COVID-19 and online teaching in higher education: A discrete choice experiment

Abstract

The Sars-Cov-2 coronavirus outbreak significantly impacted Ghana’s educational system, driving schools to close campuses and swiftly deploy online instruction. This study evaluated e-teaching in higher education amidst the Sars-Cov-2 coronavirus by using the University of Ghana as a case study. Specifically, the study investigated university instructors’ preferences for online instructional strategies to enable higher educational institutions to transit smoothly into online teaching and learning. With the help of a face-to-face questionnaire administration, this cross-sectional study used a discrete choice experiment design to capture the responses of 230-course instructors. The analysis of the survey data obtained was possible using the multinomial logit model. Our results revealed that a recorded lecture video had the highest preference among the course instructors, breakdown of teaching content for approximately 30 to 45 minutes, providing online tutorials, and online support/video tutoring from teaching assistants were the important instructional attributes to help higher educational institutions transition into online teaching and learning.

Keywords

Discrete choice experiments COVID-19 high-impact teaching practices online education instructional strategies multinomial logit model

1. Introduction

The Coronavirus Disease 2019 (COVID-19) outbreak began in Wuhan, China, and quickly spread around the world (Zhu et al., 2020). Every continent on the earth has been affected by this lethal infectious disease because of the outbreak’s rapid nature and the virus’s contagious strength. However, measures for minimizing the transmission of this contagious virus have been developed, including completely washing hands with cleaner water for at least twenty seconds, social distancing, ban on public gatherings, and temporary closing of schools, among others. Given the momentum at which the virus spreads, it was important to close all schools, to assist decrease the infection’s negative impact (Sundarasen et al., 2020). More than 100 countries ordered institutions to suspend in-person education by March 18, resulting in roughly half of the world’s entire student population not attending classes in person and not seeing their instructors face-to-face (Naveenkumar et al., 2020). School closures eventually impacted more than 87% of students worldwide (Schleicher, 2020). It also results in zero teaching and learning while instructors are being paid.

Therefore, like Ghana and other nations impacted by this virus, higher educational institutions implemented the online teaching and learning approach known as E-learning. E-learning is the creation and development of a learning experience through the use of computer technologies and systems (Matheos & Cleveland-Innes, 2018). E-learning, on the other hand, refers to the use of electronic media like the internet, CDs, mobile phones, and indeed television to facilitate remote literacy and instruction (Abidova & Guzacheva, 2020). E-learning is a complex process that incorporates users/participants, e-learning portals, content, and technology tools and design. It is distinct from traditional or other forms of education in that it focuses an equal emphasis on education and substantiated literacy (Ankerson & Pable, 2008). Even before COVID-19, the higher education system was in constant flux, with universities attempting to meet the demands, wants, and ambitions of their students. Therefore, information and communication technologies, as well as E-learning systems, are regarded as critical components of university operations, and such institutions are doing everything they can to invest more in online tools. However, blending new E-learning technologies to enhance and help both teaching and learning is one of the most remarkable challenges for universities and colleges. The conception of E-learning has numerous distinctive interpretations due to its complexities.

Previous studies have shown that E-learning has various benefits for students since it is highly focused on the students, it is extremely adaptable, and it may also improve contact among students by giving tools such as video conferences, email, chats, and forums. Likewise, internet know- approaches expedite the dispersion of content to a large number of users at the same time; scholars benefit from e-learning portals in a variety of ways, including more control over the material, greater control over time spent learning, and the opportunity to personalize the process to the student’s needs and learning goals. This may explain more contact with students, and despite some crucial limitations imposed by the existing system, E-learning may improve students’ literacy experiences.

