Multivariate Analysis Praktek & Review Jurnal

Teknik Multivariate: Praktek & Review Journal Overview

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multivariate Analysis Praktek & Review Jurnal

http://www.mathworks.com/cmsimages/62110_wl_stat_fig11_wl.jpg

Budi Hermana http://www.biologydirect.com/content/figures/1745-6150-4-13-3-l.jpg

http://piboonrungroj.files.wordpress.com/2011/07/hypotheses.png

Program Doktor Ilmu Ekonomi Universitas GUnadarma


Overview

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Univariate Analysis

https://rcenterportal.msm.edu/node/63


Overview

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Bivariate Analysis

https://rcenterportal.msm.edu/node/259


Factor

Cluster

Discriminant

Regression

SEM

Overview

WHY MULTIVARIATE ANALYSIS? Multivariate analysis consists of a collection of methods that can be used when several Variable

measurements are made on each individual or object in one or more samples Units (research units, sampling units, or experimental units) or observations

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Overview

Selecting a Multivariate Technique • Dependency – dependent (criterion) variables and independent (predictor) variables are present

• Interdependency – variables are interrelated without designating some dependent and others independent Cooper and Schindler; Business Research Method (8th edition)


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Overview

Dependency Techniques • • • • • •

Multiple regression Discriminant analysis Multivariate analysis of variance (MANOVA) Linear structural relationships (LISREL) Conjoint analysis – Simalto+Plus Cooper and Schindler; Business Research Method (8th edition)


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Overview

Interdependency Techniques • Factor analysis

• Cluster analysis • Multidimensional Scaling (MDS)

Cooper and Schindler; Business Research Method (8th edition)


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Overview Y

X

3

1

5 7 9 11 13 15 17 19 21

2 3 4 5 6 7 8 9 10

Apakah X berhubungan dengan Y? Regresi:

Y = 1 + 2X r=1

Jika:

X adalah jumlah burung camar terbang di lepas pantai

Y adalah jumlah burung camar terbang di lepas pantai

?

Cooper and Schindler; Business Research Method (8th edition)


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Overview

?

Deduksi Teori Riset sebelumnya Data/ Fakta

Hipotesis

Verifikasi

Research Gap

Pengukuran

Research Question

Pengujian Hipotesis

Formulasi Hipotesis

Hasil/Diskusi

Pemilihan Alat Uji? Premis

State of the Art

Road Map

Kontribusi


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

Factor Analysis

Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

Principal component analysis is a multivariate technique for transforming a set of related (correlated) variables into a set of unrelated (uncorrelated) variables that account for decreasing proportions of the variation of the original observations. Principal components is essentially a method of data

to produce a small number of derived variables that can be reduction that aims

used in place of the larger number of original variables to simplify subsequent analysis of the data



Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis Factor analysis, like principal component analysis, is an attempt to

a set of data in terms of a smaller number of dimensions than one begins with, but the procedures used to explain

achieve this goal are essentially quite different in the two methods.

If the factor model holds but the variances of the specific variables are small, we would expect both forms of analysis to give similar

results. Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis Factor analysis (more properly exploratory factor analysis) is concerned with whether the covariances or correlations between a set of observed variables can be explained in terms of a smaller

unobservable constructs known either as latent variables or common factors. number

of

Uji Validitas Konstruk (Pengujian instrumen/ Kuisener) Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis Application of factor analysis involves the following two

1

2

stages:

Determining the number of common factors needed to adequately describe the correlations between the observed variables, and estimating how each factor is related to each observed variable (i.e., estimating the factor loadings) Trying to simplify the initial solution by the process known as factor rotation Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

SEM

Factor Analysis

Types of factor analysis Exploratory Factor Analysis

Confirmatory Factor Analysis

Common factor analysis (CFA)

Principal component analysis (PCA)

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis EFA, traditionally, has been used to explore the possible underlying factor structure of a set of observed variables without imposing a preconceived structure on the outcome

CFA allows the researcher to test the hypothesis that a relationship between the observed variables and their underlying latent construct(s) exists

Exploratory or Confirmatory Factor Analysis? Diana D. Suhr, Ph.D

Rex Kline (2013) Exploratory and Confi rmatory Factor Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

PCA assumes that the common variance (C) becomes maximized and there is no unique variance (A and B) in each variable.

