Page 10 - rohana_journal_No_13-2020-PRINT_03_of_03 -edit-2023_07-corre

Research Journal of the University of Ruhuna, Sri Lanka- Rohana 13, 2021

Further Average Variance Extracted (AVE) values above 0.5 indicate that the

measures have convergent validity (Bagozzi & Yi, 1988). The current study with the

lowest AVE of 0.558 adheres to this as well. Refer to Table 2 for more details.

In order to measure the discriminant validity “The Fornell-Larker criterion” can be
used. According to the Fornell-Larker criterion, each construct’s AVE should be

higher than its squared correlations with other constructs (Fornell & Larker, 1981).

Refer to Table 03. Please note that values along the diagonal represent square roots

of the AVE values. Adherence to these three criteria signals the main forms of
reliability and validity for the constructs used in the study.

The study also checked for multi-collinearity, since it can have adverse effects on

the study. The highest correlation recorded is 0.588 while the highest VIF value
recorded is 1.345 (Refer to Table 4). These values are well below the accepted

levels of 0.7 and 10 respectively (Hair et al., 2013).

If all factor level VIF’s resulting from full collinearity tests are equal to or lower
than 3.3 the model can be considered free of common method bias (Kock, 2015).

The highest VIF value recorded in the current model is 2.81, thus we can determine

that the model is a common method bias-free.

The SRMR is defined as the difference between the observed correlation and the
model implied correlation (Hu & Bentler, 1999). Thus, it allows assessing the

average magnitude of the discrepancies between observed and expected correlations
as an absolute measure of (model) fit criterion (Henseler et al., 2014). A value less

than 0.10 (or 0.08, conservative) is considered a good fit (Hu & Bentler, 1999).
SRMR Values for both the saturated model and the estimated model for the current

study are below 0.08 signalling a good fit. Refer to Table 5.

vi