Factors influencing adoption of online teaching by school teachers: A study during COVID‐19 pandemic – Sangeeta – – Journal of Public Affairs


The COVID‐19 pandemic was declared a national emergency in most of the countries in early 2020. This pandemic forced counties all over the world to adopt a sequence of emergency management mechanisms (Zhang, Wang, Yang, & Wang, 2020). Government of different countries initiated measures such as lockdown of cities, shutting down of educational institutions as well as implementation of strict social distancing measures. Till April 13, 2020, approximately 1.725 billion learners were affected due to closure of schools in response to the pandemic. Consistent with UNESCO monitoring, 192 countries have implemented nationwide closure of academic institutions and 5 have implemented local closures, impacting about 99.9% of the world’s student population (UNESCO Report, 2020). A strategic move named as “Suspending Classes without Stopping Learning” was initiated by Chinese Government (Zhang et al., 2020) and later followed by governments of other countries to shift to online teaching while schools were closed. Previous outbreaks of infectious diseases like swine flu have prompted widespread school closings round the world, with varying levels of effectiveness (Barnum, 2020). If school closures occur late relative to a pandemic, they are less effective and should not have any impact in the least (Zumla, Yew, & Hui, 2010). Additionally, in some cases, the re‐opening of schools after a period of closure has resulted in increased infection rates. As closures tend to occur concurrently with other interventions like public gathering bans, it is often difficult to live the precise impact of schools closures (Barnum, 2020).

The crisis is usually coupled with opportunities and it’s time to understand the complete potential of technology for learning within the wake of this medical emergency and keeping the students’ safety in mind alongside their academic concern, different stakeholders within the education space are endorsing online learning in order that the training only grow and do not recede. Online teaching as a response to pandemics and COVID‐19, in particular, actually started in China through their “school’s out, class’s in” response as an initiative to mitigate the academic loss due to the disease (Zhou, Wu, Zhou, & Li, 2020). For schools in India, April is actually the beginning of new session and as a result, schools feared loss of teaching hours. Therefore, schools in India followed by the guidelines of state and union government and initiated virtual classes to bridge the gap. Majority of personal schools and other educational institutions initiated mandatory virtual classes, and thus, teachers are unfailingly sharing their lessons over Skype call, Zoom call, Google hangouts, Microsoft teams or the other virtual class options to stay the training on.

This research is an effort to understand the factors which encourage adoption of online teaching at the time of COVID‐19 pandemic induced lockdown.

COVID‐19 pandemic motivated the academic institutions and schools to go online. Teachers as well as students are exposed to new platforms such as Microsoft teams, Google hangouts, Zoom and others. In order to conduct classes smoothly, proper protocols and directions were given to the students as well as parents to facilitate adaptation to this novel channel of learning (Saxena, 2020). A number of previous research studies have empirically validated factor leading to online learning but majority of these studies have focussed on higher education (Mei, Brown, & Teo, 2018; Mosunmola, Mayowa, Okuboyejo, & Adeniji, 2018; Tseng, Lin, Wang, & Liu, 2019). There are negligible studies focusing on adoption of online teaching. To the best of the author’s understanding, this is the first study in Indian context to understand the factors leading to online teaching due to pandemic emergency.

There are a number of challenges from the perspectives of learners and academicians to move from offline to online mode of learning. Engaging students and indulging them in teaching‐learning progression is yet another stumbling block in adoption of online teaching. Developing content that not only covers the curriculum but also engages learners is need of the hour. Adoption of online teaching, though laudable during this COVID‐19 time, but it is also important to develop and enhance the quality of teaching and courses delivered during pandemic. An operative, competent, and resourceful educational system is mandatory to deliver online delivery of classes. There are end number of technological issues like downloading issues, installing apps, low internet‐speed, issues related login ids, inaudible voice, video and so on. Even learners also find e‐teaching boring and less interactive. Therefore, the study has been undertaken to comprehend the perception of school teachers so that sufficient efforts could be undertaken to improve delivery, evaluation and interaction among students and teachers. This research thus provides useful information to the school administrators to implement hassle‐free online teaching at this phase of lockdown. The study adopts constructs form Unified Theory of Acceptance and Use of Technology (UTAUT) by Venkatesh, Morris, Davis, and Davis (2003) but revised by Dwivedi, Rana, Jeyaraj, Clement, and Williams (2019). Based on the findings, practical insights are provided to school authorities to facilitate adoption, acceptance and use of online teaching during pandemic outbreak.

