统计分析入门与应用：SPSS+SmartPLS（第2版）"
本书是一本全面介绍SmartPLS 3的界面、操作和PLS-SEM结构方程模型的实用书，深入探讨**量表发展、中介和调节变量的应用、反映性和形成性指标的发展和模型的指定。书中新增SEM的演进、形成性的评估、中介因子的5种形态、调节效应的9种形态、测量恒等性、多群组分析呈现的范例、被中介的调节（中介式调节）和被调节的中介（调节式中介）等内容。
本书可作为统计分析和多变量分析的教科书，也是Hair、Black、Babin和Anderson所撰写的《多变量分析》（Multivariate data analysis）的**辅助参考书，更是 Hair、Hult、Ringle和Sarstedt 所撰写的A Primer on Partial Least Squares Structural Equation Modeling（PLS-SEM）的辅助参考书籍。
导 读
本书以统计分析（多变量分析）为主轴，整合了理论的介绍、量化的研究、量表的发展、
传统的统计分析、卡方检验、平均数比较、因子分析、回归分析、判别分析和逻辑回归、单因
素方差分析、多变量方差分析、典型相关分析、信度和效度分析、联合分析多元标度和聚类分
析），第二代统计技术——结构方程模型（SEM），全新改版SmartPLS 的操作和呈现。新增
了SEM的演进、PLS-SEM结构方程模型的学习范例、反映性和形成性指标与模型、二阶和高
阶因果关系、SEM 结构方程模型实例、中介和调节变量的应用。新增了形成性的评估、中介
因子的5种形态、调节效应的9种形态、测量恒等性、多组分析（MGA）呈现的范例、被中
介的调节（中介式调节）、被调节的中介（调节式中介）和论文结构与研究范例等内容。因此，
本书十分适合用于统计分析和多变量分析的课程，也希望有更多学校能够采用，让本书成为一
本协助更多人的有用的教科书。
作者在训练课程和演讲中与数百位研究人员（研究生、讲师、教授、研究机构人员）交
换意见，交流中发现众多研究人员所遇到的一些问题十分相似，我们整理并提供建议解决方式
（Q&A）如下：
.1. 在研究方法上，PLS-SEM的最新研究要求是什么？
答：
在研究方法上，长久以来，PLS-SEM的两大问题是：①缺乏一致性结果；②缺乏模型
适配指标（Model Fit）（Henseler et al.，2014）。SmartPLS 3已经提供初步解决方式，在缺
乏一致性问题上，SmartPLS 3 提供了PLSc功能（Consistent PLS Algorithm + Consistent PLS
Bootstrapping）（Dijkstra and Henseler，2015），可以提供一致性的结果，但只能用在所有构
面是反映性的情况；对于缺乏模型适配指标这个问题，SmartPLS 3 提供了SRMR模型适配指
标（Henseler et al.，2014），但只评估研究模型的适配。
请参考：
－Henseler，J.，Dijkstra，T. K.，Sarstedt，M.，Ringle，C. M.，Diamantopoulos，A.，
Straub，D. W.，Ketchen，D. J.，Hair，J. F. ，Hult，G. T. M.，and Calantone，R. J.
2014.“Common Beliefs and Reality about Partial Least Squares: Comments on R.nkk. &
Evermann（2013），” Organizational Research Methods（17:2），pp. 182-209.
－Dijkstra，T. K.，and Henseler，J. 2015. “Consistent Partial Least Squares Path
Modeling，” MIS Quarterly（39:2），pp. 297-316.
.2. PLS-SEM（SmartPLS 3）的最新功能有哪些?
答：
最新功能如下：
●一致性的PLS、PLSc（PLS Consistence）
●形成性调节构面的正确计算
●重要性——表现力映射分析（IPMA）
●多组分析（MGA：Multigroup Analysis）
●异质性（Heterogeneity） FIMIX-PLS分析
●异质性（Heterogeneity） PLS-POS分析
●PLS验证四价分析（CTA-PLS：Confirmatory Tetrad Analysis PLS）
.3. 调节构面是形成性时，交互作用项可以用自变量和调节变量交叉相乘吗?
答：
不可以，SmartPLS 3提供“2 stages交互作用项”功能，以提供调节构面是形成性时的正
确计算。
.4. 如何验证一个构面是不是形成性？
答：
SmartPLS 3 提供验证性四价分析，以验证一个构面是否为形成性。
请参考：
－Gudergan，S.，Ringle，C.M.，Wende，S.，and Will，A. 2008. “Confirmatory TetradAnalysis in PLS Path Modeling，”Journal of Business Research（61 :12），pp. 1238-1249.
.5. 管理的论文一定要有理论作为基础吗?
答：
探索性的研究不一定要有理论作为基础，因为尚在探索现象阶段。实证研究就要求有足够
的理论基础，因为在管理方面，常以理论为依据，用来说明和解释研究的现象。
请参考：
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：
峰，本书第一章关于理论的部分。
.6. 量表可以自行发展吗?
答：
当然可以，只是发展量表有一定的要求和程序，较为困难，一般的研究都会借用成熟的量表。
请参考：
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：
峰，本书第三章量表的发展。
－Shiau，W.-L.，Hsu，P.-Y.，and Wang，J.-Z. 2009. “Development of measures to assess the ERPadoption of SMEs，” Journal of Enterprise Information Management（22:1/2）， pp. 99-118.
.7. 一般论文的信效度有哪些要求?