Hee-Ju Kim (2008)

CFA assumes that there is a substantial amount of unique variance as well as reliable common variance.


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis Varians adalah akar dari standar devisia

Common Variance yaitu varians yang dibagi dengan varians lainnya; atau jumlah varians yang dapat diekstrak dengan proses factoring

Unique Variance yaitu varians yang berkaitan dengan variabel tertentu saja; jenis variabel ini tidak dapat dijelaskan dengan korelasi hingga menjadi bagian dari variabel lain; namun varians ini masih berkaitan secara unik dengan satu variabel

Error Variance yaitu varians yang tidak dapat dijelaskan lewat proses korelasi; jenis varians ini muncul karena proses pengambilan data yang salah; pengukuran variabel yang tidak tepat, dll


Factor

Cluster

Discriminant

Regression

SEM

Factor Analysis Problem formulation Construction of the Correlation Matrix Method of Factor Analysis Determination of Number of Factors Rotation of Factors Interpretation of Factors

Calculation of Factor Scores Determination of Model Fit

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

Statistics Associated with Factor Analysis • Bartlett's test of sphericity. Bartlett's test of sphericity is used to test the hypothesis that the variables are uncorrelated in the population (i.e., the population corr matrix is an identity matrix) • Correlation matrix. A correlation matrix is a lower triangle matrix showing the simple correlations, r, between all possible pairs of variables included in the analysis. The diagonal elements are all 1.


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

• Communality. Amount of variance a variable shares with all the other variables. This is the proportion of variance explained by the common factors. • Eigenvalue. Represents the total variance explained by each factor. • Factor loadings. Correlations between the variables and the factors. • Factor matrix. A factor matrix contains the factor loadings of all the variables on all the factors


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

• Factor scores. Factor scores are composite scores estimated for each respondent on the derived factors. • Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy. Used to examine the appropriateness of factor analysis. High values (between 0.5 and 1.0) indicate appropriateness. Values below 0.5 imply not. • Percentage of variance. The percentage of the total variance attributed to each factor.

• Scree plot. A scree plot is a plot of the Eigenvalues against the number of factors in order of extraction.


Factor

Cluster

Discriminant

Regression

SEM

Factor Analysis

Praktek Membaca Hasil EFA 1. Pengelompokkan item dan penamaan faktor 2. Pengujian validitas kontruk pada kuisener

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis Kemiringan/Slope yang curam Scree plot

Cenderung 1 Faktor Eigen value 2.731 2.218 0.442 0.341 0.183 0.085

% of variance 45.520 36.969 7.360 5.688 3.044 1.420

Faktor yang terbentuk adalah yang nilai eigenvalue-nya > 1

3.0

Cumulat. % 45.520 82.488 89.848 95.536 98.580 100.000

2.5

Eigenvalue

Fact or 1 2 3 4 5 6

2.0 1.5 1.0 0.5 0.0

Cenderung 2 Faktor

1

2

3 4 5 Component Number

6


Factor Analysis

Factor

Cluster

Discriminant

Menu Factor Analysis

Regression

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

“Click” pada untuk melihat grafik: “Scree plot”

“Click” untuk memasukkan contoh butir pertanyan (8 item) yang akan direduksi/dikelompokkan menjadi beberapa faktor


Factor Analysis

Ada 2 component/factor yang nilai eigen valuenya di atas 1

8 item pertanyaan mengelompok dalam 2 faktor

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis


Factor Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Matriks rotasi menunjukan pengelompokkan yang sama. Matriks ini biasanya digunakan jika ada beberapa butir pada matriks pertama (component matriks) yang sulit dimasukanan ke faktor satu atau dua karena nilainya relatif tidak berbeda jauh

Butir 1, 2, 3, 4

Factor 1

Butir 5, 6, 7, 8

Factor 2

Penamaan faktor?

