methodologies
an illustration of conjoint analysis
stage 1: objectives of the conjoint analysis
stage 2: design of the conjoint analysis
stage 3: assumptions in conjoint analysis
stage 4: estimating the conjoint model and assessing overall
model fit
stage 5: interpreting the results
stage 6: validation of the results
a managerial application: use of a choice simulator
summary 327 . questions 330 . suggested readings
references
chapter 7 multiple discriminant analysis and logistic
regression
what are discriminant analysis and logistic regression?
discriminant analysis
logistic regression
analogy with regression and manova
hypothetical example of discriminant analysis
a two-group discriminant analysis: purchasers versus
nonpurchasers
a geometric representation of the two-group discriminant
function
a three-group example of discriminant analysis: switching
intentions
the decision process for discriminant analysis
stage 1: objectives of discriminant analysis
stage 2: research design for discriminant analysis
selecting dependent and independent variables
sample size
division of the sample
stage 3: assumptions of discriminant analysis
impacts on estimation and classification
impacts on interpretation
stage 4: estimation of the discriminant model and assessing
overall fit
selecting an estimation method
statistical significance
assessing overall model fit
casewise diagnostics
stage 5: interpretation of the results
discriminant weights
discriminant loadings
partial f values
interpretation of two or more functions
which interpretive method to use?
stage 6: validation of the results
validation procedures
profiling group differences
a two-group illustrative example
stage 1: objectives of discriminant analysis
stage 2: research design for discriminant analysis
stage 3: assumptions of discriminant analysis
stage 4: estimation of the discriminant model and assessing
overall fit
stage 5: interpretation of the results
stage 6: validation of the results
a managerial overview
a three-group illustrative example
stage 1: objectives of discriminant analysis
stage 2: research design for discriminant analysis
stage 3: assumptions of discriminant analysis
stage 4: estimation of the discriminant model and assessing
overall fit
stage 5: interpretation of three-group discriminant analysis
results
stage 6: validation of the discriminant results
a managerial overview
logistic regression: regression with a binary dependent
variable
representation of the binary dependent variable
sample size
estimating the logistic regression model
assessing the goodness-of-fit of the estimation model
testing for significance of the coefficients
interpreting the coefficients
calculating probabilities for a specific value of the independent
variable
overview of interpreting coefficients
summary
an illustrative example of logistic regression
stages 1, 2, and 3: research objectives, research design, and
statistical assumptions
stage 4: estimation of the logistic regression model and
assessing overall fit
stage 5: interpretation of the results
stage 6: validation of the results
a managerial overview
summary 434 . questions 437 . suggested readings
references
chapter 8 anova and manova
manova: extending univariate methods for assessing group
differences
multivariate procedures for assessing group differences
a hypothetical illustration of manova
analysis design
differences from discriminant analysis
forming the variate and assessing differences
a decision process for manova
stage 1: objectives of manova
when should we use manova?
types of multivariate questions suitable for manova
selecting the dependent measures
stage 2: issues in the research design of manova
sample size requirements—overall and by group
factorial designs—two or more treatments
using covariates—ancova and mancova
manova counterparts of other anova designs
a special case of manova: repeated measures
stage 3: assumptions of anova and manova
independence
equality of variance–covariance matrices
normality
linearity and multicollinearity among the dependent
variables
sensitivity to outliers
stage 4: estimation of the manova model and assessing overall
fit
estimation with the general linear model
criteria for significance testing
statistical power of the multivariate tests
stage 5: interpretation of the manova results
evaluating covariates
assessing effects on the dependent variate
identifying differences between individual groups
assessing significance for individual dependent variables
stage 6: validation of the results
summary
illustration of a manova analysis
example 1: difference between two independent groups
stage 1: objectives of the analysis
stage 2: research design of the manova
stage 3: assumptions in manova
stage 4: estimation of the manova model and assessing the overall
fit
stage 5: interpretation of the results
example 2: difference between k independent groups
stage 1: objectives of the manova
stage 2: research design of manova
stage 3: assumptions in manova
stage 4: estimation of the manova model and assessing overall
fit
stage 5: interpretation of the results
example 3: a factorial design for manova with two independent
variables
stage 1: objectives of the manova
stage 2: research design of the manova
stage 3: assumptions in manova
stage 4: estimation of the manova model and assessing overall
fit
stage 5: interpretation of the results
summary
a managerial overview of the results
summary 498 . questions 500 . suggested readings
references
section iii analysis using interdependence techniques
chapter 9 grouping data with cluster analysis
what is cluster analysis?