答：
量表信度部分，主要检验个别项目的信度，以多元相关平方（Squared Multiple Correlations，
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 4
2023/9/19 21:37:04
. 导 读
V
SMC）值作为观察标准值。理想的SMC值需大于0.5，表示测量指标具有良好的信度。潜在变
量组成信度（Composite Reliability，CR），是指构面内部变量的一致性，一般而言，其值需大于0.7（Hair et al., 2010）。研究中的潜在变量的组成信度值皆大于0.9，代表构面具有良好的内部一致性。
在收敛效度方面，检验因素负荷量、各个构面的组成信度和平均方差萃取量（Hair et al., 2010；
Shiau and Luo, 2013）。因素负荷量需大于 0.7，各测量构面的组成信度的值需大于 0.7（CR>0.7）。
当所有构面平均方差萃取量的值均大于建议值阈值0.5（Hair et al., 2010），则具有收敛效度。区
别效度主要是检验测量变量对于不同构面间的区别程度。各构面间平均方差萃取量的平方根值
均需大于测量不同构面间之相关系数（Hair et al., 2010; Shiau and Luo, 2013）。
请参考:
－Hair，J.F.，Black，W.C.，Babin，B.J.，and Anderson，R.E. 2010. Multivariate data
analysis: A global perspective（7th ed.），Upper Saddle River，NJ: Pearson Prentice
Hall.
－Shiau，W.-L.，and Luo，M.M. 2013.“Continuance intention of blog users: the impact
of perceived enjoyment，habit，user involvement and blogging time，”Behaviour &
Information Technology（BIT）（32:6），pp.570-583.
－Shiau，W.-L.，and Chau，P.Y.K. 2015，Understanding behavioral intention to use a
cloud computing classroom: A multiple model comparison approach，Information &
Management，Available online 6 November 2015，ISSN 0378-7206，http://dx.doi.
org/10.1016/j.im.2015.10.004.
.8. 为何要使用PLS？
答：
相较于LISREL和AMOS的SEM，PLS方法对于测量标度（Measurement Scales）、样本
数大小（Sample Size）和残差分布（Residual Distributions ）的要求较低。
Ringle et al.整理使用PLS 方法的理由如表0-1所示。
表0-1 使用PLS方法的理由
资料来源 : Ringle，C.M.，Sarstedt，M.，and Straub，D.W.. 2012. “Editor’s Comments: A Critical Look at
the Use of PLS-SEM in MIS Quarterly，” MIS Quarterly（36: 1），pp. iii-xiv.
请参考:
－Ringle，C. M.，Sarstedt，M.，and Straub，D.W. 2012.“Editor’s Comments: A CriticalLook at the Use of PLS-SEM in MIS Quarterly，”MIS Quarterly（36: 1），pp. iii-xiv.
－Shiau，W.-L.，and Luo，M.M. 2012. “Factors Affecting Online Group Buying Intention
and Satisfaction: A Social Exchange Theory Perspective，”Computers in Human Behavior
（28:6），pp.2431-2444.
－Hair，J.F.，Sarstedt，M.，Ringle，C.M.，and Mena，J.A. 2012. “An Assessmentof the Use of Partial Least Squares Structural Equation Modeling in MarketingResearch，”Journal of the Academy of Marketing Science（40:3），pp. 414-433.
－Shiau，W.-L.，and Chau，P.Y.K. 2015.“Understanding behavioral intention to use acloud computing classroom: A multiple model comparison approach，”Information &
Management，Available online 6 November 2015，ISSN 0378-7206，http://dx.doi.
org/10.1016/j.im.2015.10.004.
.9. CB-SEM和 PLS-SEM有何不同？
答：
以变量的协方差Covariance结构进行分析，称为Covariance-Base SEM（CB-SEM），常
用的软件工具有LISREL、EQS、AMOS。
以变量的主成分结构进行分析使用做最小平方法（Partial Lease Square，PLS），称为
PLS-SEM，常用的软件工具有SmartPLS、PLS-Graph、VisualPLS。
请参考:
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：碁峰，本书第15章。
－Hair，J.F.，Sarstedt，M.，Ringle，C.M.，and Mena，J.A. 2012. “An Assessmentof the Use of Partial Least Squares Structural Equation Modeling in MarketingResearch，”Journal of the Academy of Marketing Science（40:3），pp. 414-433.
－Shiau，W.-L.，and Chau，P.Y.K. 2015.“Understanding behavioral intention to use acloud computing classroom: A multiple model comparison approach，” Information &
Management，Available online 6 November 2015，ISSN 0378-7206，http://dx.doi.
org/10.1016/j.im.2015.10.004.
.10. CB-SEM和 PLS-SEM的使用适用于哪些场景？最小样本需求如何？
答：
CB-SEM 技术强调全部的适配，主要是在检测理论的适用性，适合进行理论模型的检测（验
证性）。CB-SEM（LISREL、EQS、AMOS）所需要的样本最小值介于100～150之间，最好
有题项总数的10倍。
PLS-SEM ，PLS的部分，它的设计主要是在解释变异（检测因果关系是否具有显著的关系），
适合进行理论模型的建构（探索性），也用来验证所探讨推论的因果关系。PLS对于样本的需
求为：样本数一定要大于所提出的题项总数，最好有题项总数的10倍。
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 6
2023/9/19 21:37:05
请参考:
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：峰，本书第15章。
－Gefen，D.，Rigdon，E. E.，and Straub，D. 2011. “An Update and Extension to SEM
Guidelines for Administrative and Social Science Research，”MIS Quarterly（35 :2），
pp. iii-xiv.
.11. 一般研究常用的模型比较有哪些？
答：
在相同的模型中，一般常用的为巢状模型（Nested Model）比较。
在不同的模型中，常用的为成对巢状F检验（Pairwise Nested F-tests） 。
请参考:
－Shiau，W.-L.，and Chau，P.Y.K. 2012.“Understanding blog continuance: a model comparison
approach，” Industrial Management & Data Systems（112: 4），pp. 663-682.
.12. Reflective（反映性）和Formative（形成性）的观察变量有何不同?
Reflective（反映性）和Formative（形成性）的模型有何不同?
