E-learning in formal education under forced conditions using SDT and tam

Anindita Nandi, Smita Mehendale. E-learning in formal education under forced conditions using SDT and TAM. Cardiom-etry; Issue 22; May 2022; p. 268-276; DOI: 10.18137/cardiom-etry.2022.22.268276; Available from: http://www.cardiometry. net/issues/no22-may-2022/E-learning_formal_education

In recent times, e-learning or Information & Communications Technology (ICT) has increasingly gained 268 | Cardiometry | Issue 22. May 2022

When high motivation is observed in students, they tend to grasp topics in a more consistent manner and achieve higher academic performance. We try to understand the more consistently the need to feel ownership of one’s behavior”), competency (“the need to produce desired outcomes and to experience mastery”) & relatedness (“the need to feel connected to others”) on satisfaction in students pursuing formal education via e-learning. The perceptions and attitude of the instructor regarding e-learning plays a determining role in students’ e-learning usage;

2    Literature review

The TAM, developed for studying the intention of any new technology, is developed on the principles mentioned in the “Theory of Reasoned Action,” as observed by Davis. Since then, it is used meticulously by many researchers in developing the predictive frameworks about the usage intention of any emerging technology, for example - mobile devices, Internet, open software or enterprise systems in disciplines like business, medicinal sector, educational sector, and economical sector, as researched by Malaquias et al. In the past studies, TAM had been employed to a vast array of purposes, such as measuring the perceptions of simulation-based learning amongst clinical students, or evaluating the acceptance of mobile technologies by teachers, etc [8].

Since our study is linked to students’ emotions and attitude regarding the acceptance of e-learning under forced situation (Compulsion of online education due to COVID-19 pandemic), SDT is incorporated to associate few external factors that complete and elevates the drafted framework. In recent years, the TAM blended with SDT has been extensively used in the literature. Strong compatibility in them has been found, as in an evaluator study by Nikou et al [9].

2.1    TAM

Fig. 1.Original TAM model

Research suggests that the two factors majorly influencing the ATT of the users toward using virtual-based learning are PU and PEOU, as studied by Davis [11]. Apart from these two factors, in several of this domain-related research, many other external constructs have been included to modify the TAM, affecting e-learning acceptance. These external constructs can be related to supporting technology, learner characteristics, subjective norms or aiding environment, etc. Previous research has emphasized the importance of analyzing and reviewing the impact of the motivational drivers on usage intention of e-learning systems, as recorded by Huang, but there lies a research gap that can be detected in the current pandemic situation COVID-19. Hence, further research is required for extensively understanding the influencing factors towards usage intention of e-learning for formal education under forced conditions [12].

2.2    SDT

2.3    SDT and TAM

3    Research frameworks 3.1    Autonomy

Autonomy is pertinent to self-regulation, which is in variation to forced regulation by external instruments. Autonomy reflects the distinctive acts that control an individual’s behavior. Although SDT is more focused on individual’s approval of the act, it also supports contingencies. Sense of ‘learning autonomy’ is observed in students in the education field, as researched by Nikou et al. In his research, Racero et al.

Fig. 2. The conceptual framework

H1 : AUTO has a significant relation with PEOU in e-learning

H2 : AUTO has a significant relation with PU in e-learning

3.2    Relatedness

H3 : REL has a significant relation with PEOU in e-learning

H4 : REL has a significant relation with PU in e-learning

3.3    Competency

H5 : COMP has a significant relation with PEOU in e-learning

H6 : COMP has a significant relation with PU in e-learning

3.4    Instructor characteristics

H7 : INST has a significant relation with PU in e-learning

H8 : INST has a significant relation with SATIS in e-learning

3.5    PEOU and PU

H9 : PEOU has a significant impact on SATIS in e-learning

H10 : PU has a significant impact on SATIS in e-learning

3.6    Satisfaction

H11 : SATIS has a significant impact on students’ ATT in e-learning

3.7    Attitude

Attitude refers to “the degree to which a person has a positive or negative feeling towards e-learning systems.” It was stated by Salloum et al. the relationship of PEOU and PU with learner’s attitude towards usage intention of e-learning. Hence, in our framework, we are analyzing the influence of satisfaction on the attitude of users.

4    Research methodologies

This research has primarily used an online survey to accumulate the data from the desired respondents. The main reason for employing an online survey format is to target the appropriate sample and reach out to them considering the COVID-19 situation. The sample size comprises UG and PG students from various Indian universities currently undergoing online education after universities suspend physical classes. A total of 250 students were invited to fill the survey. Convenience sampling was adopted for the unbiased collection of data. 221 students responded in total, out of which 22 them were found unsuitable for analysis. Therefore, only 199 responses were used for further analysis. The online questionnaire used for this survey consisted of a five-point Likert Scale, where 1 represented ‘Strongly Disagree’ and 5 represented ‘Strongly Agree.’

