REFERENSI PENTING METODE SEM

SEBELUM 2001

1988 ANDERSON & GERBING Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach

1993 STEIGER Measures of Fit in Structural Equation Modeling: An Introduction

1995 HATCHER Latent-Variable, Structural Equation Modeling with PROC CALIS

1996 BROWN Assessing Specific Mediational Effects in Complex Theoretical Models

1997 BACON Using Amos for Structural Equation Modeling in Market Research

1998 HOX An Introduction to Structural Equation Modeling

1998 MACLEAN & GRAY Structural Equation Modeling in Market Research

1998 PEARL Graphs, Causality and Structural Equation Models

1998 SPIRTES ETAL Using Path Diagrams as a Structural Equation Modeling Tool

2000 OUD & FOLMER A Structural Equation Approach to Spatial Dependence Models

BUKU: Handbook of Structural Equation Modeling

The first comprehensive structural equation modeling (SEM) handbook, this accessible volume presents both the mechanics of SEM and specific SEM strategies and applications. The editor, contributors, and editorial advisory board are leading methodologists who have organized the book to move from simpler material to more statistically complex modeling approaches. 

Sections cover the foundations of SEM; statistical underpinnings, from assumptions to model modifications; steps in implementation, from data preparation through writing the SEM report; and basic and advanced applications, including new and emerging topics in SEM. Each chapter provides conceptually oriented descriptions, fully explicated analyses, and engaging examples that reveal modeling possibilities for use with readers' data. Many of the chapters also include access to data and syntax files at the companion website, allowing readers to try their hands at reproducing the authors' results.

BUKU: Principles and Practice of Structural Equation Modeling, Third Edition

This bestselling text provides a balance between the technical and practical aspects of structural equation modeling (SEM). Using clear and accessible language, Rex B. Kline covers core techniques, potential pitfalls, and applications across the behavioral and social sciences. Some more advanced topics are also covered, including estimation of interactive effects of latent variables and multilevel SEM. The companion Web page offers downloadable syntax, data, and output files for each detailed example for EQS, LISREL, and Mplus, allowing readers to view the results of the same analysis generated by three different computer tools.

 

New to This Edition

*Thoroughly revised and restructured to follow the phases of most SEM analyses.

*Syntax, data, and output files for all detailed research examples are now provided online.

*Chapter on computer tools.

*Exercises with answers, which support self-study.

*Topic boxes on specialized issues, such as dealing with problems in the analysis; the assessment of construct measurement reliability; and more.

*Updated coverage of a more rigorous approach to hypothesis and model testing; the evaluation of measurement invariance; and more.
*”Troublesome” examples have been added to provide a context for discussing how to handle various problems that can crop up in SEM analyses.

SMART LIBRARY: Koleksi 50 + Buku Metodologi Penelitian

Aplikasi Teknik Pengambilan Keputusan dalam Manajemen Rantai Pasok; Teori dan Aplikasi Sistem Pakar dalam Teknologi Manajerial; SEM dengan Lisrel 8.8, Konsep & Tutorial; Riset Pemasaran; Metode Penelitian Kombinasi (Mixed Methods); Analisis Ekonometrika & Statistika dengan Eviews (Ed 2), Riset Bisnis dengan Analisis Jalur SPSS; Pemodelan Sumber Daya Perikanan dan Kelautan; Konsep Dasar Riset Pemasaran & Perilaku Konsumen;

Analisis Sistem Dinamis, Lingkungan Hidup, Sosial, Ekonomi, Manajemen; Konsep dan Aplikasi SEM berbasis Varian dalam Penelitian Bisnis; Metode Riset Kualitatif dalam PR & Marketing Communication; Analisis SWOT Teknik Membedah Kasus Bisnis; An Introduction to DEA, A Tool for Performance Measurement; Statistika untuk Penelitian; Riset Pemasaran dan Konsumen; Analisis Input Output; Marketing Research;