答：
测量模型是观察变量对于潜在构面的关联性，主要可以分成两种关系：
－反映性的观察变量：所观察的变量可以直接反映到潜在变量上，是属于单向的关联性。
－形成性的观察变量：它是探讨形成潜在构面的原因。
反映性模型的题项会呈现构面，题项改变不会造成构面的改变，而构面改变会造成题项改
变。题项是有可换性的，题项有相同或类似的内容，应用在同一个主题之下，删除题项不会改
变构面的概念 。
形成性模型的题项定义了构面的特征，如果题项改变，构面也会跟着改变。题项不具有互
换性，题项没有相同或是类似的内容，删除题项有可能会改变构面的概念。
请参考:
－Ringle，C. M.，Wende，S.，and Will，A. 2005. SmartPLS2.0（M3），Hamburg:
University of Hamburg.（http://www.smartpls.de）
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：峰，本书第19章。
.13. 一般研究中，二阶的模型有哪些？
答：
二阶（Second order）的反映性模型与形成性模型是属于阶层式潜在变量模型（Hierarchical
latent variable Model） 最简单的模型。二阶的反映性与形成性模型与一阶的反映性与形成性模
型结合，会形成四种模型：分别是模型一，反映性-反映性；模型二，反映性-形成性；模型三，
形成性-反映性；模型四，形成性-形成性。
请参考:
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：峰，本书第19章。
.14. 投稿时，常被要求提供CMV，什么是CMV呢?
答：
CMV的全名是共同方法变异（Common Method Variance），又称同源方差，是指收集数
据时，同一个方法（来源）可能导致的偏差（Bias），又称为共同方法偏差（Common Method
Bias）。目前常用 Harman’s Single Factor Test 和Marker Variable Test来检测CMV。
请参考:
－Shiau，W.-L.，and Luo，M.M. 2012. “Factors Affecting Online Group Buying Intentionand Satisfaction: A Social Exchange Theory Perspective，”Computers in Human Behavior
（28:6），pp.2431-2444.
－Avus C.Y. Hou and Wen-Lung Shiau（2019）“Understanding Facebook to Instagrammigration: a push-pull migration model perspective”，Information Technology & People，
Forthcoming DOI: https://doi.org/10.1108/ITP-06-2017-0198
－Podsakoff，P. M.，MacKenzie，S. B.，Lee，J.-Y.，and Podsakoff，N. P. 2003.
“Common Method Biases in Behavioral Research: A Critical Review of the Literature and
Recommended Remedies，”Journal of Applied Psychology（88:5），pp. 879-903.
－Malhotra，N. K.，Kim，S. S.，and Patil，A. 2006. “Common Method Variance in ISResearch: A Comparison of Alternative approaches and a Reanalysis of Past Research，”
Management Science（52:12），pp. 1865-1883.
－（For PLS-SEM）
Ronkko，M.，Ylitalo，Y.，2011. PLS marker variable approach to diagnosing andcontrolling for method variance. In: Proceedings of the International Conference onInformation Systems（ICIS 2011）. Paper 8.
Chin，Wynne; Thatcher，Jason Bennett; and Wright，Ryan T..2012.“Assessing CommonMethod Bias: Problems with the ULMC Technique，”MIS Quarterly，（36: 3）pp.1003-1019.
－（for CB-SEM）
Serrano，C.，and Karahanna，E. 2016. “The compensatory interaction between usercapabilities and technology capabilities in influencing task performance: an empiricalassessment in telemedicine consultations”MIS Quarterly（40:3），pp. 597-622.
.15. 什么是无响应偏差（None Response Bias） ?
答：
在收集数据时，没有响应的数据会产生偏误，称为无响应偏差。一般处理的方式是，将回
收的数据分成前期和后期的数据作检验，利用t或卡方检验前后期响应无显著差异，以显示无
响应偏差对本研究的影响并不严重。
请参考:
－Shiau，W.-L.，and Luo，M.M. 2012. “Factors Affecting Online Group Buying Intentionand Satisfaction: A Social Exchange Theory Perspective，” Computers in Human Behavior
（28:6），pp.2431-2444.
.16.收集的数据呈现非正态分布（Non-normal Distribution），或违反基本假设，例如，
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 8
2023/9/19 21:37:05
方差齐性和独立性，该如何处理？
答：
1）在一般情况下，多变量分析的书都会建议进行数据的转换，将非正态分布的数据转换
成正态的分布。
2）若两组的方差不一样，建议使用Welch’s t-test。
3）若是使用ANOVA分析，方差齐性有问题时，建议使用Games-Howell事后检验。
4）当自变量不是正态分布时，也可以将区间数据转换成顺序数据，使用Whitney-Manu-
Wilcoxon检验。
5）当自变量不是正态分布时，将区间数据转换成顺序数据。若是多于两组要检验，建议
使用Kruskal-Wallis替代ANOVA检验。
请参考:
－López，X.，Valenzuela，J.，Nussbaum，M.，and Tsai，C.-C. 2015. “Some
recommendations for the reporting of quantitative studies，”Computers & Educat ion（91），
pp.106-110.
－Hair，J.F.，Black，W.C.，Babin，B.J.，and Anderson，R.E. 2010. Multivariate data
analysis: A global perspective（7th ed.）. Upper Saddle River，NJ: Pearson Prentice Hall.
