Page 46 - IJET_July2021_final

has no effect size. Therefore, the result effect size. The “attitude” construct has
of “personality” hypothesis i.e. H04 is the lowest effect on a student’s decision
cross verified by both t-statistics and related to studies.
Table-9: Effect Size (F )
2
Intention: Effect
Factors R 2 Size Inference
excluded Intention
Attitude 0.588 0.03 Small
Motivation 0.539 0.1525 Large
Perception 0.566 0.085 Medium

Personality 0.6 0 No effect
Source: Primary Data
Predictive relevance (Q2) model accuracy which may be calculated
using the following formula (Hair, Hult,
Predictive relevance is confirmation of
Ringle, & Sarstedt, 2016):
Sum of squared prediction errors (SSE)
Q2 = 1-
Sum of squared observations (SSO)

Table-10 shows the result of nearer to “1” is considered best for
blindfolding calculations where cross model accuracy. The Q2 value derived
validity redundancy option provides the from the calculations is between 0 and
value of predictive relevance (Hair, Hult, 1, therefore model predictive relevancy
Ringle, & Sarstedt, 2016). The value is good.
of Q2 should be above “0” and a value

Table-10: Cross-Validity Redundancy

Construct SSO SSE 1-SSE/SSO

Intention 862 457.819 0.469
Source: Primary Data

Scale Items Statistics factors that have obtained a mean
score below “4” are going towards
Table-11 shows the mean and
standard deviation score for each of an “agree” continuum. Therefore, it
the constructs. In likert scale “4” was can be concluded that students have
quoted as “neutral” and “1” was quoted developed an intention (post COVID-19
as “strongly agree”. This means that crises) to study via distance mode.

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