aplikasi permodelan persamaan berstruktur · pdf filedependence. sem. manova. multiple reg....

Rosseni DinFakulti Pendidikan

Universiti Kebangsaan Malaysiarosseni@yahoo.com

http://rosseni.wordpress.com

PENILAIAN e-PEMBELAJARAN

HEURISTICS EVALUATION

Rosseni DinFakulti Pendidikan

Universiti Kebangsaan Malaysiarosseni@yahoo.com

http://rosseni.wordpress.com

APLIKASI PERMODELAN PERSAMAAN BERSTRUKTUR

Katakan anda ingin mengkajitentang ePembelajaran

(1) E-Pembelajaran(2) Pembelajaran Bermakna

(3) Stail Pembelajaran

ePembelajaranHibrid

eLatihanBermakna

StailPembelajaran

Pilihan

Sorotan Literatur Kerangka Konsep

Learning Style Preference

(LSP)

Tactual

Group

Individual Kinesthetic

Auditory

Visual

Content

Delivery

Structure

Service

Outcome

Hybrid

e-Training

(HiT)

Meaningful

e-Training

(MeT)

Cooperativity

Intentionality

Construction

Activity

Authenticity

Katakan anda ingin melihat pertalianePembelajaran terhadap eLatihanbermakna dan stail pembelajaran

ePembelajaran Hibrid (i) kandungan, (ii) penyampaian (iii) perkhidmatan (iv) hasil (v) struktur

eLatihan Bermakna (i) kandungan, (ii) penyampaian (iii) perkhidmatan (iv) hasil (v) struktur

Stail Pembelajaran Pilihan (i) visual, (ii) auditori (iii) kinestetik (iv) tactile (v) berkumpulan (vi) individu

Bina RQ dari sorotan literatur.

Cari indikator untuk mengukurePembelajaran

Pastikan instrumen ePembelajransah dan boleh dipercayai

Sorotan Literatur Kerangka Konsep

Learning Style Preference

(LSP) Tactual

Group

Individual Kinesthetic

Auditory

Visual

Content

Delivery

Structure

Service

Outcome

Hybrid

e-Training

(HiT)

Meaningful

e-Training

(MeT)

Cooperativity

Intentionality

Construction

Activity

Authenticity

RQ1 RQ2

RQ3

RQ4

Di peringkat “analisis data” apakah teknik statistik yang

sesuai digunakan dengan data dan RQ?

Antara analisis multivariate yang biasa digunakan untukkajian pertalian:

Relationships?

interdependence

dependence

SEM

MANOVA

multiple reg

multivariate reg

canonical corr

PCA

CFA

cluster analysis

multidimen scaling

correspondence an

Model Multiple Regression

Hybrid eTrainingDelivery

Content

Structure

1

Model MultivariateRegression

Delivery

Content

Structure

Face-to-Face Delivery

Online Delivery

Measurement Model + Measurement Model = Structural Model

Hypothesized Structural Model of HiTs and MeT Relationship

MeT

coop e11

1

inten e21

const e31

activ e41

authen e51

HiTs

outcme11

serve12

struce13

delivere14

contente15

1

1

1

1

1

1 e16

1

Basic Terms• latent variables•measurement model• structural model•measurement error

• factor loading• direct causal effect• indirect effect

MeT

coop1.85

e61.00

1

inten.99

e71.411

const3.23

e82.51 1

activ.91

e9

2.011

authen1.30

e10

2.91

1

3.50

HiTs

outcm.66

e5

serv.94

e4

struc.93

e3

deliver1.30

e2

content1.82

e1

Normed Chi-Square 4.112RMSEA .121CFI .869TLI .842p .000

1.00

1

.821

1.001

.831 .88

1

.22

.51

e17

1

LSP

group

.89

e15

tactil

.52

e14

kines

.57

e13

audio

.84

e12

visual

1.60

e11

1.00

1

.60

1

.90

1

.94

11

.10 -.20

1.03

e181

1.01

Bagaimanamenggunakan SEM

untuk menjawab soalankajian?

PengenalanKepadaAplikasi SEM

Carta Alir Pendekatan SEM Konsep dan istilah asas

Antara situasi penyelidikan yang bolehmemanfaatkan aplikasi SEM

1. Pertalian antara beberapa IVs dengan satu DV

2. Pertalian antara beberapa IVs dengan beberapa DV

3. Pengukuran unobserved concepts

4. Perkaitan antara beberapa DV yang saling berkaitan

1. Pertalian antara beberapa IVs dengan satu DV

HiTs

serve2

struce3

delivere4

contente5

11

1

1

1

outcme11

Hypothesized Five-Factor Measurement Model for HiT

2. Pertalian antara beberapa IVs dengan beberapa DV

Hypothesized Structural Model of HiTs and MeT Relationship

MeT

coop e11

1

inten e21

const e31

activ e41

authen e51

HiTs

outcme11

serve12

struce13

delivere14

contente15

1

1

1

1

1

1 e16

1

ePembelajaranHibrid

eLatihanBermakna

StailPembelajaran

Pilihan

3. Pengukuran unobserved concepts

4. Perkaitan antara beberapa DV yang saling berkaitan

Hypothesized Result of Structural Relationships among HiTs, MeT & LSP

MeT

coop e1.52

inten e2.75

const e3.84

activ e4.84

authen e5

.95

HiTs

outcme11

serve12

struce13

delivere14

contente15

Normed Chi-Square 2.394RMSEA .081CFI .945TLI .929p .000

.89

.87

.84

.80.79

LSP

Group

e6

tactil

e7

kines

e8

visual

e9

audio

e10

.62.51.84.67.75

.49

-.25

e16

.57

-.95.43

-.25

.37

.15

e17

Sorotan literatur tertentu dapat membantu andamengusulkan RQ seperti berikut:

RQ1.Is the measurement scale for hybrid e-Training (HiT) construct-valid?

RQ2.Is the measurement scale for meaningful e-Training (MeT) psychometrically-sound?

RQ3.Are the psychometric properties for learning style preference (LSP) measure reasonable?

RQ4. Does a relationship exist among learning style preference (LSP), hybrid e-training (HiT) and meaningful e-training (MeT)?

RQ1. Is the measurement scale for hybrid e-Training construct-valid?

RQ1.1: Are learners’ acceptance of hybrid e-traininginfluenced by five factors: content, delivery, service, outcome and structure?

RQ1.2: Does each indicator has a nonzero loading on the hypothesized (targeted) factor?

RQ1.3: Does each indicator has a zero loading on the other (non targeted) factors.

RQ1.4: Are the five factors correlated? RQ1.5: Are the error terms uncorrelated?

SEM

RQ1. Is the measurement scale for Hybrid e-Training construct-valid?

RQ1.1: Are learners’ acceptance of hybrid e-training influenced by five factors: content, delivery, service, outcome and structure?

RQ1.2: Does each indicator has a nonzero loading on the hypothesized (targeted) factor?

RQ1.3: Does each indicator has a zero loading on the other (non targeted) factors.

RQ1.4: Are the five factors correlated?

RQ1.5: Are the error terms uncorrelated?

HiTs

serve12

struce13

delivere14

contente15

11

1

1

1

outcme111

Konsep dan istilah asas dalam SEM

exogenous variable endogenous variable observed; manifest; indicator variables (square) unobserved variable; latent variable; factors; construct (oval) measurement model structural model parameter estimates; path coefficient; factor loading measurement errors; residuals identification; just identified

atas tumpuan anda moga ilmu yang dikongsi dapatdimanfaatkan.

Terima kasih …