.17.投稿文章后，收到审查意见“缺乏贡献或贡献不足”时，该如何处理？
答：
理论是概念（concept）和它们之间关系的叙述，用来说明现象发生的原因和发生的过程，
因此，理论上的贡献是增加我们对于概念和它们之间关系的了解（知识）。一篇好的管理（期
刊）文章要有对知识的重要贡献。典型的管理（期刊）文章一般都会先讨论理论上的贡献，接
着讨论研究和实务上的意涵（Implication for Research and Practice）。实务上的意涵是经过确
认而需要说明实务上的问题；研究上的意涵是经过确认而说明未来需要调查的现象。
Ladik and Stewart（2008）对于一篇有所创新（Innovation）的文章，给予不同的贡献程度
评分（1最少，8最多），按从小到大排列如下：
1）直接复制（先前的研究）（Straight Replication）
2）复制和延伸（Replication and Extension）
3）延伸新的理论/将方法运用到新领域（Extension of a New Theory/Method in a New
Area）
4）整合性观点（例如汇总分析） [Integrative Review（e.g.，meta-analysis）]
5）发展新理论，以解释旧现象（Develop a New Theory to Explain an Old Phenomenon -Compete
One Theory Against Another - Classic Theory Testing）
6）确认新的现象（Identification of a New Phenomenon）
7）发展大的融合，也就是整合（Develop a Grand Synthesis - integration）
8）发展新理论以预测新的现象（Develop a New Theory that Predicts a New Phenomenon）
大部分的研究贡献度都落在评分2～5，贡献度评分6～8的研究相对较少，也较不容易
完成和发表。
当研究者完成一篇论文，若是因未能写出贡献而被拒绝刊登是很可惜的。研究者不能寄望
审查者会自己找出贡献，而是要清楚地写出文章的贡献，可能是在理论、方法或文本上的贡献。
理论的贡献（Theoretical Contribution）包含原创性（Originality or Novelty）和效用性（Utility）
两方面。关于理论上的贡献，一般需要呈现出原创性和效用性，理论贡献的原创性或新奇性与
理论的意涵（Theoretical Implication）息息相关。理论的意涵是基于理论的延伸，是理论贡献
中必要且合理的一部分。换句话说，理论的贡献基于存在理论，理论的意涵也就理所当然成为
理论贡献的一部分。理论的意涵又常与科学上的有用相关，科学上的有用是理论贡献的最重要
部分，所以，需要呈现出理论上的效用（Utility）。
具体的贡献可以写在摘要（Abstract）、绪论（Introduction）、讨论（Discuss）、结论
（Conclusion）、理论和实务上的意涵（Implication for Research and Practice）中。例如，在绪
论中探讨A影响B的重要性时，许多文章都会叙述过去的研究而很少探讨A影响B（研究缺
口），未讨论A影响B对哪些人是重要的（Ladik and Stewart 2008），也未讨论A影响B对
于知识（理论上）的贡献。文章在投稿前，务必再次确认已经清楚地写出文章的贡献。
请参考:
－Ladik，D. M.，and Stewart，D. W. 2008. “The contribution continuum，”Journal of theAcademy of Marketing Science（36:2）， pp.157-165.
－.gerfalk，P. J. 2014. “Insufficient theoretical contribution: a conclusive rationale forrejection?”European Journal of Information Systems（23:6），pp.593-599.（Editorial）
.18.投稿时，被要求做Measurement Invariance，什么是Measurement Invariance？
答：
测量不变性（Measurement Invariance）又称为测量恒等性（Measurement Equivalence）。我们
通常使用测量恒等性来确认群组间的差异是来自不同群组潜在变量的内含或意义。换句话说，
无法确立测量恒等性时，群组间的差异可能是来自测量误差，这会使比较群组的结果失效。当
测量恒等性未呈现时，会降低统计检验力，影响估计的精确性，甚至可能会误导结果。总而言之，
做多组分析时，若是未能建立测量恒等性，则所有的结果都可能是有问题的，因此，测量恒等
性在多组分析中，是必要的检测，也是必须通过的测试。PLS-SEM是使用复合模型的测量不变
性（Measurement Invariance of Composit Models，MICOM）程序来评估的。测量恒等性有设定恒
等性（Configurall Invariance）、组成恒等性（Compositional Invariance）与平均值和方差恒等性（Equal
Mean Values and Variances）。
请参考:
－Li-Chun Huang and Wen-Lung Shiau. 2017.“Factors affecting creativity in informationsystem development: Insights from a decomposition and PLS–MGA，”IndustrialManagement & Data Systems，Vol. 117 Iss: 3，pp. 442-458.
.19.投稿时，被要求说明是中介因子5种形态中的哪一种。什么是中介因子的5种形态？
答：
中介因子的5种形态
1）互补的中介[Complementary（Mediation）]
2）竞争的中介[Competitive（Mediation）]
3）完全中介[Indirect-only（Mediation）]
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 10
2023/9/19 21:37:05
4）只有直接影响（无中介）[Direct-only（Non Mediation）]
5）没有影响（无中介）[No-effect（Non Mediation）]
请参考:
－本书第20章
.20.投稿时，被要求说明调节项的估计计算方式（Caculation Method）。什么是调节项
的估计计算方式？
答：
在SmartPLS 3.X中提供3种计算方式来处理交互作用项，分别是：
1）乘积指标法（The Product Indicator Approach）
2）二阶法（默认选项）（The Two-stage Approach）
3）正交法（The Orthogonalizing Approach）
使用变量来产生交叉乘项有3种选项：
1）非标准化（Unstandardized）
2）均值中心化（Mean Centered）
3）标准化（Standardized）（默认选项）
产生交叉乘项有3×3共9种组合Caculation method计算方式
1）非标准化+ 乘积指标
2）非标准化+二阶法
3）非标准化+正交法
4）均值中心化+乘积指标
5）均值中心化+二阶法
6）均值中心化+ 正交法
7）标准+乘积指标
8）标准+二阶法（默认选项）
9）标准+ 正交法
请参考:
－本书第20章
.21.投期刊论文，经常被要求说明PLS-SEM 是否是适当的分析方法，我们提供最新说明
和最新参考文献如下：
在过去的几十年中，基于协方差的结构方程模型是分析观测变量和潜在变量之间复杂关系
的主要方法。PLS-SEM方法近年来在营销管理、组织管理、国际管理、人力资源管理、信息
系统管理、运营管理、管理会计、战略管理、酒店管理、供应链管理和运营管理等诸多领域都
发生了很大的变化，成为重要的多变量分析方法之一。PLS-SEM 放宽了使用 CB-SEM 估计模
型的最大可能性方法所需的正态分布假设要求。此外，PLS-SEM 有使用较小样本估计更复杂
的模型的能力（Hair et al., 2019; Shiau et al., 2019; Khan et al., 2019；萧文龙，2018）。
与CB-SEM相比，PLS-SEM更适合于一些研究，包括：a. 研究目标为理论发展的探索性
研究；b.分析用于预测；c. 结构模型复杂；d. 结构模型包括一个或多个形成性构面； e. 样本
量较小是由于母体较少；f.分布缺乏正态性；g.研究需要潜在的分数进行后续分析。（Gefen
et al., 2011; Hair et al., 2019; Shiau et al., 2019; Khan et al., 2019；萧文龙，2018）。
请参考：
－Gefen，D; Straub，Detmar W.; and Rigdon，Edward E.. 2011. “An Update and Extension
to SEM Guidelines for Admnistrative and Social Science Research，”MIS Quarterly，（35:
2） pp.iii-xiv.