Lihat kemiripan substansi pertanyaan dalam satu faktor


Factor Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis untuk uji validitas konstruk (misal pada kuisener)

Contoh: Menurut model UTAUT dari Venkantesh (2003), contoh konstruk/variabel yang digunakan yaitu Performance Expectancy yang diukur dengan 4 butir pertanyaan dan Effort Expectancy yang diukur dengan 4 pertanyaan. 4 Butir pernyataan untuk variabel “Performance Expectancy”

4 Butir pernyataan untuk variabel “Effort Expectancy”


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Factor Analysis

Masukkan 4 butir pertanyaan untuk satu variabel (Performance Expectancy)

Berdasarkan pertimbangan praktis, perhitungan validitas dengan analisis faktor ini dilakukan per varibel. Jadi, jika ada 3 variabel maka dilakukan tiga kali perhitungan Validitas konstruk dilakukan sekaligus dalam Structural Equation Model yaitu pada “measurement model”


Factor

Cluster

Discriminant

Regression

SEM

Factor Analysis

Mengelompok dalam satu faktor, artinya benar 4 pertanyaan tersebut mengukur satu variabel yang sama yaitu “Performance Expectancy

Uji statistik

Meta-Analysis


Factor Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Contoh penyajian hasil uji validitas dan reliabilitas kuisener


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Classification

Classification Cluster Analysis & Discriminant



Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Classification Statistical techniques concerned with classification are essentially of two types.

Cluster analysis to uncover groups of observations from initially unclassified data

Discriminant function analysis works with data that is already classified into groups to derive rules for classifying new (and as yet unclassified) individuals on the basis of their observed variable values. Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

Cluster Analysis 1

SEM

Meta-Analysis

Cluster analysis

Distance and similarity measures Euclidean distance

Euclidean distances are the starting point for many clustering techniques, but care is needed if the variables are on very different scales, in which case some form of standardization will be needed

2

Method of Clustering Agglomerative hierarchical techniques k-means clustering Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Cluster Analysis

Regression

SEM

Meta-Analysis

Cluster analysis

Agglomerative hierarchical techniques

clustering techniques that proceed by a series of steps in which progressively larger groups are formed by joining together groups formed earlier in the process. more and more individuals are linked together to form larger and larger clusters of increasingly dissimilar elements

to determine the stage at which the solution provides the best description of the structure in the data, i.e.,

determine clusters.

the

number

of Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

Cluster Analysis

SEM

Meta-Analysis

Cluster analysis

Agglomerative Hierarchical Techniques Dendrogram

Nested Clusters 5

6

0.2

4 3

0.15

4

2 5

0.1

2 1

0.05

3 0

1

3

2

5

4

6

Dendogram records the sequences of merges or splits

1


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Cluster Analysis Contoh pengelompokkan 214 negara berdasarkan jumlah penduduk dan nilai PDB per kapita (sumber: Data World Bank) Data distandarisasi terlebih dahulu (dikonversi ke nilai Z pada distribusi normal)


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Cluster Analysis Nilai Z score (hasil konversi otomatis)

Menu yang digunakan


Cluster Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Sebagian tampilan “Dendogram” yang menunjukkan pengelompokka negara berbentuk diagram pohon Ada berapa klaster/kelompok negara? Dapat juga dibuat sub klaster! Nama Klaster? Contoh: • Lower income, lower-middle income, middle income, dst • Kelompok negara berpenduduk besar dengan pendapatan tinggi, ……, negara kecil dengan pendapatan kecil. dst


Factor

Cluster Analysis

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Cluster analysis

k-means clustering Method of clustering that produces a partition of the data into a particular number of groups set by the investigator

To minimize the variability within clusters and maximize variability between clusters. Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Cluster Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis


Cluster Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Cluster analysis untuk segmentasi pasar

Survey terhadap 50 konsumen (misal yang memilih beberapa merek produk elektronik). Tampilan data editor

Sumber data: Santoso (2014)

Latihan ini hanya menggunakan 3 variabel saja yaitu usia, gaji, dan tingkat konsumsi. Nilai yang dimasukkan dalam analisis klaster adalah nilai yang sudah dikonversi ke nilai Z


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Cluster Analysis

“cluster membership” membuat kolom baru pada data editor yang menunjukkan setiap konsumen (responden) masuk ke klaster 1 atau 2

Jumlah klaster yang diinginkan ditetapkan sebanyak 2 klaster saja


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Cluster Analysis

VS

Dengan nilai semula (tanpa standarisasi)

Nama/deskripsi segmen Segmen 1

Segmen 2

? ?