The rest of this paper is organized as follows: Section 2 reports the theoretical background and hypotheses formulation. Research methodology, measurement items to carry out the survey, sampling and data collection procedures are discussed in Section 3. Section 4 includes the statistical analysis and hypotheses testing followed by Section 5 discussing the empirical findings in detail and excerpts implications, limitations and directions for future research.


2.1 Theories on acceptance IT/IS and use

Different models for the introduction and adoption of information technology innovations have been elucidated by previous researchers such as Social Cognitive Theory (SCT) (Bandura, 1986), the Technology Acceptance Model (TAM) (Davis, 1989), the Theory of Planned Behavior (Ajzen, 1991), extended TAM (Venkatesh & Davis, 2000), the model combining TAM and the Theory of Planned Behavior (Taylor & Todd, 1995), and the Model of PC Utilization (Thompson, Higgins, & Howell, 1991) and UTAUT and UTAUT2 (Venkatesh et al., 2003; Venkatesh, Thong, & Xu, 2012).

Among these, UTAUT and UTAUT2 by Venkatesh et al. (2003) and Venkatesh et al. (2012) have been applied widely in various domains to understand users’ behaviour concerning different technologies (Appendix A). This research study adopts the modified UTAUT (Dwivedi et al., 2019). Based on the empirical findings (Tandon & Kiran, 2019; Tseng et al., 2019), four moderators were also excluded. Additionally, the user attitude was also included in this research based on revised UTAUT (Dwivedi et al., 2019). The significance of attitude in explaining technology acceptance is also based on previous research studies (Kim et al., 2009; Tandon, Kiran, & Sah, 2016). Previous models like TRA, TPB and DTPB have also validated attitude as a construct in their theory.

The inclusion of attitude in models of IS/IT acceptance is consistent with TRA (Ajzen and Fishbein (1980). The TRA model claims that attitude completely mediates the relationship between beliefs and intention (Taylor and Todd, 1995). TAM hypothesized that the positive attitude is developed when any technology is easy to comprehend (Davis, Bagozzi, & Warshaw, 1989). Attitude has also been validated by previous researchers that used UTAUT also (Aboelmaged, 2010; Alshare & Lane, 2011; Chen & Lu, 2011; Mosunmola et al., 2018; Rana, Dwivedi, Lal, Williams, & Clement, 2017; Tandon et al., 2016). Based on the literature reviewed, direct effect of attitude on behavioural intention is also proposed in this study.

2.2 Hypotheses development

This research study develops hypotheses based on UTAUT and validates performance expectancy, effort expectancy, facilitating conditions, social influence and attitude.

Mosunmola et al. (2018) re‐evaluated UTAUT in adoption of mobile learning and found that model holds good for adoption of mobile learning. The study by Mei et al. (2018) confirmed facilitating conditions as a strongest determinant of intention to adopt Web 2.0 for language learning. Tseng et al. (2019) also corroborated that constructs of UTAUT2 hold equally good for adoption on MOOC programmes by teachers. Nikou and Economides (2019) validated extended TAM and found perceived ease of use, facilitating conditions as determinants of behavioural intention to use mobile‐based assessments. The study by Wong (2016) also confirmed facilitating conditions as a vital construct for adoption of educational technology by teachers. Teo and Noyes (2014) validated UTAUT model to comprehend pre‐service teachers’ intentions to use information technology. Teo et al. (2019) also emphasized that perceived usefulness and facilitating conditions simplify the process of adoption of Web 2.0 for teaching. Pynoo et al. (2011) applied UTAUT to study digital learning by secondary school teachers and found performance expectancy and social influence as significant drivers. The studies by Mosunmola et al. (2018), Tandon and Kiran (2018) and Jairak, Praneetpolgrang, and Mekhabunchakij (2009) validated positive impact of social influence on behavioural intention but the study by Nassuora (2012) reported insignificant impact of social influence. As most of the related literature has led to inconsistent findings concerning the impact of constructs of UTAUT, the model needs to be validated further to understand the comprehensiveness of the model as a theoretical underpinning for examining factors leading to adoption of online teaching by teachers during pandemic outbreak.