－Khan G. F.，Sarstedt M.，Shiau W，L.，Hair J. F.，Ringle C. M.，Fritze M. P.，（2019）
“Methodological research on partial least squares structural equation modeling（PLS-SEM）:
An analysis based on social network approaches”，Internet Research，Vol. 29 Issue: 3，
pp.407-429.
－Shiau W. L.，Sarstedt M.，Hair J. F.，（2019）“Internet research using partial leastsquares structural equation modeling（PLS-SEM）”，Internet Research，Vol. 29 Issue: 3，
pp.398-406.
－Hair J. F.，Risher J. J.，Sarstedt M.，Ringle C. M.，（2019）“When to use and how to
report the results of PLS-SEM”，European Business Review，Vol. 31 Issue: 1，pp.2-24.
－萧文龙（2018）统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）（第二
版），中国台北：
峰。
.22. 什么是新兴议题? 有哪些方向?
答：
每年都会有新兴议题，可以参考最新的调查（例如Garner Group，Wall Street Journal 等）。
例如，信息系统成功模型可以参考 Petter et al.（2013）新议题和新方向。
以作者为例，电子商务和云计算都有新议题和新方向，也需要了解过去已经建立起来的知
识，例如，电子商务（Shiau and Dwivedi 2013）、知识管理（Shiau 2015）、供应链管理（Shiau，
Dwivedi，and Tsai2015）、企业信息系统（Shiau 2016）、人机互动（Shiau，Yan，and Kuo
2016）、云计算（Shiau and Chau 2016）、社会网络（Shiau，W.-L. and Dwivedi，Y.K.2017）
如表0-2所示。
表0-2 My core knowledge of MIS（digital world）
Supply Chain Management Shiau et
al.（2015）
Social Network Shiau & Dwivedi（2017）
Facebook Shiau et al.（2018）
Human Computer Interaction Shiau
et al.（2016）
Electronic Commerce Shiau &
Dwivedi（2013）
Knowledge Management Shiau（2015）
Enterprise Information System Shiau（2016）
Management Information System
Shiau et al.（2015）
Mobile Information System Shiau（2019）
请参考:
－Petter，S.，DeLone，W.D.，and McLean，E.R.2013. “Information Systems Success: The
Quest for the Independent Variables，”Journal of Management Informat ion Systems（29:4），
pp. 7-62.
－Wen-Lung Shiau and Yogesh K. Dwivedi（2013），“Citation and co-citation analysis toidentify core and emerging knowledge in electronic commerce research，” Scientometrics，
94（3），1317-1337.（SSCI）
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 12
2023/9/19 21:37:05
－Wen-Lung Shiau，Yogesh K. Dwivedi，and Chia-Han Tsai（2015），“Supply chain
management: exploring the intellectual structure”Scientometrics 105（1），215-230.（SSCI，
2014 IF= 2.183，5-year IF is 2.316，Ranks Q1，10 /85 in Information Science & Library
Science.）
－Wen-Lung Shiau，Shu-Yi Chen，and Yu-Cheng Tsai（2015） “Key management
information systems issues: Co-citation analysis of journal articles，”International Journal
of Electronic Commerce Studies，Vol.6，No.1，pp.145-162（EI）.
－Wen-Lung Shiau（2015）“Exploring the intellectual structure of knowledge management:
A co-citation analysis”，International Journal of Advancements in Computing Technology（IJACT）（7:1），pp.9-16.
－Wen-Lung Shiau（2016），“The intellectual core of enterprise information systems: A co-
citation analysis，”Enterprise Information Systems，2016，10（8），815-844.（SCI，2015
IF= 2.269.）
－Wen-Lung Shiau，Chang Ming Yan，and Chen-Chao Kuo（2016），“The Intellectual
Structure of Human Computer Interaction Research，”Journal of Information Science and
Engineering（JISE），32（3），703-730.（SCI）
－Wen-Lung Shiau and Yogesh K. Dwivedi，（2017） “Co-citation and cluster analyses of
extant literature on social networks”，International Journal of Information Management（37（5），390-399.（SSCI）
－Wen-Lung Shiau，Yogesh K. Dwivedi，He-Hong Lai（2018），“Examining the core
knowledge on facebook,”International Journal of Information Management 43（December
2018），Pages 52-63（SSCI，2017 IF= 4.516，5-year Impact Factor: 4.810 ）
－Wen-Lung Shiau，Chang-Ming Yan，Bang-Wen Lin（2019），“Exploration into
the Intellectual Structure of Mobile Information Systems，”International Journal of
Information Management 47（August 2019），Pages 241-251（SSCI，2017 IF=4.516，5-year
Impact Factor: 4.810 ）
.23. SEM 能做什么研究？顶级期刊还接受SEM文章吗？
答：
请参考顶级期刊中，MISQ，ISR 和 JAIS 部分的SEM 多用途范例：
MISQ
Experiment+SEM
Johnston，Warkentin，M.，and Siponen，M. 2015. “An enhanced fear appeal rhetorical
framework: leveraging threats to the human asset through sanctioning rhetoric，”MIS
Quarterly（39:1），pp. 113-134.
Surveys
Schmitz，P. W.，Teng，J. T. C.，and Webb，K. J. 2016. “Capturing the complexity of
malleable it use : adaptive structuration theory for individuals，”Social Science Electronic
Publishing（40:3），pp. 663-686.