Even though systems and technologies have aided in the creation and extension of educational possibilities, the use of E-learning in advanced education, and scholars’ evaluations of its effectiveness, have piqued the interest of numerous researchers. A study on scholars’ opinion of the perpetration and integration of E-learning portals and platforms using the Technology Acceptance Model (TAM) as a theoretical foundation revealed that all scholars responded that the E-learning module they took was useful and easy to use, stating emphatically that they understood the contents of the study and effortlessly navigated and manoeuvre through documents with ease (Zhelezniakova & Lapteva, 2021). In general, studies on the use of E-learning in higher education show that it is beneficial, effective, and has a positive impact on student performance. The vast majority of academic instructors are of the view that E-learning can revolutionize the educational process by enhancing student coordination and communication, as well as giving versatility and a deeper understanding of lectures (Mailizar et al., 2020). However, the absence of peers’ physical presence can lead to feelings of loneliness, as well as diminished motivation and delayed feedback or support from teachers when students need it throughout the learning process (Harapan et al., 2020). E-learning is further hampered by poor internet quality. This is due to the high cost of internet access in Africa, as well as the poor quality of service (Stork et al., 2014). This will result in pauses and interruptions in the teaching and learning process, and students who are unwilling to study will hide behind the excuse of bad internet access and refuse to participate in the learning process (Bingimlas, 2009).

Nevertheless, these challenges can be addressed with the help of instructors who adjust their tutoring styles to the requirements of their students. According to the findings of research conducted at the start of the epidemic, 66.9% of respondents (students and instructors) thought E-learning was new to them and that they were not prepared for a completely online experience (Chang & Fang, 2020). While most studies focus on students’ positive attitudes toward online learning, similar studies have found that students do not regard online courses as having the same value as classroom courses (Elzainy et al., 2020) and that students prefer hybrid learning (a mix of in-person and online programs) over traditional learning. Bao (2020) is of the view that contingency plans for unforeseen circumstances, division of teaching content into smaller units, adjustment of teaching speed, sufficient support from faculty and teaching assistants, and high-quality participation are the high-impact online education fundamentals. However, she did not quantify these claims. Our study fills this knowledge gap by providing quantitative evidence in this regard. To the best of our knowledge, this is the first research of its kind to explore university instructors’ preferences regarding online teaching strategies as they transition smoothly into E-education amid the COVID-19 pandemic. This article revealed that a contingency plan to deal with an online education platform shutdown in the case of a recorded lecture video, breakdown of teaching content for approximately 45 minutes, providing online tutorials, and online support/video tutoring from teaching assistants were important instructional strategies to help higher educational institutions transit smoothly unto online teaching and learning. The trade-offs course instructors would make regarding current online learning instructional strategies were also highlighted in this study. Instructors were willing to trade-off the contingency plan to deal with online education platform shutdown in the case of recorded lecture audio, online support from teaching assistants in the case of providing consultation, and phase of teaching in the case of offline self-learning.

2. Methods

2.1 Survey

This study was conducted in compliance with the guidelines for the design and execution of discrete choice experiment (DCE) studies (Hauber et al., 2016; Lancsar & Louviere, 2008). A crucial aspect of DCE design is the selection of attributes and their corresponding levels. The number of attributes and levels to be used in a DCE design is critical because it influences survey results. The researcher must include all important features and attribute levels to avoid respondents forming substantial conclusions based on missed qualities or levels. The recognition of the attribute and its levels should take into account the perspective and experiences of the target population, as well as their cultural and linguistic contexts (Mangham et al., 2009). Therefore, in this study, an extensive literature review about online teaching and learning amidst COVID-19 was conducted to identify an initial list of online instructional attributes (Kohn et al., 2010; Astri, 2017; Bao, 2020; Harapan et al., 2020; Zaharah et al., 2020; Mailizar et al., 2020). These attributes were reviewed by 10 higher education specialists (instructors) and a DCE expert through a focus group discussion. Six relevant attributes and their corresponding plausible levels (one of the attributes at three levels and five of the remaining attributes at two levels) were identified as appropriate through the focus group discussion as well as from a policy standpoint to include in our DCE. The six attributes used in this study are a contingency plan for dealing with an online education platform shutdown, breakdown of teaching content, online instructional information, online support from teaching assistants, strengthening of students’ active learning outside of class, and phase of teaching.