Kozma (2011) emphasized that “for smooth learning, the school curriculum should increasingly be interwoven with ICT, and students should be given opportunities to use advanced technological tools and digital resources for creative and innovative problem solving,” (p. 115). The study by Basilaia and Kvavadze (2020) confirmed that immediate transition to e‐learning by schools due to COVID 19 was successful in Georgia. Bao (2020) presented a case study of Peking University and emphasized upon the contingency plan to deal with the pandemic types situations. König, Jäger‐Biela, and Glutsch (2020) studied the impact of school closures in Germany and empirically validated that competence of teachers and opportunities provide to them to acquire digital competence are the significant factors in adoption of online teaching.

On the basis of above discussion, following hypotheses have been posited:

H1.Performance expectancy positively influences behavioural intention.

H2.Effort expectancy positively influences behavioural intention.

H3.Facilitating conditions positively influences behavioural intention.

H4.Social Influence positively influences behavioural intention.

H5.Performance expectancy positively influences attitude towards online teaching.

H6.Effort expectancy positively influences attitude towards online teaching.

H7.Facilitating conditions positively influences attitude towards online teaching.

H8.Social influence positively influences attitude towards online teaching.

H9.Attitude positively influences behavioural intention to adopt only e teaching.

H10.Behaviour intention positively influences actual use of online teaching.

H11.Facilitating conditions positively influences actual use of online teaching.

H12.Attitude positively influences actual use of online teaching.

Figure 1 has been proposed on the basis of above discussion.


Proposed model


3.1 Measurement development

After a thorough literature revision, a survey instrument was elaborated based on established measurement scales. The study adopted the modified UTAUT model validated by Dwivedi et al. (2019). The study adopts the scale items from Venkatesh et al. (2003). The scale items on attitude were adopted from the previous study of Mosunmola et al. (2018). This condensed model help us to understand the perception of school teachers regarding delivery of online classes (Appendix B).

3.2 Sample and data collection

An item screening test was conducted with an expert panel of 12 University Professors, teachers of schools, and researchers to confirm the face validity of the scale items. This panel suggested minor amendments in language and applicability as well as alternatives where applicable, and the scales were modified accordingly. According to Onwuegbuzie and Collins (2007), mixed method sampling is highly imperative where the respondents are unknown and difficult to reach. Therefore, non‐probability sampling techniques such as convenience, purposive (also known as judgmental), and snowball sampling methods, were used to contact respondents. An online survey was preferred due to ease in assembling the data and maintaining anonymity with respondents. This technique not only reduces bias (Llieva et al., 2002; Andrews, Nonnecke, & Preece, 2003) but also helps the researcher to get complete responses as respondents answer all the required questions (Andrews et al., 2003). This helps to get complete responses. Further, an online survey saves responses into a data file directly, thereby, reducing the burden of inputting the data and emitting transcription errors (Evans & Mathur, 2005). The questionnaire was composed of two parts. The initial part comprised of demographic details of the respondents followed by second part which comprised of key constructs.

For sample selection, websites of prominent schools located in North India were visited. Those schools who were conducting online classes were considered in this study. An online link covering scale items along with an invitation letter was mailed to school teachers. Simultaneously, an invitation letter as well as link of the questionnaire was also mailed to the Deans/Heads of the schools to circulate the link among those teachers who were conducting online classes.

To control for the social desirability bias, respondents were assured about their response anonymity and motivated to respond sincerely as much as possible (De Leeuw, Hox, & Dillman, 2008; Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Since, an online survey was carried out to collect data, common method bias could emerge due to high correlation among constructs. In order to reduce common method bias, all constructs were subjected to a principal component factor analysis with varimax rotation. The results of unrotated factor analysis revealed six factors with each construct accounting for 57% of variation. Thus, no specific factor was noticeable (Podsakoff et al., 2003) indicating the common method bias is acceptable in data set.

