You are on page 1of 5

Indices for Measurement Model Analysis using Partial Least Square SEM (PLS – SEM)

Assessment Test Name of Index Level of Acceptance Literature Support

1. Reliability 1. Internal Consistency Reliability Cronbach Alpha > 0.7 Robinson, Shaver & Wrightsman (1991)

Composite Reliability > 0.708 Hair et al (2010), Hair et al (2014)

2. Convergent 1. Average Variance Explained (AVE) AVE score > 0.5 Hair et al (2010), Hair et al (2014)
Validity
2. Factor Loadings Loadings for indicators > 0.708 Hair et al (2014)

3. Discriminant 1. Cross-Loadings Assessment Cross-loadings scores differ by 0.1 Vinzi, Henseler, Chin & Wang (2010)
Validity
2. Fornell and Larcker criterion AVE > r2 Hair et al (2010), Hair et al (2014)
(1981)
3. HTMT criterion (2014) HTMT.85, HTMT.90, HTMTinference Henseler et al (2014)
HTMT.85 – Kline (2011)
HTMT.90 – Gold, Malhotra, Segar (2001)
Indices for Structural Model Analysis using Partial Least Square SEM (PLS – SEM)

Assessment Test Name of Index Level of Acceptance Literature Support


1. Collinearity Variace Inflator Factor (VIF) VIF < 3.3 / VIF < 5.0 Diamantopoulos & Sigouw (2006), O’Brien
(2007), Rogerson (2001)

2. Path Co-efficient Path Co-efficient p value < 0.05, t value > 1.96 Hair et al., (2014)

3. R2 Co-efficient of determination 0.75 – Substantial Hair et al., (2014),


0.50 – Moderate
0.25 - Weak
4. f2 Effect size to R2 0.35 – Large effect size Hair et al., (2014), Cohen (1988)
0.15 – Medium effect size
0.02 – Small effect size
5. Q2 Stone-Geisser Q2 Predictive Value larger than 0 indicates that Stone (1974), Geisser (1974), Hair et al., (2014)
relevance exogenous constructs have predictive
relevance over endogenous construct
MEASUREMENT MODEL ANALYSIS

ASSESSMENT OF INTERNAL CONSISTENCY AND CONVERGENT


VALIDITY

Sample Table for Internal Consistency and Convergent Validity Reporting


Construct Item Loading CR AVE Convergent
Validity
(AVE > 0.5)
XXX XXX1 0.870 0.902 0.697 YES

XXX2 0.862

XXX3 0.849

XXX4 0.752

YYY YYY1 0.856 0.924 0.802 YES

YYY2 0.921

YYY3 0.909

ZZZ ZZZ1 0.902 0.929 0.814 YES

ZZZ2 0.907

ZZZ3 0.890
ASSESSMENT OF DISCRIMINANT VALIDITY

Sample Table for Loading and cross loading of constructs


Items XXX YYY ZZZ
XXX1 0.703 0.381 0.343
XXX2 0.752 0.386 0.438
XXX3 0.817 0.487 0.457
XXX4 0.722 0.464 0.449
YYY1 0.523 0.798 0.662
YYY2 0.460 0.742 0.542
YYY3 0.498 0.808 0.580
ZZZ1 0.351 0.482 0.759
ZZZ2 0.459 0.663 0.790
ZZZ3 0.419 0.597 0.793

Sample Table for Fornell and Larcker criterion


XXX YYY ZZZ
XXX 0.783
YYY 0.622 0.774
ZZZ 0.721 0.647 0.824
Note: Diagonal elements shaded and highlighted in bold represent the square root of AVE. Off diagonal
elements are simple bivariate correlations between the constructs. HOFC denotes Higher Order Formative
Construct.

Sample Table for Heterotrait Monotrait (HTMT) Criterion for Discriminant Validity
XXX YYY ZZZ
XXX -
YYY 0.806 -
ZZZ 0.657 0.668 -
Criteria: Discriminant validity is established at HTMT 0.85 / HTMT0.90
STRUCTURAL MODEL ANALYSIS

Sample table for collinearity assessment


INT
XXX 1.428
YYY 1.258
ZZZ 1.465
Criteria: VIF < 3.3 (Diamantopoulus & Sigow, 2006)
VIF < 5.0 (O’Brien, 2007; Rogerson, 2001)

Sample table for Path Coefficient (Inclusive of Moderation)


Direct Effect (ß) Standard Error T-statistic P value

XXX  YYY 0.335 0.084 4.000** 0.000

XXX*ZZZ  YYY 0.257 0.084 3.070** 0.002

Sample table for indirect effect report (Mediation)


Indirect Standard T Statistic Confidence Confidence p-Value Result
Effect Error Interval Interval
(|O/STERR|)
(STERR) Low Up

XXX  YYY  ZZZ 0.079 0.031 2.514* 0.029 0.151 0.012 Significant
*p<0.05, **p<0.01

Sample table for Determination of Co-efficient (R2), Effect size (f2) and Predictive
Relevance (Q2)
Co-efficient Predictive
of Relevance Effect Size f 2
Determination
R2 Q2 ZZZ Effect Size
ZZZ 0.421 0.326
YYY 0.370 0.301 0.148 Small to Medium
XXX 0.080 Small to Medium
Note: i) For interpretation of effect size, Hair et al., (2014) suggest the following (0.02 – small effect size,
0.15 – Medium effect size, 0.35 – Large effect size).
ii) R2 score interpretation (0.75 – Substantial, 0.50 – Moderate, 0.25 – weak)
iii) Q2 score (value larger than 0 indicates that the exogenous construct have predictive relevance
over endogenous constructs)

You might also like