The option selection to be utilized in the choice task, which is defined by a range of choices, attributes, and associated levels, is aided by a well-defined experimental design framework. This is a highly effective strategy for altering attributes and their related levels to rigorously test a specific hypothesis of interest (Louviere et al., 2000). The instructional attributes and their corresponding plausible levels resulted in 96 treatment combinations. However, since it will not be possible to present all the 96 experimental combinations to respondents as it will result in information overload, the JMP Pro (Version 16.0) statistical software was used to reduce the 96 experimental combinations to 12 choice sets of size two (alternatives/instructional strategies). These choice sets were further divided into two blocks to minimize participants’ cognitive burden as frequently encountered in DCE studies (Rajasulochana et al., 2016). An example of a choice question utilized in the DCE survey is presented in Table 1.

Table 1
Example of a DCE question used in the DCE survey

Attributes	Instructional strategy A	Instructional strategy B
A contingency plan to deal with online education platform shutdown	Recorded lecture video	Recorded lecture audio
Breakdown teaching content	45 minutes	60 minutes
Online instructional information	Adopt normal speech (as used in traditional in-class teaching)	Slow down speech
Online support from teaching assistants	Providing consultations	Providing online support (online video tutoring, answering questions for students via email, SAKAI, WhatsApp)
Strengthen students’ active learning outside class	Providing home works/exercises	Providing reading assignments
Phase of teaching	Offline self-learning (reading course-specific literature and submitting short papers based on their key materials before class)	Online teaching (discussion section for students to exchange their understanding based on their reading)
Which of the instructional strategies A and B would you be more likely to choose?	$\bigcirc$	$\bigcirc$

Please bear in mind that you are choosing instructional strategies to promote student literacy focus and engagement in the aftermath of the COVID-19 epidemic to make a smooth transition to online learning. Please indicate which option (instructional strategy A or B) you prefer in the table below?

The discrete choice experiment is a statistical technique that is survey-based and used to quantitatively investigate individual stated preferences in a hypothetical setting (occasionally real). Disciplines, where choice designs have been widely used, include education, health economics, environmental valuation, and other disciplines where there are no effective market institutions to observe consumers’ decisions and examine stakeholder preferences (Wang et al., 2017; Schneider, 2021). The administration of DCE is usually done through a questionnaire that contains typically three parts; respondent information (i.e., sociodemographic variables like income, age, education, and gender as well as anything relevant to the analysis of their decisions), an introduction to the DCE task (it enables respondents to understand the choice sets or choice tasks that they are responding to, why they are doing that, and how to do it correctly), and the DCE itself (i.e., choice tasks that usually contain at least two different hypothetical or real scenarios presented to respondents to indicate the preferences). A detailed description of how to select from possible alternatives was presented to study participants before they were asked to choose between options A or B for each choice question as shown in Table 1. The initial-draft survey questionnaire was also tested in a pilot. For DCE studies, a necessary component of the survey questionnaire is Pre-testing. As a result, researchers can practice and assess participants’ understanding of choice sets, attribute descriptions, and attribute levels, as well as their motivations for selecting specific choices during the actual survey (Adamowicz et al., 1998). For purpose of this study, university of Ghana instructors was tasked to decide on several two sets of instructional strategies or alternatives (i.e., choice sets of size two), which as before were generated using experimental design. These several two sets of instructional strategies were considered to be the principal factors influencing the instructor’s choice of instructional strategies thereby making trade-offs. To enable a smooth transition to online learning, the study will uncover crucial instructional features that will aid boost learning attention and involvement of students.

2.2 Sample size and study population

The outbreak of COVID-19 was unexpected, forcing universities to quickly shift all their existing courses online in a matter of days. A fully online course generally requires an elaborate lesson plan design, teaching resources like audio and video content, and technical support teams. However, most faculty members were struggling because they needed more prior preparation, expertise in teaching online, or help from educational technology teams. Therefore, in this article, we selected instructors as study participants based on their observations and experience of online teaching due to the sudden emergence of COVID-19.