Using the above‐mentioned methodology, a total of 652 responses were received in the survey. Few responses which had odd figures and incomplete were discarded, resulting in 643 usable responses. Kline (2010) suggested that a sample of 200 responses or larger is suitable for a complicated path model.

In the sample, there is a fair inclusion of respondents across gender—34.83% males and 65.17% females, and good representation of each age group, education level, designation, and experience of conducting online classes. Table 1 reports the characteristics of respondents in more details.

Respondents’ characteristics
Category N = 643 N %
Male 224 34.83
Female 419 65.17
Age N %
25–35 190 29.5
36–45 328 51
Above 45 125 19.5
Education N %
Graduate 186 29
Postgraduate 399 61
Others 58 10
Designation N %
Primary class teacher (up to class 5) 154 23.96
Secondary class teacher (from class 5 to 10) 358 55.67
Lecturer (class 11 and 12) 131 20.37
Experience of taking online classes N %
2 weeks 139 21.62
3–6 weeks 284 44.16
More than 6 weeks 220 34.22


The data analysis process was conducted by means of a two‐step analytical approach. In the first phase, a confirmatory factor analysis (CFA) assessed the measurement model including reliability, validity and fit. Secondly, a structural equation model (SEM) estimated the structural model to test the hypotheses.

4.1 Validating the measurement model

A CFA (Amos) was conducted on all the measurement constructs. The results showed that model fit was satisfactory (χ2/df = 4.688, GFI = 0.899, CFI = 0.881, TLI = 0.895, NFI = 0.902, RMSEA = 0.079) as per Byrne (1994) recommendations. Further, factor loadings were used to assess the indictors’ reliability and 0.50 was taken as a minimum threshold for the retention of measurement items (Fornell & Larcker, 1981). As shown in Table 3, all standardized factor loadings were above 0.50 confirming item reliability and factor unidimensionality (Table 2). Further, convergent validity was assessed through item loadings, composite reliability (CR), and average variance extracted (AVE) of each construct. Table 3 shows that AVE and CR for each construct is above the minimum suggested cut‐off level that is, AVE > 0.50 and CR > 0.70, thereby confirming convergent validity (Bagozzi & Yi, 1991). Further, as can be seen in Table 3, the results also indicated satisfactory discriminant validity since all constructs are more strongly correlated with their own items compared to the other constructs’ items (Fornell & Larcker, 1981).

Measurement model
Std. estimate Std. error Critical ratio Average variance extracted Composite reliability
Performance expectancy PE1 0.689
Mean = 4.424 PE2 0.795 0.081 17.771 0.574 0.843
SD = 0.821 PE3 0.806 0.066 17.964
PE4 0.735 0.066 16.606
Effort expectancy EE1 0.733
Mean = 4.096 EE2 0.806 0.061 18.99 0.544 0.826
SD = 0.895 EE3 0.737 0.056 17.502
EE4 0.668 0.055 15.872
Facilitating conditions FC1 0.806
Mean = 4.137 FC2 0.675 0.048 17.154 0.563 0.837
SD = 0.918 FC3 0.756 0.046 19.493
FC4 0.759 0.054 19.567
Social influence SI1 0.664
Mean = 4.344 SI2 0.82 0.077 17.866 0.625 0.868
SD = 0.853 SI3 0.899 0.071 18.946
SI4 0.761 0.063 16.819
Attitude AT1 0.64
Mean = 4.389 AT2 0.721 0.094 14.146 0.516 0.761
SD = 0.768 AT3 0.787 0.095 14.81
Behavioural intention BI1 0.573
Mean = 4.208 BI2 0.855 0.122 14.894 0.613 0.821
SD = 0.880 BI3 0.883 0.129 14.982
Actual use AU1 0.847
Mean = 4.362 AU2 0.814 0.053 18.871 0.549 0.778
SD = 0.881 AU3 0.518 0.05 12.571
Correlations matrix
Performance expectancy Effort expectancy Facilitating conditions Social influence Behavioural intention Attitude Actual use
Performance expectancy .757
Effort expectancy .630**

Facilitating conditions .522**


Social influence .477**



Behavioural intention .441**




Attitude .461**





Actual use .357**






  • Correlation is significant at the .01 level (2‐tailed).