Penelitian Bisnis Paradigma Kuantitatif; Teknik dan Aplikasi Pengambilan Keputusan Kriteria Majemuk; SPSS Statistik Parametrik; Riset Eksperimen dengan Excel 2007 dan Minitab 15; Buku Pintar Minitab 15; Olah Data Skripsi dan Penelitian dengan SPSS 19; Metodologi Penelitian Ekonomi Islam; Analytic Hierarchy Process; Buku Saku SPSS, Analisis Statistik Data; Decision Making with The ANP, Economic, Political, Social & Technological Application with BOCR;

BUKU: Longitudinal Structural Equation Modeling

Featuring actual datasets as illustrative examples, this book reveals numerous ways to apply structural equation modeling (SEM) to any repeated-measures study. Initial chapters lay the groundwork for modeling a longitudinal change process, from measurement, design, and specification issues to model evaluation and interpretation. 

Covering both big-picture ideas and technical "how-to-do-it" details, the author deftly walks through when and how to use longitudinal confirmatory factor analysis, longitudinal panel models (including the multiple-group case), multilevel models, growth curve models, and complex factor models, as well as models for mediation and moderation. 

User-friendly features include equation boxes that clearly explain the elements in every equation, end-of-chapter glossaries, and annotated suggestions for further reading. The companion website (www.guilford.com/little-materials) provides datasets for all of the examples--which include studies of bullying, adolescent students' emotions, and healthy aging--with syntax and output from LISREL, Mplus, and R (lavaan).

BUKU: Structural Equation Modeling With Lisrel, Prelis, and Simplis: Basic Concepts, Applications, and Programming

This book illustrates the ease with which various features of LISREL 8 and PRELIS 2 can be implemented in addressing research questions that lend themselves to SEM. Its purpose is threefold: (a) to present a nonmathmatical introduction to basic concepts associated with SEM, (b) to demonstrate basic applications of SEM using both the DOS and Windows versions of LISREL 8, as well as both the LISREL and SIMPLIS lexicons, and (c) to highlight particular features of the LISREL 8 and PRELIS 2 progams that address important caveats related to SEM analyses.

This book is intended neither as a text on the topic of SEM, nor as a comprehensive review of the many statistical funcitons available in the LISREL 8 and PRELIS 2 programs. Rather, the intent is to provide a practical guide to SEM using the LISREL approach. As such, the reader is "walked through" a diversity of SEM applications that include both factor analytic and full latent variable models, as well as a variety of data management procedures.

Uji Fit Model pada SEM

Fit model adalah rasio atau koefisien yang digunakan untuk menguji apakah model SEM yang dibentuk dapat digunakan untuk estimasi ( jika untuk uji hipotesis aja maka melihat probabilitasnyanya aja sudah cukup). ada banyak sekali koefisien untuk digunakan dalam pengujian model fit, namun yang penting adalah :

CHI-SQUARE (syarat : probabilitas > 5 %)
RMSEA ( syarat : < 0.08)
GFI ( syarat : > 0.9)
AGFI (syarat : > 0.9)
CMIN/DF ( syarat : < 2)
TLI ( syarat: > 0.95)
CFI (syarat: > 0.95)

Lalu bagaimana jika model awal tidak memenuhi fit model ?
Jawabannya adalah buat model revisinya yaitu :
1. menghilangkan hubungan yang tidak signifikan dalam model
2. melakukan korelasi antara varianse atau antar variabel 

Beberapa Kekurangan SEM

SEM mengsyaratkan jumlah sampel yang cukup besar 100-200 (minimal), jika sampel dibawah 100 maka disarankan menggunakan software PLS (partial least square). beda PLS dan SEM adalah PLS menggunakan asumsi distribusi data yang tidak normal sedangkan SEM dengan asumsi distribusi data yang normal.

