Desktop version

Home arrow Education arrow New Ways of Studying Emotions in Organizations

Source

Construct Reliability

In PLS-SEM applications, the recommended measure of reliability is the composite reliability score which, unlike Cronbach alpha, does not assume tau equivalence and therefore “prioritizes indicators according to their reliability during model estimation” (Hair, Ringle, & Sarstedt, 2011b). All constructs in the research model showed acceptable levels of composite reliability being above 0.70 (see Table 5).

Construct Validity

To assess the validity of the measurement or outer model in PLS-SEM it is necessary to examine both convergent and discriminant validity as described by Henseler et al. (2009) and Hair et al. (2011a).

Convergent Validity

Convergent validity is concerned with the extent to which a set of indicators are representative of the construct they are meant to measure; and, whether a construct and its associated indicators are significantly distinct or different from other constructs in the overall model (Vogt, 1993, p. 44). The test for convergent validity is done by assessing the average variance

Table 5. Smart PLS Reliability and Validity Results.

AVE

SQRT

AVEa

Composite

Reliability

R

Square

Cronbach

Alpha

Communality

Redundancy

AFFCT

0.58

0.76

0.89

0.34

0.86

0.58

0.19

CONSC

0.57

0.75

0.79

0.71

0.60

0.57

0.40

ER

0.60

0.77

0.81

0.64

0.66

0.60

0.38

INFL

0.63

0.79

0.84

0.73

0.70

0.63

0.46

INT

0.61

0.78

0.83

0.71

0.69

0.61

0.43

IS

0.49

0.70

0.75

0.48

0.49

0.49

0.23

ITL

0.84

0.92

0.94

0.47

0.91

0.84

0.13

JOBSAT

0.92

0.96

0.96

0.32

0.92

0.92

0.30

MOT

0.62

0.79

0.83

0.75

0.69

0.62

0.46

OEI

0.59

0.77

0.93

0.00

0.92

0.40

0.00

SA

0.60

0.78

0.82

0.74

0.67

0.60

0.44

aSQRT indicates the square root of the AVE, which is necessary to calculate the Fornell- Larcker criterion.

extracted (AVE) of each construct in the PLS-SEM output. An AVE value of 0.50 or more signifies that the latent variable explains at least half of the variance of its indicators and therefore demonstrates an acceptable level of convergent validity (Hair et al., 2011a). All the constructs in the research model showed an AVE of 0.50 or more, which denotes an acceptable level of convergent validity (see Table 5).

 
Source
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >

Related topics