The survey questionnaires were administered face-to-face, which solely involved course instructors at the University of Ghana who has engaged in online teaching due to the COVID-19 pandemic. The random sampling approach was used to sample 230 course instructors. In what follows, we mention that the required degree of accuracy of computed probabilities determines the minimum permissible sample size for simple random sampling. Here we assume that $g$ is the true choice proportion of the population under study, $a$ is the allowable deviation as a percentage between $\hat{g}$ and $g$ , $\alpha$ be the level of confidence in the estimations such that $\text{Pr}(|\hat{g}-g|\leqslant ag)\geqslant\alpha$ for any given $n$ . The minimum sample size according to Hensher et al. (2005) is given by

$\displaystyle n\geqslant\frac{q}{ga^{2}}\left[{\Phi^{-1}\left({1-\frac{\alpha}% {2}}\right)}\right]^{2}.$ (1)

Here, $q=1-g$ and $\Phi^{-1}(1-\frac{\alpha}{2})$ represents the inverse cumulative distribution function of a standard normal distribution reported at $(1-\frac{\alpha}{2})$ . We postulate that the probability of the population choosing either one of the alternatives is 50%. Thus, when $g=$ 50% (0.5), $q=$ 1 $-$ 0.5 $=$ 0.5, $a=$ 0.05 and $\Phi^{-1}(1-\frac{\alpha}{2})=$ 1.96. Hence, the estimated sample size desired for the study is given by

$\displaystyle n\geqslant\frac{0.5}{0.5\,\times\,0.05^{2}}\left[{1.96}\right]^{2}$ $\displaystyle n\geqslant 1536.$

However, to effectively assess the needed minimum number of respondents, we let $N=\frac{n}{k}$ whereas before, $n$ is the minimum sample size and $k=12$ is the number of choices sets available (Hensher et al., 2005). Hence

$\displaystyle N=\frac{1536}{12}=128.$

$N=$ 128 reflects the minimum number of participants required for the DCE study. This indicates that 128 participants were chosen to take part in the research. According to this criterion, a minimum sample size of 128 respondents was required, and our anticipated sample size of 230-course instructors was enough for this DCE study.

2.3 Choice model specification and statistical analysis

Random Utility Theory (RUT) underpins discrete choice experiment modelling. Per the perspective of RUT, the utility $(U_{ci})$ represents a respondent $i$ derived from choosing an alternative in a choice set $c$ which comprises a deterministic component $V_{ci}$ and an unobservable random term $\varepsilon_{ci}$ can be formulated as

$\displaystyle U_{ci}=V_{ci}+\varepsilon_{ci}$ (2)

Here $V_{ci}$ of each alternative is a function of the instructional attributes of the alternative itself and the characteristics of the course instructors. Following McFadden (1978), a utility can be formulated as

$\displaystyle U_{ci}=\alpha+\sum_{q=1}^{Q}{\beta_{q}X_{cqi}}+\sum_{w=1}^{W}{% \gamma_{w}Z_{wi}+\sum_{q=1}^{Q}{\sum_{w=1}^{W}{\delta_{qw}X_{cqi}Z_{wi}+}}}u_{ci}$ (3)

where the choice set $c$ denotes instructional strategy $A$ or $B$ , and $i=$ 1, 2, $\ldots$ , $N$ is the number of respondents. Here, $X$ and $Z$ are vectors of $Q$ attribute levels and $W$ characteristics respectively. The parameter $\delta_{qw}$ measures the utility variation by a particular response attribute while the parameters $B_{q}$ represents the utility associated with the $q^{th}$ attribute. The random component which indicates unobservable effects while deciding on a specific instructional option is denoted by $\mu_{ci}$ . The respondent’s choice of the most useful instructional technique associated with the highest utility is termed “random maximization”. The gain in utility from choosing instructional strategy $A$ over instructional strategy $B$ by an individual $i$ can be formulated as

