Page 44 - IJET_July2021_final

Table-5: Discriminant Validity

Construct Attitude Intention Motivation Perception Personality

Attitude 0.89
Intention 0.48 0.90
Motivation 0.40 0.71 0.80
Perception 0.40 0.70 0.73 0.85

Personality 0.70 0.51 0.43 0.60 0.87
Source: Primary Data
Table-5 shows the results of Fornell discriminant validity of variables is also
lacker criterion i.e. the square root of confirmed.
average variance explained (AVE) is Table-6 shows the value of path
compared with other constructs latent coefficients, where positive value
variable value (Hair, Hult, Ringle, & denotes positive relationship between
Sarstedt, 2016). For example, the square the independent variable and
root of attitude’s AVE value should be dependent variable. The path coefficient
greater than latent variable scores value lies between 0 to 1, which could be
of other constructs in order to meet positive or negative (Hair, Hult, Ringle,
discriminant validity criteria (Wong, & Sarstedt, 2016). The highest path
2019). Table-5 shows that AVE value coefficient value is 0.386 which means
of each construct is greater than the that path “Motivation → Intention” has
latent variable value of other constructs the strongest relationship.
(specific row and column), therefore
Table-6: Path coefficient values of all the paths in Model

Path Path coefficient

Attitude → Intention 0.168

Motivation → Intention 0.386

Perception → Intention 0.332
Personality → Intention 0.032
Source: Primary Data

Multicollinearity variance inflation value (VIF) is
calculated. The VIF value should be
The constructs should not have less than “5” to prove that there is no
collinearity among themselves; collinearity among the constructs (Hair,
therefore to check collinearity issue Hult, Ringle, & Sarstedt, 2016).

34 Indian Journal of Educational Technology
Volume 3, Issue 2, July 2021