Mixed method Qual+Quan（SEM）
Zhang，X.，and Venkatesh，V. 2017. “A nomological network of knowledge management
system use: antecedents and consequences，”MIS Quarterly（41:4），pp. 1275-1306.
Mixed method Qual+Quan（SEM）
Srivastava，Shirish C.; Chandra，Shalini 2018. “Social presence in virtual worldcollaboration: an uncertainty reduction perspective using a mixed methods approach，”MIS
Quarterly（42:3），pp. 779-803.
ISR
A survey experiment（SEM）
Wang，J.，Li，Y.，& Rao，H. R. 2017. “Coping responses in phishing detection: aninvestigation of antecedents and consequences，”Information Systems Research（28:2），
pp. 378-396.
A survey experiment（SEM）
Breward，M.，Hassanein，K.，& Head，M. 2017. “Understanding consumers’ attitudes
toward controversial information，”Information Systems Research（28:4），pp. 760-774.
Qual+Quan（SEM）
Sarker，S.，Ahuja，M.，& Sarker，S. 2018.“Work-Life Conflict of Globally Distributed
Software Development Personnel: An Empirical Investigation Using Border Theory，”
Information Systems Research（29:1），pp. 103-126.
A survey experiment（SEM）
Robert Jr，L. P.，Dennis，A. R.，& Ahuja，M. K. 2018. “Differences are different:
Examining the effects of communication media on the impacts of racial and gender diversityin decision-making teams，”Information Systems Research（29:3），pp. 525-545.
JAIS
Focus group +SEM
Crossler，R. E.，& Posey，C. 2017. “Robbing Peter to Pay Paul: Surrendering Privacyfor Security’s Sake in an Identity Ecosystem，”Journal of the Association for Information
Systems（18:7），pp. 487-515.
Experiment & SEMYou，S.，& Robert，L. 2018. “Emotional attachment，performance，and viabilityin teams collaborating with embodied physical action（EPA） robots，”Journal of theAssociation for Information Systems（19:5），pp. 377-407.
建议引用参考数据如下:
－Dijkstra，T. K.，and Henseler，J. 2015. “Consistent Partial Least Squares PathModeling，”MIS Quarterly（39: 2），pp. 297-316.
－Gefen，D.，Rigdon，E. E.，and Straub，D. 2011.“An Update and Extension to SEMGuidelines for Administrative and Social Science Research，”MIS Quarterly（35 :2），
pp. iii-xiv.
－Gudergan，S.，Ringle，C.M.，Wende，S.，and Will，A. 2008. “Confirmatory Tetrad
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 14
2023/9/19 21:37:05
Analysis in PLS Path Modeling，” Journal of Business Research（61:12）， pp. 1238-
1249.
－Hair，J.F.，Black，W.C.，Babin，B.J.，and Anderson，R.E. 2010. Multivariate data
analysis: A global perspective（7th ed.）. Upper Saddle River，NJ: Pearson Prentice Hall.
－Hair，J.F.，Hult，G.T.M.，Ringle，C.M.，and Sarstedt，M. 2013. A Primer on Partial
Least Squares Structural Equation Modeling. Thousand Oaks: Sage.
－Hair，J.F.，Sarstedt，M.，Ringle，C.M.，and Mena，J.A. 2012. “An Assessment
of the Use of Partial Least Squares Structural Equation Modeling in Marketing
Research，”Journal of the Academy of Marketing Science（40:3），pp. 414-433.
－Henseler，J.，Dijkstra，T. K.，Sarstedt，M.，Ringle，C. M.，Diamantopoulos，
A.，Straub，D.W.，Ketchen，D.J.，Hair，J.F.，H ult，G.T.M.，and Calantone，R.J.
2014. “ Common Beliefs and Reality about Partial Least Squares: Comments on R.nkk. &
Evermann（2013）” Organizational Research Methods（17:2），pp. 182-209.
－Petter，S.，DeLone，W.D.，andMcLean，E.R.2013. “Information Systems Success: The
Quest for the Independent Variables，” Journal of Management Information Systems （29:4），
pp. 7-62.
－Ringle，C.M.，Wende，S.，and Will，A. 2005. SmartPLS2.0（M3），Hamburg:
University of Hamburg.（http://www.smartpls.de）
－Shiau，W.-L.，and Chau，P.Y.K. 2012. “Understanding blog continuance: a model
comparison approach，” Industrial Management & Data Systems（112:4），pp. 663- 682.
－Shiau，W.-L.，Hsu，P.-Y.，and Wang，J.-Z. 2009. “Development of measures to assess
the ERP adoption of SMEs，”Journal of Enterprise Information Management（22:1/2），
pp. 99-118.
－Shiau，W.-L.，and Luo，M.M. 2012. “Factors Affecting Online Group Buying Intention
and Satisfaction: A Social Exchange Theory Perspective，” Computers in Human Behavior（28:6），pp.2431-2444.
－Shiau，W.-L.，and Luo，M.M. 2013. “Continuance intention of blog users: the impact
of perceived enjoyment，habit，user involvement and blogging time，”Behaviour &
Information Technology（32:6），pp.570-583.
－Shiau，W.-L. 2015. “An Evolution，Present，and Future Changes of Cloud Computing
Services，” Journal of Electronic Science and Technology（13:1），pp. 54-59.
－Shiau，W.-L.，and Chau，P.Y.K.（2016） “Understanding behavioral intention to use
a cloud computing classroom: A multiple model-comparison approach”，Information &
Management Vol. 53 Iss: 3，pp 355–365（doi:10.1016/j.im.2015.10.004）（SSCI，2015
IF= 2.163，5-year Impact Factor: 3.175，Ranks Q1，25/144 - Information Science &
Library Science）.