Dengan nilai Z (standarisasi)

Usia klaster 1 = Rata-rata + z x standar deviasi

Usia klaster 1 = 30,12 – 0.6711 x 6,043 Usia klaster 1 = 26,06 Usia klaster 2 = 30,12 + 1.095 x 6,043 Usia klaster 2 = 36,74 tahun dst untuk variabel gaji dan tingkat konsumsi


Factor

Cluster

Discriminant

Regression

Discriminant

Discriminant Analysis

SEM

Meta-Analysis


Discriminant

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Discriminant analysis

The eigen value represents the ratio of the between-group sums of squares to the within-group sum of squares of the discriminant scores.

The canonical correlation is simply the Pearson correlation between the discriminant function scores and group membership coded as 0 and 1.

The lambda coefficient is defined as the proportion of the total variance in the discriminant scores not explained by differences among the groups

The “Wilk’s Lambda” provides a test for assessing the null hypothesis that in the population the vectors of means of the five measurements are the same in the two groups Sabine Landau and Brian S. Everitt (2004), A Handbook of Statistical Analyses using SPSS.


Factor

Cluster

Discriminant

Regression

Discriminant

SEM

Meta-Analysis

Discriminant analysis Dependent non-metric? Independent variables metric or dichotomous?

No

Inappropriate application of a statistic

Yes Ratio of cases to independent variables at least 5 to 1?

No


Yes Number of cases in smallest group greater than number of independent variables?

No


Yes Run discriminant analysis, using method for including variables identified in the research question.

No Sufficient statistically significant functions to distinguish DV groups?

Yes

False


Factor

Cluster

Discriminant

Discriminant

Regression

SEM

Meta-Analysis

Discriminant analysis

Stepwise method of entry used to include independent variables?

Yes

No Entry order of variables interpreted correctly?

No Yes

Relationships between individual IVs and DV groups interpreted correctly?

Yes

No

False

False


Factor

Cluster

Discriminant

Discriminant

Regression

SEM

Meta-Analysis

Discriminant analysis Cross-validated accuracy is 25% higher than proportional by chance accuracy rate?

No

False

Yes Satisfies preferred ratio of cases to IV's of 20 to 1

No

True with caution

Yes Satisfies preferred DV group minimum size of 20 cases?

No

True with caution

Yes DV is non-metric level and IVs are interval level or dichotomous (not ordinal)?

Yes True

No

True with caution


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Discriminant

Predictor (Metric/Continous Variable)

Pembuktian isu gender dalam prilaku penggunaan internet atau adopsi TIK; analisis kesenjangan digital (digital divide) antar kelompok masyarakat atau antar wilayah/regional

Performance Expectancy Effort Expectancy Internet Self-Efficacy

Jenis Kelamin Internet Anxiety Social Influence Supporting Condition

Kategori dengan 2 kelompok: Pria dan Wanita


Discriminant

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Discriminant

2 Kategori (pria & Wanita)

Untuk menampilkan tabel hasil klasifikasi (melihat ketepatan/ tingkat prediksi secara deskriptif)


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Discriminant Statistik Uji

Struktur Matrix

Fungsi/Persamaan Diskriminan

Urutan variabel berdasarkan “discriminating power” dari yang tertinggi ke yang terendah Y = -0.131-0.058-0.202+0.162+0.957–0.212

y = 1  pria, y = 2  Wanita

“Internet self-efficacy merupakan prediktor yang paling besar kontribusinya dalam membedakan pria dan wanita berdasarkan prilaku penggunaan internet”


Factor

Cluster

Discriminant

Regression

SEM

Discriminant

Fungsi diskriminant dapat memprediski jenis kelamin dari responden berdasarkan prilaku penggunaan internet (yang diukur dengan 6 prediktor) dengan tingkat akurasi 68,8%

Meta-Analysis


Factor

Cluster

Discriminant

Regression

Regression

Regression Analysis

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Regression

Regresi Probit dan Logit Binary Outcome

Ya/Tidak Menang/Kalah Bangkrut/Tidak bangkrut Sehat/Tidak sehat Demokratis/Otoriter Sentralisasi/Desentralisasi y* is unobserved, as the underlying latent propensity that y=1