SEM membutuhkan banyak jurnal karena semakin banyak hipotesis maka semakin banyak jurnal yang dibutuhkan

SEM sebaiknya digunakan untuk variabel variabel yang berbasis data pada "persepsi" seperti MSDM atau marketing atau sumber data primer dengan kuestionare (namun minimal skala data intervel). Variabel unobserved adalah variabel yang pengukurannya menggunakan indikator (lebih dari 1 indikator), namun untuk SEM sebaiknya indikator > 3. Variabel unobserved biasanya disebut : latent variabel. Untuk variabel observed sebaiknya menggunakan model path analysis (sama dengan model SEM, hanya tidak menggunakan indikator pada variabelnya). Variabel observed biasanya disebut : manifest atau proxy, biasanya digunakan pada penelitian yang menggunakan sumber data sekunder dengan skala data rasio.

Logika Pemodelan SEM

SEM merupakan suatu perluasan (extension) dari beberapa tehnik multivariate, khususnya regresi berganda dan analisis faktor. (Hair, dkk, 1992 : 426). Defenisi yang relatif sama adalah bahwa SEM : “tidak lebih “ dan “tidak kurang” dari path analytical model dengan “latent variables”. SEM = Path Anaysis + latent variables (Confirmatory factor analysis) (Mueller, 1996:129).

Jadi SEM dapat didefenisikan sebagai tehnik analisis multivariate yang menganalisis hubungan yang melibatkan variabel intervening.

Pertanyaan yang sering muncul adalah mengapa kita menggunakan SEM ? regresi saja udah cukup ! salah satu kelebihan SEM adalah melibatkan variabel intervening (variabel antara). Konsep intervening tidak sama dengan variabel moderating (walaupun ada beberapa tulisan yang menggunakan kedua istilah ini bergantian atau memiliki makna yang sama). 

Variabel intervening tidak sama dengan moderating. Variabel moderating biasanya disebut variabel kontrol, yaitu variabel yang "disisipkan" dalam model untuk melihat dampak variabel itu dalam suatu hubungan antara variabel Y dan X. jadi sifat variabel moderating ini tidak kuat (tidak ditunjang teori dan penelitian terdahulu yang kuat). 

PLS dan SEM

• Structural Equation Model (SEM) adalah covariance based, sedangkan Partial Least Square (PLS) adalah component based
• Kapan saat membutuhkan SEM dengan PLS (selanjutnya kita sebut dengan PLS saja)?
  • Model penelitian mengindikasikan lebih dari satu var dependen
  • Data tidak bersifat multivariate normal
  • Sampel kecil atau jumlah cases terbatas
  • Model penelitian melibatkan item formatif dan item refleksif sekaligus.
• Kelebihan PLS adalah kemampuannya memetakan seluruh jalur ke banyak variabel dependen dalam satu model penelitian yang sama dan menganalisis semua jalur dalam model struktural secara simultan. (Fornell and Bookstein, 1982; Barclay, Higgins, and Thompson, 1995; Gefen, Straub, and Boudreau, 2000). Kelebihan lainnya adalah hanya memerlukan sedikit cases daripada SEM (Chin and Newsted, 1999; Gefen, Straub, and Boudreau, 2000).
• Dalam PLS data tidak perlu memenuhi asumsi multivariate normal.
• Analisis SEM mengasumsikan seluruh item/indikator adalah reflektif. Sedangkan PLS bisa reflektif dan formatif.

[SMART Book] SEM: Konsep, Aplikasi dan Tutorial

Structural Equation Modeling: Konsep, Aplikasi dan Tutorial [SMART Publishing]

BUKU: A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM)

A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM), by Hair, Hult, Ringle, and Sarstedt, provides a concise yet very practical guide to understanding and using PLS structural equation modeling (PLS-SEM). PLS-SEM is evolving as a statistical modeling technique and its use has increased exponentially in recent years within a variety of disciplines, due to the recognition that PLS-SEM’s distinctive methodological features make it a viable alternative to the more popular covariance-based SEM approach. 