－Wen-Lung Shiau and Yogesh K. Dwivedi（2013），“Citation and co-citation analysis to
identify core and emerging knowledge in electronic commerce research，” Scientometrics，
2013，94（3），1317-1337.（SSCI）
－Wen-Lung Shiau，Yogesh K. Dwivedi，and Chia-Han Tsai（2015），“Supply chainmanagement: exploring the intellectual structure，”Scientometrics 105（1），215-230.（SSCI，
2014 IF= 2.183，5-year IF is 2.316，Ranks Q1，10 /85 in Information Science & Library
Science.）
－Wen-Lung Shiau，Shu-Yi Chen，and Yu-Cheng Tsai（2015）“Key managementinformation systems issues: Co-citation analysis of journal articles”，International Journal
of Electronic Commerce Studies，Vol.6，No.1，pp.145-162（EI）.
－Wen-Lung Shiau（2015）“Exploring the intellectual structure of knowledge management:
A co-citation analysis”，International Journal ofAdvancements in Computing Technology（IJACT）（7:1），pp.9-16.
－Wen-Lung Shiau（2016），“The intellectual core of enterprise information systems: A co-
citation analysis”，Enterprise Information Systems，2016，10（8），815-844.（SCI，
2015 IF= 2.269.）
－Wen-Lung Shiau，Chang Ming Yan，and Chen-Chao Kuo（2016），“The IntellectualStructure of Human Computer Interaction Research”，Journal of Information Science and
Engineering（JISE），32（3），703-730.（SCI）
－Wen-Lung Shiau and Yogesh K. Dwivedi，（2017）“Co-citation and cluster analyses ofextant literature on social networks”，International Journal of Information Management，
37（5），390-399.（SSCI）
－Wen-Lung Shiau，Yogesh K. Dwivedi，He-Hong Lai（2018），“Examining the coreknowledge on facebook”，International Journal of Information Management 43（December
2018），Pages 52-63（SSCI，2017 IF= 4.516，5-year Impact Factor: 4.810 ）
－Wen-Lung Shiau，Chang-Ming Yan，Bang-Wen Lin（2019），“Exploration intothe Intellectual Structure of Mobile Information Systems”，International Journal ofInformation Management 47（August 2019），Pages 241-251（SSCI，2017 IF= 4.516，5-year
Impact Factor: 4.810 ）
－Huang，L.-C. and Shiau，W.-L.（2017），“Factors affecting creativity in informationsystem development: Insights from a decomposition and PLS–MGA”，IndustrialManagement & Data Systems，Vol. 117 Iss: 3，pp. 442 - 458 -（SCI）
－萧文龙.多变量分析最佳入门实用书：SPSS＋LISREL（SEM）[M]. 2版. 中国台北：
峰（硕博士论文的引用次数已经超过1 200次），2009.
－萧文龙.统计分析：SPSS中文版+PLS-SEM（SmartPLS）[M]中国台北：峰，2013.
－萧文龙.统计分析入门与应用SPSS（中文版）+ SmartPLS 3（PLS-SEM）[M].中国台北：
峰，2016.
－萧文龙.统计分析入门与应用：SPSS中文版+SmartPLS 3（PLS-SEM）[M]. 5版.中
国台北：
峰，2018.
－萧文龙 & 陈世智. AMOS结构方程模式最佳入门实用书[M].中国台北：
峰，2018.
统计分析入门与应用：SPSS+SmartPLS（第2版） 5校 文前.indd 16
2023/9/19 21:37:06
目 录
第1章 统计分析简介与数量方法的基础 ··············································1
1.1 统计分析简介 ··············································································1
1.2 理论 ··························································································2
1.3 量表简介 ··················································································12
1.4 抽样 ························································································16
1.5 统计分析的基础统计学 ································································17
1.6 常用的统计分析（多变量分析或称为数量方法） ································29
第2章 SPSS的基本操作 ·······························································34
2.1 SPSS简介 ················································································34
2.2 SPSS软件的菜单介绍 ··································································36
2.3 数据的输入 ···············································································40
2.4 数据分析与输出结果 ···································································45
2.5 实用范例 ··················································································50
第3章 量表的编制、信度和效度 ······················································64
3.1 量表的编制 ···············································································64
3.2 量表的信度和效度 ······································································69
3.3 量表发展实例 ············································································69
3.4 探索性和验证性研究的信度和效度 ·················································70
3.5 探索性因素分析和验证性因素分析的比较 ·········································77
3.6 研究作业 ··················································································78
第4章 检视数据与描述性统计 ·························································79
4.1 检视数据 ··················································································79
4.2 描述性统计分析 ········································································105
第5章 相关分析 ·········································································114
5.1 相关分析 ·················································································114
5.2 积差相关系数 ···········································································115
5.3 相关系数 ··············································································119
5.4 点二系列相关 ···········································································123
5.5 斯皮尔曼等级相关 ·····································································126
5.6 偏相关 ····················································································128
5.7 部分相关 ·················································································132
第6章 卡方检验 ·········································································138
6.1 卡方检验 ·················································································138
6.2 适配度检验 ··············································································138
6.3 独立性检验 ··············································································142
6.4 同质性检验 ··············································································148
第7章 平均值比较(t检验)··························································155
7.1 平均值比较（各种t检验的应用）·················································155
7.2 平均值分析 ··············································································155
7.3 单一样本t检验 ·········································································159
7.4 独立样本t检验 ·········································································163
7.5 成对样本t检验 ·········································································165
第8章 因子分析 ·········································································171
8.1 因子分析 ·················································································171
8.2 因子分析的基本统计假设 ····························································172
8.3 因子分析的检验 ········································································172
8.4 选取因子的数量 ········································································172
8.5 因子的转轴和命名成为构面 ·························································173
8.6 样本的大小和因子分析的验证 ······················································175
8.7 因子分析在研究上的重要应用 ······················································175
8.8 研究范例 ·················································································176
第9章 回归分析 ·········································································186
9.1 回归分析 ·················································································186
9.2 回归分析的基本统计假设 ····························································186
9.3 找出最佳的回归模型 ··································································187
9.4 检验回归模型的统计显著性 ·························································188
9.5 共线性问题 ··············································································189
9.6 验证结果 ·················································································189