$\displaystyle U_{Bi}-U_{Ai}=\sum_{q=1}^{Q}{\beta_{q}\left({Y_{\textit{Bqi}}-Y_% {\textit{Aqi}}}\right)}+\sum_{q=1}^{Q}\sum_{w=1}^{W}\delta_{qw}\left({Y_{% \textit{Bqi}}-Y_{\textit{Aqi}}}\right)Z_{wi}+\left({u_{\textit{Bi}}-u_{\textit% {Ai}}}\right)$ (4)

Now, suppose a university instructor chooses instructional strategy $B$ if $U_{\textit{Bi}}-U_{\textit{Ai}}>0$ . Then Eq. (4) can be reformulated with the probability

$\displaystyle P\left[{U_{\textit{Bi}}-U_{\textit{Ai}}>0}\right]=P\left[{\sum_{% q=1}^{Q}{\beta_{q}\left({Y_{\textit{Bqi}}-Y_{\textit{Aqi}}}\right)}+\sum_{q=1}% ^{Q}{\sum_{w=1}^{W}{\delta_{qw}\left({Y_{\textit{Bqi}}-Y_{\textit{Aqi}}}\right% )Z_{wi}+\left({u_{\textit{Bi}}-u_{\textit{Ai}}}\right)}}>0}\right]=P\left[{% \left({u_{\textit{Ai}}-u_{\textit{Bi}}}\right)<\sum_{q=1}^{Q}{\beta_{q}\left({% Y_{\textit{Bqi}}-Y_{\textit{Aqi}}}\right)}+\sum\limits_{q=1}^{Q}{\sum\limits_{% w=1}^{W}{\delta_{qw}\left({Y_{\textit{Bqi}}-Y_{\textit{Aqi}}}\right)Z_{wi}}}}\right]$ (5)

For instance, if we assume a distribution for $Y^{\prime}B_{j}=U_{\textit{Ai}}-U_{\textit{Bi}}$ say, logistic, then Eq. (5) may be formulated in terms of a cumulative logistic distribution:

$\displaystyle P\left[{U_{\textit{Bi}}-U_{\textit{Ai}}>0}\right]=F\left[{\sum_{% k=1}^{Q}{\beta_{q}\left({Y_{\textit{Bqi}}-Y_{\textit{Aqi}}}\right)}+\sum_{q=1}% ^{Q}{\sum_{w=1}^{W}{\delta_{qw}\left({Y_{\textit{Bqi}}-Y_{\textit{Aqi}}}\right% )Z_{wi}}}}\right]$ (6)

$\displaystyle F\left({Y^{\prime}B_{j}}\right)=\frac{e^{Y^{\prime}B_{j}}}{1+e^{% Y^{\prime}B_{j}}}.$ (7)

Similar to the present study, this formula for the multinomial logit is sometimes used to describe models with two alternatives (Train, 2009). The fit of the model to the utility function is

$\displaystyle\text{Pr}(Y=1|X)=\beta_{0}+\beta_{1}\textit{Cplan}_{1}+\beta_{2}% \textit{Cplan}_{2}+\beta_{3}\textit{Bteaching}_{1}+\beta_{4}\textit{Bteaching}% _{2}+\beta_{5}\textit{Bteaching}_{3}+\beta_{6}\textit{Oinfo}_{1}+\beta_{7}% \textit{Oinfo}_{2}+\beta_{8}\textit{Osupport}_{1}+\beta_{9}\textit{Osupport}_{% 2}+\beta_{10}\textit{Alearning}_{1}+\beta_{11}\textit{Alearning}_{2}+\beta_{12% }\textit{Pteaching}_{1}+\beta_{13}\textit{Pteaching}_{2}+\varepsilon$ (8)

$Y$ describes the choice between instructional strategies $A$ or $B$ ; $\beta_{q}=$ Vector of utility coefficients reflecting participants’ preference for each instructional attribute level on average; $\textit{Cplan}=$ Contingency plan; $\textit{Bteaching}=$ Breakdown of teaching content; $\textit{Oinfo}=$ Online instructional information; $\textit{Osupport}=$ Online support; $\textit{Alearning}=$ Strengthen students’ active learning outside class; $\textit{Pteaching}=$ Phase of teaching; $\varepsilon=$ Error term.