cluster analysis as a multivariate technique
conceptual development with cluster analysis
necessity of conceptual support in cluster analysis
how does cluster analysis work?
a simple example
objective versus subjective considerations
cluster analysis decision process
stage 1: objectives of cluster analysis
stage 2: research design in cluster analysis
stage 3: assumptions in cluster analysis
stage 4: deriving clusters and assessing overall fit
stage 5: interpretation of the clusters
stage 6: validation and profiling of the clusters
an illustrative example
stage 1: objectives of the cluster analysis
stage 2: research design of the cluster analysis
stage 3: assumptions in cluster analysis
employing hierarchical and nonhierarchical methods
step 1: hierarchical cluster analysis (stage 4)
step 2: nonhierarchical cluster analysis (stages 4, 5, and
6)
summary 561 . questions 563 . suggested readings
references
chapter 10 mds and correspondence analysis
what is multidimensional scaling?
comparing objects
dimensions: the basis for comparison
a simplified look at how mds works
gathering similarity judgments
creating a perceptual map
interpreting the axes
comparing mds to other interdependence techniques
individual as the unit of analysis
lack of a variate
a decision framework for perceptual mapping
stage 1: objectives of mds
key decisions in setting objectives
stage 2: research design of mds
selection of either a decompositional (attribute-free)
or compositional (attribute-based) approach
objects: their number and selection
nonmetric versus metric methods
collection of similarity or preference data
stage 3: assumptions of mds analysis
stage 4: deriving the mds solution and assessing overall
fit
determining an object’s position in the perceptual map
selecting the dimensionality of the perceptual map
incorporating preferences into mds
stage 5: interpreting the mds results
identifying the dimensions
stage 6: validating the mds results
issues in validation
approaches to validation
overview of multidimensional scaling
correspondence analysis
distinguishing characteristics
differences from other multivariate techniques
a simple example of ca
a decision framework for correspondence analysis
stage 1: objectives of ca
stage 2: research design of ca
stage 3: assumptions in ca
stage 4: deriving ca results and assessing overall fit
stage 5: interpretation of the results
stage 6: validation of the results
overview of correspondence analysis
illustrations of mds and correspondence analysis
stage 1: objectives of perceptual mapping
identifying objects for inclusion
basing the analysis on similarity or preference data
using a disaggregate or aggregate analysis
stage 2: research design of the perceptual mapping study
selecting decompositional or compositional methods
selecting firms for analysis
nonmetric versus metric methods
collecting data for mds
collecting data for correspondence analysis
stage 3: assumptions in perceptual mapping
multidimensional scaling: stages 4 and 5
stage 4: deriving mds results and assessing overall fit
stage 5: interpretation of the results
overview of the decompositional results
correspondence analysis: stages 4 and 5
stage 4: estimating a correspondence analysis
stage 5: interpreting ca results
overview of ca
stage 6: validation of the results
a managerial overview of mds results
summary 623 . questions 625 . suggested readings
references
section iv structural equations modeling
chapter 11 sem: an introduction
what is structural equation modeling?