9.7 研究范例 ·················································································190
第10章 判别分析与逻辑回归 ························································208
10.1 判别分析 ···············································································208
10.2 逻辑回归 ···············································································217
第11章 单变量方差分析 ·······························································225
11.1 单变量方差分析简介 ·································································225
11.2 单因素方差分析的设计 ······························································225
11.3 方差分析的基本假设条件 ···························································226
11.4 单变量方差分析 ····································································226
11.5 单变量方差分析范例 ·································································229
11.6 单因素方差分析范例 ·································································235
11.7 重复测量 ···············································································240
11.8 单变量协方差分析——控制变量 ··················································245
11.9 单变量协方差分析——前后测设计 ···············································251
第12章 变量方差分析 ·································································257
12.1 多变量方差分析 ·······································································257
12.2 多变量方差分析的基本假设 ························································257
12.3 多变量方差分析和判别分析的比较 ···············································257
12.4 多变量方差分析与单变量方差分析的比较 ······································258
12.5 样本大小的考虑 ·······································································258
12.6 多变量方差的检验 ····································································258
12.7 双因子交互作用下的处理方式 ·····················································259
12.8 多变量方差分析范例：双因子交互作用显著 ···································262
12.9 多变量方差分析范例：双因子交互作用不显著 ································286
第13章 典型相关 ·······································································301
13.1 典型相关 ···············································································301
13.2 典型相关分析的基本假设 ···························································301
13.3 典型函数的估计 ·······································································302
13.4 典型函数的选择 ·······································································302
13.5 冗余指数 ···············································································302
13.6 解释典型变量 ·········································································303
13.7 验证结果 ···············································································303
13.8 典型相关与其他多变量计数的比较和应用 ······································303
13.9 典型相关的范例 ·······································································304
第14章 联合分析、多维标度方法和聚类分析 ····································315
14.1 联合分析 ···············································································315
14.2 多维标度方法 ·········································································319
14.3 聚类分析 ···············································································326
第15章 结构方程模型之偏最小二乘法 ·············································336
15.1 结构方程模型 ·········································································336
15.2 偏最小二乘法 ·········································································342
15.3 结构方程模型 ·········································································344
15.4 偏最小二乘法的结构方程模型 ·····················································346
15.5 协方差形式结构方程模型和方差形式结构方程模型的比较·················349
15.6 当代结构方程模型研究（论文）需要呈现的内容·····························352
第16章 SmartPLS统计分析软件介绍 ·············································353
16.1 SmartPLS统计分析软件的基本介绍··············································353
16.2 将Excel（.xls）转成csv文件（.csv） ···········································354
16.3 基本功能介绍 ·········································································356
16.4 基本功能实际操作 ····································································361
第17章 偏最小二乘结构方程模型的学习范例 ····································377
17.1 一因三果的模型 ·······································································377
17.2 三因一果的模型 ·······································································388
17.3 单一间接路径的模型 ·································································399
17.4 多重间接路径的模型 ·································································409
17.5 多重直接和间接路径的模型 ························································424
第18章 结构方程模型的反映性模型实例 ··········································436
18.1 PLS-SEM结构方程模型的各种标准 ··············································436
18.2 PLS-SEM 研究（论文）需要展示的内容 ········································441
18.3 PLS-SEM实例——量表的设计与问卷的回收 ··································443
18.4 结构方程模型之反映性模型范例 ··················································445
第19章 结构方程模型之形成性模型 ················································459
19.1 反映性与形成性模型的比较 ························································459
19.2 反映性和形成性的模型设定错误 ··················································461
19.3 反映性和形成性模型的判断 ························································462
19.4 反映性和形成性模型的范例 ························································463
19.5 方差形式结构方程模型研究（论文）需要呈现的内容·······················468
19.6 形成性构面测量模型的评估标准 ··················································469
19.7 结构方程模型之形成性模型实例 ··················································471
19.8 阶层式潜在变量模型 ·································································488
第20章 交互作用、中介和调节（干扰） ··········································492
20.1 交互作用 ···············································································492
20.2 中介效果的验证 ·······································································497
20.3 调节（干扰）效果的验证 ···························································514
20.4 写作参考范例 ·········································································554
第21章 SmartPLS 3进阶应用介绍 ·················································561
21.1 一致性的偏最小二乘法 ······························································561
21.2 重要性与绩效的矩阵分析（Importance-performance Matrix Analysis，
IPMA） ··························································································570
21.3 多组分析 ···············································································572
21.4 异质性 ··················································································580
21.5 偏最小二乘验证性四元体分析 ·····················································594
第22章 被中介的调节（中介式调节）和被调节的中介（调节式中介）
分析 ·············································································597
第23章 研究流程、论文结构与在期刊上发表的建议 ···························612
23.1 研究流程 ···············································································612
23.2 论文结构 ···············································································612
23.3 研究在期刊上发表的建议 ···························································615
"本书以统计分析(多变量分析)为主轴，整合了理论的介绍、量化的研究、量表的发展、传统的统计分析、卡方检验、平均值比较、因子分析、回归分析、判别分析和逻辑回归、单因素方差分析、多变量方差分析、典型相关分析、信度和效度分析、联合分析多元标度和聚类分析，第二代统计技术——结构方程模型(SEM)。
本书内容包括SmartPLS 3基本操作、PLS-SEM 结构方程模型的学习范例、反映性和形成性指针与模型的指定、二阶和高阶因果关系、SEM 结构方程模型实例、中介因子的5种形态、调节效果的9种形态、测量恒等性、多组分析呈现的范例、论文结构与研究范例。
本书实用性较强，主要引导学员从学习基础理论知识开始，到独立完成一项研究专题、写好研究生论文以及成功发表论文。本书适用于统计学专业的大学本科、研究生以及统计分析领域的研究人员阅读。"
萧文龙，毕业于美国Polytechnic Univ.硕士，台湾中央大学企管博士，任职浙工大管院特聘教授。擅长SPSS资料分析和结构方程模型，多次登上SmartPLS官网介绍的学者。