Dependent variable bersifat kategori dengan dua level

Where τ is the threshold


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Regression

Logit vs Probit The difference between Logistic and Probit models lies in this assumption about the

distribution of the errors Standard logistic distribution of errors

Park (2010) & Moore (2013)

Normal distribution of errors

Hasilnya cenderung hampir sama


Factor

Cluster

Discriminant

Regression

SEM

Regression

β0+ β1X

Logit = log odds

When x changes one unit, the logit (log odds) changes β1 units When x changes one unit, the odds changes eβ1 units Contoh

Sekretariat pasca melakukan tes masuk program pasca berdasarkan tidak parameter yaitu test masuk berbasis komputer, IPK calon pada saat S1, dan akreditasi program studi dari calon. Hasil seleksi adalah diterima atau ditolak.

Meta-Analysis


Logit Regression

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Logit Regression

Dummy variable


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Logit Regression

Faktor IPK menunjukkan peluang lebih tinggi untuk diterima dibandingkan hasil test masuk Log odd (B=0.804)) IPK > log odd Test (0.02) Exp(B) untuk IPK (2.235) > Exp(B) untuk test (1.002) Calon dari program studi terakreditasi A lebih tinggi dibandingkan dengan calon dari program studi tidak terakreditasi (kategori yang dijadikan referensi/pembanding) Peluangnya 4,718 kali dibandingkan calon dari program studi tidak terakreditasi


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Logit Regression

Output tabel tersebut mirip dengan tabel klasifikasi hasil analisis diskriminant

Logit regression menjadi teknik alternatif dengan tujuan analisisnya yang hampir sama dengan analisis diskriminant. Perbedaannya, semua prediktor pada analisis diskriminant harus berskala metrik atau kotinyu (skalanya minimal interval)


Regression

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multinomial & Ordinal Regression

Regresi logit bisa diperluas jika variabel respon (dependent variable) terdiri dari lebih dari 2 tingkat, atau r > 2 r  Nominal Multinomial Logistic Regression Models

r  Ordinal Ordered (ordinal) Logistic Regression Models


Multinomial

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Contoh Regresi Multinomial

Lulusan SMA yang akan melanjutkan ke perguruan tinggi mempunyai tiga pilihan program pendidikan tinggi, yaitu universitas, sekolah tinggi, dan vokasi. Bagaimana kecenderungan (peluang) pilihan lulusan SMA tersebut berdasarkan jenis kelamin, status ekonomi orang tua, status SMA (negeri atau swasta), serta nilai ujian (misal nilai UN untuk Matematika, IPS, dan IPA). Y  3 kategori yang bersifat nomonal (Universitas, Sekolah Tinggi, Vokasi)

Variabel eksogenus (X) terdiri dari 3 variabel yang bersifat kategorikal (dummy variable) dan 3 skor ujian yang bersifat kontinyu/Metrik


Multinomial

Factor

Cluster

Discriminant

Regression

Contoh Regresi Multinomial

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multinomial

Kategori 2 (Sekolah Tinggi) Sebagai referensi/pembanding


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multinomial

Exogenous variable/predictor yang bersifat kategorikal (variabel dummy)

Predictor yang skalanya kontinyu ditempatkan sebagai covariate


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multinomial

“Perempuan dan status ekonomi rendah cenderung memilih program vokasi, dan yang matematikanya lebih baik cenderung memilih sekolah tinggi dan universitas”


Panel Data Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multivariate untuk “Panel Data”

Panel data analysis is a method of studying a particular subject within multiple sites, periodically observed over a defined time frame.

Cross-sectional time-series analysis Analisis longitudinal (ada unsur waktu) vs cross-sectional ? Kinerja perusahaan pada sektor manufaktur dalam 5 tahun terakhir Perbandingan daya saing negara di dunia dalam 3 tahun terakhir (cross-country analysis) Spatial Dua Dimensi Temporal

Cross-sectional unit (perusahaan, negara, orang, dll) Periodic observations (Time Span)

Xij


Panel Data Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Multivariate untuk “Panel Data”

Contoh: sampel perusahaan sebanyak 60 dilihat kinerja keuangannya selama 3 tahun, misal dengan menggunakan satu dependent variabel dan 4 independent variable 180 data (baris)  60 x 3 Perusahaan Tahun