The scale instruments have been derived from the constructs in previously conducted studies. They are modified to align with this research. The survey instrument includes 6 items of Autonomy (AUTO), 5 items of Relatedness (REL) & 5 items of Competency (COMP) adapted from Racero et al., 4 items of Instructor Characteristics (INST) from Al-Fraihat et al. Ibrahim et al., 6 items of PEOU & 5 items of PU from Racero et al., 5 items of SATIS from Sunkara et al.; Al-Fraihat et al. and 5 items of Attitude (ATT) adapted from Salloum et al.

The data collected has been analyzed using SPSS version 25. The various analysis techniques employed for this research are ‘Linear Regression’ and ‘Multiple Regression test.’ Data characteristic of the demographic information in Table 1.

Table 1

Demographic profile of the sample

Descriptive Statistics
S. No.	Variable	Group	Frequency	%
1)	Gender	Male	115	57.79
		Female	84	42.21
		Others	-	-
2)	Course	MBA / BBA	153	76.88
		M.Tech / BE / B.Tech	27	13.57
		MCA / BCA	12	6.03
		MSc / BSc	4	2.01
		M.Com / B.Com	3	1.51
3)	Age	Under 25	162	81.41
		Over 25	37	18.59

Source: Author’s findings and calculations

5    Data analysis and discussion 5.1    Cronbach’s alpha

The reliability of the construct is tested to confirm and verify that the results obtained are consistent and reliable. The consistency of items scale used for each of the constructs are measured by reliability analysis in SPSS through Cronbach’s Alpha test. The admissible value is 0.7 was suggested by Nunnally. It is observed that Cronbach’s alpha is above 0.7 for the constructs, and the overall value is 0.908. Hence, we can consider the questionnaire to be a reliable measurement instrument in Table 2.

Table 2

Reliability statistics

Scale	Cronbach’s Alpha	No. of Items
Autonomy	0.808	4
Relatedness	0.827	5
Competency	0.796	5
Instructor Characteristics	0.783	4
PEOU	0.866	6
PU	0.802	5
Satisfaction	0.814	5
Attitude	0.818	5
Overall value	0.908	39

Source: Author’s findings and calculations

Since multivariate problems often arise in these practical problems, many variables have significant relationships that might increase the complexity of analysis. So, Exploratory Factor Analysis (EFA) is carried out to address this problem. In carrying out EFA, the sampling adequacy is tested through KMO measures. The value obtained is computed as 0.930, which perfectly lies in the range defined by Kaiser, 1974. Bartlett’s Test of Sphericity, indicating a substantial relationship among the constructs.

5.2    Regression analysis

In consideration of the TAM and SDT framework, eleven hypotheses have been drafted. Regression statistics have been used to test the significance of each hypothesis. The summary of the regression data in Table 3 with certain observations like β value is high for high t-value and small p-value.

H1: Regression Analysis carried out to test the hypothesis with Autonomy (AUTO) as the predictor variable and PEOU as the dependent variable, we obtained (β=0.799, t=21.186, p<0.05). Hence, we can conclude that a significant relationship exists between AUTO and PEOU. The obtained outcome of R2 is 0.695, indicating 69.5% of the variation in PEOU is explained by AUTO, which measures the overall stability of the association and reflects the extent to which a predictor factor is associated with a dependent factor. The unstandardized coefficient of β illustrates the degree to which the predictor variable predicts the dependent variable. For every one unit of change in AUTO, a 0.799 unit of change can be observed in PEOU. Hence, the hypothesis is accepted.

The obtained result proves the existence of a significant relationship between AUTO, REL, COMP(predic-tor variable), and PEOU (dependent variable) with the calculated value of (R2=0.725) indicating that the predictor variables explain 72.5% variation in the PEOU. The unstandardized coefficient of β illustrates the extent to which predictor variables predict the dependent variables. AUTO (β=0.636) has the strongest effect on PEOU, followed by REL (β=0.151). VIF (Variance Inflation Factor) and Tolerance are used to measure the presence of multicollinearity. VIF >10 and Tolerance < 0.2 indicates multicollinearity. From Table 5, we can see VIF values are between 2-4, and Tolerance values are in