estimation of multiple interrelated dependence
relationships
incorporating latent variables not measured directly
defining a model
sem and other multivariate techniques
similarity to dependence techniques
similarity to interdependence techniques
the emergence of sem
the role of theory in structural equation modeling
specifying relationships
establishing causation
developing a modeling strategy
a simple example of sem
the research question
setting up the structural equation model for path analysis
the basics of sem estimation and assessment
six stages in structural equation modeling
stage 1: defining individual constructs
operationalizing the construct
pretesting
stage 2: developing and specifying the measurement model
sem notation
creating the measurement model
stage 3: designing a study to produce empirical results
issues in research design
issues in model estimation
stage 4: assessing measurement model validity
the basics of goodness-of-fit
absolute fit indices
incremental fit indices
parsimony fit indices
problems associated with using fit indices
unacceptable model specification to achieve fit
guidelines for establishing acceptable and unacceptable fit
stage 5: specifying the structural model
stage 6: assessing the structural model validity
structural model gof
competitive fit
comparison to the measurement model
testing structural relationships
summary 678 . questions 680 . suggested readings
appendix 11a: estimating relationships using path analysis
appendix 11b: sem abbreviations
appendix 11c: detail on selected gof indices
references
chapter 12 applications of sem
part 1: confirmatory factor analysis
cfa and exploratory factor analysis
a simple example of cfa and sem
a visual diagram
sem stages for testing measurement theory validation with
cfa
stage 1: defining individual constructs
stage 2: developing the overall measurement model
unidimensionality
congeneric measurement model
items per construct
reflective versus formative constructs
stage 3: designing a study to produce empirical results
measurement scales in cfa
sem and sampling
specifying the model
issues in identification
avoiding identification problems
problems in estimation
stage 4: assessing measurement model validity
assessing fit
path estimates
construct validity
model diagnostics
summary example
cfa illustration
stage 1: defining individual constructs
stage 2: developing the overall measurement model
stage 3: designing a study to produce empirical results
stage 4: assessing measurement model validity
hbat cfa summary
part 2: what is a structural model?
a simple example of a structural model
an overview of theory testing with sem
stages in testing structural theory
one-step versus two-step approaches
stage 5: specifying the structural model
unit of analysis
model specification using a path diagram
designing the study
stage 6: assessing the structural model validity
understanding structural model fit from cfa fit
examine the model diagnostics
sem illustration
stage 5: specifying the structural model
stage 6: assessing the structural model validity
part 3: extensions and applications of sem
reflective versus formative measures
reflective versus formative measurement theory
operationalizing a formative construct
distinguishing reflective from formative constructs
which to use—reflective or formative?
higher-order factor analysis
empirical concerns
theoretical concerns
using second-order measurement theories
when to use higher-order factor analysis
multiple groups analysis
measurement model comparisons
structural model comparisons
measurement bias
model specification
model interpretation
relationship types: mediation and moderation
mediation
moderation
longitudinal data
additional covariance sources: timing
using error covariances to represent added covariance
partial least squares
characteristics of pls
advantages and disadvantages of pls
choosing pls versus sem
summary 778 . questions 781 . suggested readings
references
index