Perusahaan P1 P2 P3 P4 P5 P6 …. P58 P59 P60

P1 2011 P1 2012 2013 P1 P2 Y2012 Y2011 Y2011 2013 P2 2012 2013 P2 …. ……. P60 2011 ……. … P60 …. 2012 2013 P60

Y

10 25 15 X1201110X12012 21 22 …. Wide-Format 21 …. 19 …. 21

X1 22 15 10 22 X12013 19 16 … 18 21 …. 23

X2

X3 X4 Long-Form Data 25 32 15 Format 20 25 30 22 26 32 28 15 27 ………. X4 2013 23 22 18 19 22 31 …. …. …. 29 17 23 …… 22.…….. 17 26 18 17 31

Yit = a + bX1it + cX2it + dX3it + eX4it


Problems of heteroskedasticity and autocorrelation

Panel Data Analysis

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Model – Model Analisis “Panel Data”

Constant Coefficient Model Fixed Effect Model

Ordinary least squares (pooled) regression Least Squares Dummy Variable Model

Random Effect Model Error Component Model

LIMDEP, STATA, SAS, EViews

Random Parameter Model

Dynamic Model Robust Panel Model

Covariance Structure Model

SPSS  Tricky (SPPS command; Wide vs Long-Form Format) Analysis  Generalized Linear Model  Generalized Estimating Equation Robert Yaffee (2003). A Primer for Panel Data Analysis.


Factor

Cluster

Discriminant

Regression

SEM

SEM

Structural Equation Model

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

SEM SEM is a statistical technique for simultaneously

testing and estimating causal relationships among multiple independent and dependent constructs (Gefen et al. 2000) SEM is a statistical technique for testing and estimating those causal relationships based on statistical data and qualitative causal assumptions (Urbach and Ahlemann, 2010) SEM  A Second Generation of Multivariate Analysis First Generation of Multivariate Analysis

MANOVA. dll Cluster Analysis Factor Analysis Discriminant Analysis Multiple Regression

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

SEM

Nature

Answers a set of interrelated research questions in a single, systematic, and comprehensive analysis Supports latent variables

Not directly measured

common factor underlying factor

Relationships between the Latent Variable, which has to be derived from

theoretical considerations Structural model SEM

Measurement model Relationship between the empirically observable indicator variables and the Latent Variable


SEM

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis


SEM ~ “Analisis faktor dengan menggunakan SEM”

Factor

Cluster

Discriminant

Regression

SEM

Measurement Model Kontribusi (variansi) indikator PE1 terhadao Latent Variabel PE Korelasi antar PE dengan EE

Koefisien regresi untuk indikator ISE sebagai independent variable terhadap latent variabel ISE

Meta-Analysis


SEM

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Measurement Model & Structural Model

Koefisien regresi untuk variabel latent sebagai exogenous variable

Squared multiple Correlation antara Performance dengan ISE dan Effort

~ padanan r2 pada analisis regresi konvensional


Factor

Cluster

Discriminant

SEM

Diagrammatic Syntax

Gefen, D., D.W. Straub, & M.C. Boudreau. 2000. Structural Equation Modeling and Regression: Guidelines For Research Practice.

Regression

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

SEM

1

Nature of the fit statistics

Statistical assumptions

Analyses’ objectives

Two general approaches

Covariance-based structural equation modeling (CBSEM) LISREL, AMOS, EQS, SEPATH, RAMONA Uses a maximum likelihood (ML) function to minimize the difference between the sample

covariance and those predicted by the theoretical model

2

The component-based approach PLS Minimizes the variance of all the dependent variables instead of explaining the covariation

Urbach and Ahlemann (2010). Structural Equation Modeling in Information Systems Research Using Partial Least Squares


SEM Factor Cluster Discriminant Regression SEM Meta-Analysis


SEM

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Comparative Analysis between Techniques



SEM

Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Capabilities by Research Approach



Factor

Cluster

Discriminant

Regression

SEM

Framework for applying (PLS) in structural equation modeling

Urbach and Ahlemann (2010). Structural Equation Modeling in Information Systems Research Using Partial Least Squares