This text—the only comprehensive book available to explain the fundamental aspects of the method—includes extensive examples on SmartPLS software, and is accompanied by multiple data sets that are available for download from the accompanying website (www.pls-sem.com).

Langkah-Langkah dalam Pemodelan SEM

Dalam membuat permodelan SEM perlu dilakukan langkah-langkah berikut ini:

1.      Pengembangan model teoritis

Langkah pertama yang harus dilakukan dalam mengembangkan model SEM adalah mengembangkan sebuah model penelitian dengan dukungan teori yang kuat melalui berbagai telaah pustaka dari sumber-sumber ilmiah yang berhubungan dengan model yang dikembangkan. Tanpa dasar teoritis yang kuat, SEM tidak bisa digunakan. SEM tidak digunakan untuk mempengaruhi sebuah teori kausalitas yang sudah ada teorinya, karena dengan pengembangan sebuah teori yang berjustifikasi ilmiah merupakan syarat utama dalam menggunakan model SEM

2.      Pengembangan diagram alur (path diagram) untuk menunjukkan hubungan kausalitas (sebab akibat). Model penelitian yang telah dibangun pada tahap pertama akan digambarkan pada sebuah path diagram yang akan mempermudah untuk melihat hubungan-hubungan kausalitas yang akan diuji. Dalam path diagram hubungan antar konstruk akan dinyatakan melalui anak panah. Anak panah yang lurus menunjukkan sebuah hubungan kausal yang langsung antar satu konstruk dengan konstruk yang lainnya. Sedangkan garis lengkung antar konstruk dengan anak panah pada setiap ujungnya menunjukkan korelasi. Konstruk-konstruk dalam path diagram dapat dibedakan menjadi dua kelompok, yaitu:

·         Konstruk Eksogen (exogenous construct), dikenal dengan source variable atau independent variable yang tidak diprediksi oleh variabel-variabel yang lain yang terdapat dalam model. Konstruk eksogen adalah konstruk yang dituju oleh garis dengan satu ujung anak panah.

·         Konstruk Endogen (endogenous construct) yang merupakan faktor-faktor yang diprediksi oleh satu atau beberapa konstruk. Konstruk endogen dapat memprediksi satu atau beberapa kosntruk endogen lainnya, tetapi konstruk endogen hanya dapat berhubungan kausal dengan konstruk endogen.

Konsultasi Disertasi/Tesis dengan SEM

Pengantar Structural Equation Model

Structural Equation Modeling (SEM) merupakan teknik analisis multivariat yang dikembangkan untuk menutupi keterbatasan yang dimiliki oleh model-model seperti: analisis regresi, path analysis (analisis jalur), dan confirmatory factor analysis (analisis faktor konfirmatori) (Hox dan Bechger, 1998).

SEM merupakan teknik statistik yang digunakan untuk membangun dan menguji model statistik yang umumnya dalam bentuk model hubungan kausalitas. SEM merupakan teknik perpaduan (hybrid) yang menggabungkan aspek dari analisis regresi, path analysis (analisis jalur), dan confirmatory factor analysis (analisis faktor konfirmatori).

Sebagai teknik statistik multivariat, penggunaan SEM memungkinkan peneliti untuk melakukan pengujian terhadap bentuk hubungan tunggal (i.e. regresi sederhana), regresi berganda, hubungan rekursif maupun hubungan resiprokal, atau bahkan terhadap variabel laten maupun variabel yang diobservasi/ diukur langsung.

Salah satu keunggulan SEM adalah kemampuannya dalam membuat model yang mengandung variabel laten atau variabel–variabel yang tidak diukur (tidak diobservasi) secara langsung (seperti "kecerdasan", “kinerja”, dan "sikap terhadap merek"). Variabel-variabel laten tersebut diestimasi dalam model melalui variabel-variabel terukur (measured variable) yang diasumsikan mempunyai hubungan dengan variabel-variabel laten tersebut.

