Examining the interplay of internet banking service quality, e-satisfaction, e-word of mouth and e-retention: a post pandemic customer perspective

Figures & data

Figure 1. Estimation model.

Table 1. Operational definitions of research constructs.

Construct	Operational definition
Internet banking service quality	The perception by an e-customer of how well the performance of internet banking met their prior expectations.
E-customer satisfaction	The subjective evaluation of whether an e-customer is happy after using internet banking services.
E-word of mouth	Any personal communication by a customer to other customers, people or prospects about their experiences with a financial service provider over internet based platforms.
E-customer retention	The bank’s ability to turn online banking customers into repeat customers over a known period time and the ability to keep them from switching.

Table 2. Fit evaluation for the structural model.

Model fit Index	Description	Thresholds	Obtained
X^2/df	Ratio of the model Chi square to degrees of freedom	X^2/df < 3-Perfect fit X^2/df < 5-Acceptable fit	1.856
GFI	Proportion of variance accounted for by the estimated population covariance. Scaled from 0 to 1, the closer to 1, the better fit.	0.95 < GFI < 1-Perfect fit 0.90 < GFI < 0.95-Acceptable fit	0.955
RMR	The square root of the difference between the residuals of the sample covariance matrix and the hypothesized model. The smaller, the better fit.	0 < RMR < 0.01-Perfect fit 0.01 < RMR < 0.05-Good fit 0.05 < RMR < 0.08- Acceptable fit	0.04
RMSEA	A parsimony adjusted index. Values closer to 1, the better fit	0 < RMSEA < 0.05-perfect fit 0.05 < RMSEA < 0.08-Acceptable fit	0.057
TLI	A TLI of 0.95 indicates that the model of interest improves by 95% relative to the null model	0.95 < TLI < 1-Perfect fit 0.90 < TLI < 0.95- Acceptable fit	0.955
CFI	Compares the fit of a target model to the fit of a null model	0.97 < CFI < 1-Perfect fit 0.90 < CFI < 0.97-Acceptable fit	0.963
NFI	Relative location of the model between the saturation model and the independence model	0.95 < NFI < 1-Perfect fit 0.90 < NFI < 0.95-Acceptable fit	0.923

x²/df: ratio of model chi square to degrees of freedom; GFI: Goodness of Fit Index; RMR: Root Mean Residual; RMSEA: Root Mean Square Error of Approximation; TLI: Tucker Lewis Index; CFI: Comparative Fit Index; NFI: Normed Fit Index.

Table 3. Demographic profile.

Variable	Category	Frequency	Percentage
Gender	Male	113	43.3
	Female	148	56.7
Age	18–35 years	160	61.6
	36–55 years	57	21.8
	Above 55 years	44	16.9
Education	No formal education	0	0
	Ordinary level	32	12.3
	Diploma/Certificate	11	4.2
	Bachelor’s degree	169	64.8
	Master’s degree	36	13.8
	Doctorate degree	13	5.0
Income source	Formal employment	135	51.7
	Self employment	75	28.7
	Diaspora remittances	41	15.7
	Pension funds	10	3.8

Source: Primary data.

Table 4. Discriminant validity: the Fornell-Larcker criterion.

Constructs	1	2	3	4
E-Banking Service quality (1)	0.769
E-Customer Satisfaction (2)	0.539	0.829
E-Word of Mouth (3)	0.703	0.520	0.839
E-Customer Retention (4)	0.427	0.596	0.399	0.868

AVEs are bolded on the diagonal: construct correlations are below the diagonal.

Table 5. Psychometric properties of the measurement model.

Construct	Item	Loading	SMC	AVE	MSV	CR	CA
Internet	EBSQ1	0.701	0.491	0.591	0.494	0.935	0.940
Banking	EBSQ2	0.707	0.499
Service	EBSQ3	0.754	0.569
Quality	EBSQ4	0.815	0.665
	EBSQ5	0.815	0.676
	EBSQ6	0.822	0.648
	EBSQ7	0.814	0.663
	EBSQ8	0.787	0.620
	EBSQ9	0.774	0.599
	EBSQ10	0.693	0.480
E-Customer	ECS1	0.744	0.554	0.687	0.355	0.929	0.931
Satisfaction	ECS2	0.783	0.612
	ECS3	0.824	0.679
	ECS4	0.915	0.838
	ECS5	0.869	0.756
	ECS6	0.828	0.686
E-Word	EWM1	0.819	0.671	0.705	0.494	0.923	0.919
Of	EWM2	0.833	0.694
Mouth	EWM3	0.833	0.694
	EWM4	0.890	0.793
	EWM5	0.820	0.672
E-Customer	ECR1	0.860	0.739	0.753	0.355	0.924	0.921
Retention	ECR2	0.880	0.774
	ECR3	0.770	0.592
	ECR4	0.951	0.904

SMC: Squared Multiple Correlation; MSV: Maximum Shared Variance; AVE: Average Variance Extracted; CR: Composite Reliability; CA: Cronbach Alpha.

Figure 2. The structural model. The path diagram shows standardized estimates. Source: Survey data.

Table 6. Results of hypothesis testing.

Hypothesis	Causal Path	(1)	(2)	(3)	(4)	(5)	(6)
H1	EbSq->EcS	0.542	0.652	0.079	8.302	***	Supported
H2	EbSq->EWoM	0.596	0.639	0.077	8.260	***	Supported
H3	EcS->EWoM	0.198	0.176	0.054	3.263	0.01	Supported
H4	EWoM->EcR	0.125	0.117	0.056	2.102	0.04	Supported
H5	EcS->EcR	0.532	0.413	0.053	7.732	***	Supported

(1): standardized estimate; (2): unstandardized estimate; (3) standard error; (4): C.R/t-value; (5): p-value; (6): result; EbSq; e-banking service quality; EcS; E-Customer Satisfaction; EWoM; E-Word of Mouth; EcR; E-Customer Retention.

Data availability statement

The materials and data used in this research will be availed by the authors upon a reasonable request.