  • The italics and bold values represents the average value extracted (AVE). **

4.2 Structural model

The section examines the structural model. Table 4 also indicates the structural model reporting the theoretical associations between constructs. The results strongly support H1 and H4, H5–H7 and H9–10, but fail to lend support to H2, H3 and H8.

Structural Model
No Hypotheses Std. loading Std. error Critical ratio p Result
H1 Performance expectancy Behavioural intention 0.144 0.093 2.186 .032 Supported
H2 Effort expectancy Behavioural intention −0.351 0.061 −5.32 *** Not‐supported
H3 Facilitating conditions Behavioural intention 0.39 0.048 −6.911 *** Supported
H4 Social influence Behavioural intention 0.21 0.067 3.44 *** Supported
H5 Performance expectancy Attitude 0.145 0.074 1.614 .007 Supported
H6 Effort expectancy Attitude 0.362 0.058 4.277 *** Supported
H7 Facilitating conditions Attitude 0.222 0.04 3.487 *** Supported
H8 Social influence Attitude 0.094 0.046 1.648 .099 Not‐supported
H9 Attitude Behavioural intention 0.456 0.078 7.856 *** Supported
H10 Behavioural intention Actual use 0.52 0.063 9.057 *** Supported
H11 Facilitating conditions Actual use 0.41 0.051 8.116 *** Supported
H12 Attitude Actual use 0.28 0.077 4.638 *** Supported
  • Note: *** Significant at p < 0.001.

The model fit indices reflect a good fit to the data (χ2/df = 4.687, GFI = 0.902, CFI = 0.898, TLI = 0.901, IFI = 0.895, RMSEA = 0.077) as per recommended thresholds of Byrne (1994). Thus, it can be concluded that the model fit summary indicates that the hypothesized structural model achieved an acceptable model fit. Thereafter, the potency of direct, indirect and total effect for each path was assessed considering standardized path coefficients (β). Figure 2 and Table 4 concludes the structural model’s results by reporting standardized loadings, standard error, critical ratio and significance level for the 12 proposed claims. Overall, analysis has provided support for the acceptance of nine proposed claims, and unsupported three proposed claims.


Path relationships

4.3 Path analysis showing direct and indirect effects

The results claimed the following significant positive and negative direct effects: (a) from PE to BI (0.14); (b) from EE to BI (−0.351); (c) from FC to BI (0.39); (d) from SI to BI (0.21), (e) PE to ATT (0.145); (f) EE to ATT (0.362); (g) FC to ATT (0.222); (h) SI to ATT (0.094); (i) attitude to BI (0.456); (j) behavioural intention to actual use of online teaching (0.52); (k) FC to actual use of online teaching (0.41) and (l) ATT to actual use of online teaching (0.28) (Table 4). The study findings build an understanding about the intricated relationships among performance expectancy, effort expectancy, social influence, facilitating conditions, attitude, behavioural intention and the actual use of online teaching during pandemic outbreak.


5.1 Discussion of the results

The adoption of online teaching during pandemic outbreak was a step indeed to exploit the potential of technology for schools by keeping the safety of students and teachers without wasting their academic hours. Majority of schools initiated mandatory virtual classes on a day to day, and thus, teachers are unfailingly sharing their lessons over Skype call, Zoom call or the other virtual class options to stay the training on through the power of digital technology.

Few previous reported studies have validated TAM (Nikou & Economides, 2019) to understand the perception of teachers but the modified UTAUT validating attitude (Dwivedi et al., 2019) has not been explored much in literature. Further, attitude emerged as a significant construct having direct effect in behavioural intention. This finding is vital as it underlines the significance of individual characteristics in adoption of any technology (Dwivedi et al., 2019).