BUKU: Structural Equation Modeling with Mplus: Basic Concepts, Applications, and Programming

The first two chapters introduce the fundamental concepts of SEM and important basics of the Mplus program. The remaining chapters focus on SEM applications and include a variety of SEM models presented within the context of three sections: Single-group analyses, Multiple-group analyses, and other important topics, the latter of which includes the multitrait-multimethod, latent growth curve, and multilevel models.

Intended for researchers, practitioners, and students who use SEM and Mplus, this book is an ideal resource for graduate level courses on SEM taught in psychology, education, business, and other social and health sciences and/or as a supplement for courses on applied statistics, multivariate statistics, intermediate or advanced statistics, and/or research design. Appropriate for those with limited exposure to SEM or Mplus, a prerequisite of basic statistics through regression analysis is recommended.

BUKU: A Beginner's Guide to Structural Equation Modeling: Third Edition

This best-seller introduces readers to structural equation modeling (SEM) so they can conduct their own analysis and critique related research. Noted for its accessible, applied approach, chapters cover basic concepts and practices and computer input/output from the free student version of Lisrel 8.8 in the examples. Each chapter features an outline, key concepts, a summary, numerous examples from a variety of disciplines, tables, and figures, including path diagrams, to assist with conceptual understanding.

The book first reviews the basics of SEM, data entry/editing, and correlation. Next the authors highlight the basic steps of SEM: model specification, identification, estimation, testing, and modification, followed by issues related to model fit and power and sample size. Chapters 6 through 10 follow the steps of modeling using regression, path, confirmatory factor, and structural equation models. Next readers find a chapter on reporting SEM research including a checklist to guide decision-making, followed by one on model validation. Chapters 13 through 16 provide examples of various SEM model applications. The book concludes with the matrix approach to SEM using examples from previous chapters.

Designed for introductory graduate level courses in structural equation modeling or factor analysis taught in psychology, education, business, and the social and healthcare sciences, this practical book also appeals to researchers in these disciplines. An understanding of correlation is assumed.

BUKU: Discovering Structural Equation Modeling Using Stata

Discovering Structural Equation Modeling Using Stata is devoted to Stata’s sem command and all it can do. You’ll learn about its capabilities in the context of confirmatory factor analysis, path analysis, structural equation modeling, longitudinal models, and multiple-group analysis. The book describes each model along with the necessary Stata code, which is parsimonious, powerful, and can be modified to fit a wide variety of models. Downloadable data sets enable you to run the programs and learn in a hands-on way.

A particularly exciting feature of Stata is the SEM Builder. This graphic interface for structural equation modeling allows you to draw publication-quality path diagrams and fit the models without writing any programming code. When you fit a model with the SEM Builder, Stata automatically generates the complete code that you can save for future use. Use of this unique tool is extensively covered in an appendix, and brief examples appear throughout the text.

Requiring minimal background in multiple regression, this practical reference is designed primarily for those new to structural equation modeling. Some experience with Stata would be helpful but is not essential. Readers already familiar with structural equation modeling will also find the book’s State code useful.

BUKU: Structural Equation Modeling With AMOS: Basic Concepts, Applications, and Programming

The book opens with an introduction to the fundamental concepts of SEM and the basics of the AMOS program. The next 3 sections present applications that focus on single-group, multiple-group, and multitrait-mutimethod and latent growth curve models. The book concludes with a discussion about non-normal and missing (incomplete) data and two applications capable of addressing these issues.

Intended for researchers, practitioners, and students who use SEM and AMOS in their work, this book is an ideal resource for graduate level courses on SEM taught in departments of psychology, education, business, and other social and health sciences and/or as a supplement in courses on applied statistics, multivariate statistics, statistics II, intermediate or advanced statistics, and/or research design. Appropriate for those with limited or no previous exposure to SEM, a prerequisite of basic statistics through regression analysis is recommended.