Table 3

Regression values

Hypothesis	Predictor Variable	Dependent Variable	Un-standardized Coefficients		F	t	p	R2	Hypothesis Supported?
			β	Std. Error
H1	AUTO	PEOU	0.799	0.038	448.852	21.186	0.00	0.695	Yes
H2	AUTO	PU	0.621	0.05	154.728	12.439	0.00	0.44	Yes
H3	REL	PEOU	0.746	0.048	239.752	15.484	0.00	0.549	Yes
H4	REL	PU	0.787	0.042	349.61	18.698	0.00	0.64	Yes
H5	COMP	PEOU	0.705	0.056	156.65	12.516	0.00	0.443	Yes
H6	COMP	PU	0.855	0.042	423.645	20.583	0.00	0.683	Yes
H7	INST	PU	0.775	0.039	391.136	19.777	0.00	0.665	Yes
H8	INST	SATIS	0.805	0.04	397.444	19.936	0.00	0.669	Yes
H9	PEOU	SATIS	0.614	0.057	114.633	10.707	0.00	0.368	Yes
H10	PU	SATIS	0.916	0.034	707.907	26.607	0.00	0.782	Yes
H11	SATIS	ATT	0.843	0.039	471.292	21.709	0.00	0.705	Yes

Source: Author’s findings and calculations

Table 4

Result

Model	R	R2	Adjusted R2	Std. The error of the Estimate
1	.851	0.725	0.720	0.55404

Source: Author’s findings and calculations

Table 5

Multiple regression statistics

Coefficients
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	Collinearity Statistics
		B	Std. Error	Beta			Tolerance	VIF
1	(Constant)	0.221	0.126		1.746	0.082
	Autonomy	0.610	0.057	0.636	10.786	0.000	0.406	2.462
	Relatedness	0.152	0.077	0.151	1.969	0.049	0.341	4.146
	Competency	0.132	0.069	0.125	1.919	0.016	0.334	2.990

Source: Author’s findings and calculations the acceptance range, signifying the absence of multicollinearity Table 6 shown the result.

The obtained result proves the existence of a significant relationship between AUTO, REL, COMP, INST (predictor variable), and PU (dependent variable) with the calculated value of (R2=0.785) indicating that predictor variables explain 78.5% variation in the PU. INST (β=0.373) has the strongest effect on PU followed by COMP (β=0.357), REL (β=0.199), and AUTO (β=0.142). From Table 7, we can see VIF values are between 2-4, and Tolerance values are in the acceptance range, signifying the absence of multicollinearity. Table 8 shows the result

Table 6

Result

Model	R	R2	Adjusted R2	Std. The error of the Estimate
1	.886	0.785	0.781	0.47904

Source: Author’s findings and calculations

Table 7

Multiple regression statistics

Coefficients
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	Collinearity Statistics
		B	Std. Error	Beta			Tolerance	VIF
1	(Constant)	0.070	0.112		0.628	0.531
	Autonomy	0.040	0.049	0.142	0.805	0.022	0.402	2.487
	Relatedness	0.195	0.069	0.199	2.836	0.005	0.225	4.437
	Competency	0.369	0.064	0.357	5.772	0.000	0.290	3.450
	Instructor Characteristics	0.354	0.051	0.373	6.888	0.000	0.378	2.649

Source: Author’s findings and calculations

Table 8

Result

Model	R	R2	Adjusted R2	Std. The error of the Estimate
1	.900	0.810	0.807	0.46563

Source: Author’s findings and calculations

Table 9

Multiple regression statistics

Coefficients
Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	Collinearity Statistics
		B	Std. Error	Beta			Tolerance	VIF
1	(Constant)	0.089	0.107		0.834	0.405
	Instructor Characteristics	0.285	0.054	0.290	5.250	0.000	0.320	3.126
	PEOU	0.407	0.043	0.117	7.163	0.007	0.539	2.855
	PU	0.676	0.059	0.653	11.449	0.000	0.300	3.338

Source: Author’s findings and calculations

we can see VIF values are close to 3, and Tolerance values are in the acceptance range, signifying the absence of multicollinearity.

Thus, the proposed framework holds for this domain as proven through the data analysis.

6    Conclusions

The spread of COVID-19 has forced most of the service industry around the world to look for alternatives to sustain. The Internet became the savior, and ‘Online became the ‘New Normal.’ The education industry is also forced to shift from offline to online overnight, impacting millions of students and teachers worldwide. Students in the middle of their academic year are forced to accept and adapt to the new technology of e-learning. Under TAM’s conceptual framework with SDT, the proposed hypotheses were tested to comprehend better the influencing determinants of e-learning acceptance by students pursuing their educational degree.

With adequate infrastructure required for the smooth functioning of e-learning, further studies can be carried out to understand the technical system quality or support system quality. A comparative study can also be conducted on the cost of traditional education methods and online modes of education. Since we have only studied a sample of university students, the drivers for e-learning acceptance can differ for students with changing pedagogy. Generalization of this framework would be unsatisfactory, thus leaving a further scope of research.

Statement on ethical issues

Research involving people and/or animals is in full compliance with current national and international ethical standards.

Conflict of interest

None declared.

Author contributions

The authors read the ICMJE criteria for authorship and approved the final manuscript.