The findings of this research suggest that performance expectancy perceived by school teachers can build positive attitude as well as drive their behavioural intention to adopt online teaching during the pandemic outbreak. This finding is consistent with the previous studies (Dwivedi et al., 2019; Mosunmola et al., 2018). Effort expectancy had negative but significant relationship with behavioural intention (Liao, Shim, & Luo, 2004; Tseng et al., 2019) but a significant and positive relationship with attitude (Mosunmola et al., 2018). A possible of this could be that most of the sample comprised of senior teachers and length of service plays a vital role in shaping the attitude towards the technology (Teo & Noyes, 2014). Additionally, the impact of effort expectancy dilutes with experience (Venkatesh et al., 2003).

Moving further, facilitating conditions has a significant and positive impact on behavioural intentions as well as attitude (Jairak et al., 2009; Tseng et al., 2019; Venkatesh et al., 2003). This finding indicates that in‐house training programmes and proper equipment helps in familiarization of faculty members with novel technologies thereby facilitating their adoption.

Further, positive impact of social influence on behavioural intention was observed but surprisingly, social influence had a negative influence on attitude to adopt online teaching. This finding is consistent with previous research studies (Sumak, Polancic, & Hericko, 2010; Tseng et al., 2019; Mosunmola et al., 2018; Dwivedi et al., 2017). This finding indicates that teaching fraternity is influenced with experiences of peers who are performing some activity on a particular technology. Attitude had a direct effect on behavioural intention (Dwivedi et al., 2017) and actual use of online teaching (Dwivedi et al., 2017; Thomas, Singh, & Gaffar, 2013). The inclusion of attitude significantly enhances the explanatory power of the model as the direct impact of four independent variables was stronger on attitude as compared to BI.

5.2 Implications of the study

The findings of this research provide significant implications to the policy makers of the schools to encourage online teaching at the time of pandemic outbreak. The results of the study indicate that the administrators of schools should improve performance expectancy as it had significant relationship with behavioural intention and attitude to conduct online classes. For performance expectancy, teachers need to be trained about the benefits and usefulness of online teaching. Further, those teachers who do not understand the usefulness of the technology were unable to adopt the technology. Those teachers who are conducting classes online could convince their colleagues to adopt online teaching. A positive incentive could also be linked with those teachers who are promoting as well as motivating their colleagues to adopt online teaching. These teachers can also discuss their experience with those teachers who are reluctant to adopt the technology.

Since facilitating conditions emerge significant, this indicates that infrastructural support is well established in schools to facilitate online teaching, and it can enable behavioural intention as well as actual use. Regular query handling and training sessions for the teachers should be organized by the school administration so that the teachers are able to comprehend the nitty‐gritties of the system. School authorities need to instil a positive feeling among teachers about usefulness of the online teaching during pandemic COVID 19.

Finally, the very strong relationship between intentions and actual use suggests that an “intention‐behaviour gap” is very unlikely in the specific context of online adoption of classes. This may be explicable by the fact that, unlike other highly involving behaviours where the gap is very acute, the actual behaviour under study (i.e., using online teaching during pandemic) is not so difficult to perform and does not require excessive commitment or motivation. Instead, it fits well in very diverse lifestyles especially since most people use e‐learning and e‐teaching now for a variety of purposes.

5.3 Limitations and future directions for research

This study has some limitations. The study draws sample from North Indian schools, future studies may replicate the study in remaining parts of the country as exposure to technology varies in different parts of country. The study excluded moderators validated by UTAUT and UTAUT2 by Venkatesh et al. (2003, 2012). Future studies may validate moderating variables. Since the UTAUT theoretical model is employed to predict the acceptance of all kinds of technology, researchers expect it is conceived too broadly to capture the specifics of the e‐teaching environment. To enhance the applicability of the model and make it universally acceptable, constructs relating to education like Regulators’ support and Project team capability may also be incorporated in future studies. Nevertheless, it is necessary that the section of population who does not favour new technologies should be included in research during the phase of introducing online or blended learning into teaching staff.


Authors are thankful to two reviewers for providing valuable insights thereby improving the quality of manuscript. Funding information is not available as authors did not receive any funding for this research.


Sangeeta: Responsible of model development and data collection. Urvashi Tandon: Responsible for data analysis and write up.

Source link

Most Popular

To Top