版权页：插图：Missing data are termed missing at random （MAR） if the missing values of Y depend on X, but not on Y In other words, the observed Y values represent a random sample of the actual Y values for each value of X, but the observed data for Ydo not necessarily represent a truly ran- dom sample of all Y values. Even though the missing data process is random in the sample, its values are not generalizable to the population. Most often, the data are missing randomly within subgroups, but differ in levels between subgroups. The researcher must determine the factors determining the subgroups and the varying levels between groups.For example, assume that we know the gender of respondents （the X variable） and are asking about household income （the Y variable）. We find that the missing data are random for both males and females but occur at a much higher frequency for males than females. Even though the missing data process is operating in a random manner within the gender variable, any remedy applied to the missing data will still reflect the missing data process because gen- der affects the ultimate distribution of the household income values.A higher level of randomness is termed missing completely at random （MCAR）. In these instances the observed values of Y are truly a random sample of all Y values, with no underly- ing process that lends bias to the observed data. In simple terms, the cases with missing data are indistinguishable from cases with complete data.From our earlier example, this situation would be shown by the fact that the missing data for household income were randomly missing in equal proportions for both males and females. In this missing data process, any of the remedies can be applied without making allowances for the impact of any other variable or missing data process.Diagnostic Tests for Levels of Randomness. As previously noted, the researcher must ascer-tain whether the missing data process occurs in a completely random manner. When the data set issmall, the researcher may be able to visually see such patterns or perform a set of simple calculations（such as in our simple example at the beginning of the chapter）. However, as sample size and the num-ber of variables increases, so does the need for empirical diagnostic tests. Some statistical programsadd techniques specifically designed for missing data analysis （e.g., Missing Value Analysis in SPSS）,which generally include one or both diagnostic tests.The first diagnostic assesses the missing data process of a single variable Y by forming twogroups: observations with missing data for Y and those with valid values of Y. Statistical testsare then performed to determine whether significant differences exist between the two groupson other variables of interest. Significant differences indicate the possibility of a nonrandommissing data process.Let us use our earlier example of household income and gender. We would first form twogroups of respondents, those with missing data on the household income question and thosewho answered the question. We would then compare the percentages of gender for each group.If one gender （e.g., males） was found in greater proportion in the missing data group, we wouldsuspect a nonrandom missing data process. If the variable being compared is metric （e.g., anattitude or perception） instead of categorical （gender）, then t-tests are performed to determinethe statistical significance of the difference in the variable's mean between the two groups. Theresearcher should examine a number of variables to see whether any consistent patternemerges. Remember that some differences will occur by chance, but either a large number or asystematic pattern of differences may indicate an underlying nonrandom pattern.A second approach is an overall test of randomness that determines whether the missing datacan be classified as MCAR.
《多元数据分析(英文版)(第7版)》特色：以循序渐进方式（流水线方式）组织内容：在内容组织上，各章集中概述一个论题，每章均从基础开始并讨论应用。后面各章逐步深入。扩展各种方法应用：对"经验法则"给出解释，包括像样本容量这类重要问题。重新组织结构方程建模这一重要内容，包括结构方程建模概述、验证-性因素分析、估计和检验结构模型的相关问题，以及验证性因素分析和结构方程建模的一些高级主题，如检验更高阶因子模型、群组模型、调节变量与中间变量。
无
这本是应用类多元里一本写得比较好的，很细致，结构我比较喜欢，但是做教材不行，毕竟是英文版。我觉得适合研究生和专业人士看，本科英语好的也可尝试。
书本身质量很好，摸起来也很舒服。
书很好，内容也给力！
蛮好的书哦 原版的原汁原味些 希望能读懂啊
作为一名市场营销的研究生，能够灵活地分析数据，得出决策判断的依据，很有必要学习一下！
适合初学者自学，前提是英语基础不错，或愿意尝试
这本书是国际版，章节安排上与美国本土版本有所不同，但内容基本一致。封面用的是本土版的封面，还是显得出版社不太诚实。这本书基本没有公式，实际上是想用文字语言把多元统计的原理说清楚。网上有配套材料，包括SPSS格式的数据和各章的SPSS SYNTAX文件。如果你的英语理解能力足够好，并且有SPSS和基础统计学的基础，这本书能用作自学，像有些网友称之为多元数据分析的“圣经”也不过分。但是，如果你缺乏这些基础，自学就比较难了。其实这样的书真适合有国外教育背景的统计学家或社会科学家来翻译。希望出版社能在这方面做些努力。
挺好的一本书，就是全英文，理解有点难度，头脑一热买的
多元数据分析工人的bible
Nice book and very useful.
《多元数据分析(英文版.第7版)》是一本面向应用的经典多元数据分析教材，自1979年出版第1版至今，深受读者好评。《多元数据分析(英文版.第7版)》循序渐进地介绍了各种多元统计分析方法，并通过丰富的实例演示了这些方法的应用。书中不仅涵盖多元数据分析的基本方法，而且还介绍了一些新方法，如结构方程建模和偏最小二乘法等。... 阅读更多