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

SEM

Y

b, d, f H1

d H3 b

Z f

X

H2

Standardized Coefficient

Persamaan Struktural: Y = a + bX ….. (1) Z = c + dY ……(2) Z = e + f X .…..(3) f  pengaruh langsung X ke Z

Z = c + d Y(a + bX) Z = c + ad + bd X) bd  pengaruh tidak langsung X ke Z melalui Y


Factor

Cluster

Discriminant

Regression

SEM

SEM

Perbandingan Hasil Regresi, LISREL, dan PLS

Regresi

LISREL


PLS

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Latihan SEM

Regressi Analysis(SPSS) & AMOS

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Review Journal

Meta-Analysis

SEM

Meta-Analysis


Meta-Analysis

Factor

Cluster

Discriminant

Peta Konsep

Regression

SEM

Meta-Analysis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Meta-Analysis Bab 2 Tinjauan Pustaka

PETA KONSEP E-BANKING

2.1. Regulasi E-Banking 2.2. Teknologi E-Banking

2.3. Dampak E-Banking 2.3.Tipe Produk E-Banking 2.4. Kinerja E-Banking

Keyword


Factor

Meta-Analysis

Research on E-Banking Service Quality: State of The Art

Cluster

Discriminant

Regression

SEM

Meta-Analysis

E-Banking Quality & Adoption

Hasil Penelusuran kata kunci di Google Scholar untuk 5 tahun terakhir

Mr Z (2000) menyatakan bahwa exploratory empirical analysis, cross-sectional; Spearman rank order correlation (karena variable SDM bersifat non-metrik); cronbach alpha untuk beberapa variable organisasi dan

Meta-Analysis

Annotated Bibliography

Mister X (2009) meneliti 273 perusahaan besar, Teori teknologi informasi dan arsitektur organisasi (disain organisasi mencakup spesifikasi wewenang pengambilan keputusan, system evaluasi kinerja, dan system kompensasi). Exploratory empirical analysis, crosssectional; Spearman rank order correlation (karena variable SDM bersifat non-metrik); cronbach alpha untuk beberapa variable organisasi. Menurut Mr T (2010), komputerisasi tidak secara otomatis meningkatkan produktifitas, tetapi tetapi merupakan komponen penting dalam system yang lebih luas mengenai perubahan organisasi yang akan meningkatkan produktifitas; Jadi perubahan organisasi merupakan bagian integral dari proses komputerisasi;

Terima kasih, selamat membuat proposal, Meneliti, dan publikasi international Dipilah & Dipilih

Kutipan/Premis


Factor

Cluster

Discriminant

Regression

SEM

Meta-Analysis

Referensi Alvin C. Rencher. 2002. Methods of Multivariate Analysis (2nd Edition). Publication,

A John Wiley & Sons, Inc.

Cooper and Schindler; Business Research Method (8th edition) Diana D. Suhr . Exploratory or Confirmatory Factor Analysis?. Statistics and Data Analysis, University of Northern Colorado. Gefen, D., D.W. Straub, & M.C. Boudreau. 2000. Structural Equation Modeling and Regression: Guidelines For Research Practice. Communications of AIS Volume 4, Article 7. Hee-Ju Kim. 2008. Common Factor Analysis Versus Principal Component Analysis: Choice for Symptom Cluster Research. Asian Nursing Research , March 2008. Vol 2. No 1 Neil H. Timm. 2002. Applied Multivariate Analysis. Springer-Verlag New York, Inc Park. 2010 & Moore. 2013. Rex Kline. 2013. Exploratory and Confirmatory Factor Analysis Robert Yaffee (2003). A Primer for Panel Data Analysis. Connect, Fall 2003 Edition, New York University Sabine Landau and Brian S. Everitt. 2004. A Handbook of Statistical Analyses using SPSS. Chapman & Hall/Crc, A Crc Press Company, Washington, D.C. Urbach and Ahlemann. 2010. Structural Equation Modeling in Information Systems Research Using Partial Least Squares. Journal of Information Technology Theory and Application. Volume 11, Issue 2, pp. 5-40, June 2010. Wolfgang Härdle and Léopold Simar. 2007. AppliedMultivariate StatisticalAnalysis (2nd Edition). SpringerVerlag, Berlin Heidelberg

Multivariate Analysis Praktek & Review Jurnal

Recommend Documents