The DCE dataset was analyzed using the corresponding model. All attributes were treated as generic or qualitative variables, and effects-type coding (Bech & Gyrd-Hansen, 2005) was used for the attributes. John’s Macintosh Project Pro (JMP Pro) version 16.0 was used for all statistical analyses. Confidence intervals as well as probability values ( $p$ -values) were used to ascertain the significance of the various instructional attributes. Significant positive/negative coefficients indicate a preference for a particular level of an instructional attribute if the p-value is equal to or less than a pre-specified significance level. The relative importance of different levels of the same or completely different instructional attributes was then compared by the utility estimates (preference weights).

3. Results

3.1 Respondents’ characteristics

A total of $N=$ 230 course instructors (age ranges from 18–60+ years) who have online teaching experience were enrolled and completed the DCE survey between April and May 2021. Course instructors’ answers were duly checked for quality because they responded to several instructional scenarios. None were excluded from the analysis. Table 2 shows that most of the participants (34.8%) were from the College of Basic and Applied Sciences (CBAS), which was slightly higher than those from the College of Humanities. In the study population, females were slightly predominant (51.7%). Most of the participants were between the ages of 46 and 59 (46.3%). Of the participants, 81.7% were married, 60.9% had 1–3 children, 82.6% were full-time course instructors and 39.1% were senior lecturers.

Table 2
Demographic characteristics of the study population

Characteristics	Category	Percentage
Gender	Male	48.3
	Female	51.7
Age	18–30	5.6
	31–45	29.6
	46–59	48.3
	60+	16.5
Marital status	Single	14.8
	Married	81.7
	Widowed	3.5
Number of children	None	20.0
	1–3	60.9
	4–6	17.0
	$>$ 6	2.1
Employment status	Full-Time	82.6
	Part-Time	13.9
	Visiting Scholar	3.5
Academic rank	Assistant Lecturer	5.7
	Lecturer	33.5
	Senior Lecturer	39.1
	Associate Professor	3.9
	Professor	17.8
College	CBAS	34.8
	Education	9.6
	Health	21.3
	Humanities	34.4

Table 3

Multinomial logit estimates on online instructional strategy preferences

Attributes	Coefficient	SE	95% confidence interval
A contingency plan to deal with online education platform shut down (Recorded lecture audio)	$-$ 0.258	0.030	$-$ 0.319	$-$ 0.199
Breakdown teaching content (30 minutes)	$-$ 0.001	0.046	$-$ 0.092	0.089
Breakdown teaching content (45 minutes)	0.112	0.044	0.026	0.199
Online instructional information (Adopt normal speech)	$-$ 0.047	0.028	$-$ 0.104	0.008
Online support from teaching assistants (Provide consultation)	$-$ 0.061	0.028	$-$ 0.118	$-$ 0.006
Strengthen students’ active learning outside class (Provide home works/exercises)	0.053	0.028	$-$ 0.002	0.109
The phase of teaching (Offline self-learning)	$-$ 0.067	0.028	$-$ 0.124	$-$ 0.011
Number of observations	2760
AICc	1841.196
BIC	1877.723
$-$ 2*Log-Likelihood	1827.114

3.2 Discrete choice experiment results

A utility coefficient is statistically significant provided that its associated p-value is less than or equal to a 1% or 5% pre-specified significance level. Table 3 shows that the preference weights of the various instructional attributes have the anticipated sign at the 95% confidence interval, but not all are significant. The marginal probabilities (MP) and marginal utilities for each utility coefficient in the model is reported. The marginal probability is the probability that an individual selects online instructional attribute A over online instructional attribute B with all other attributes set to their mean levels. Moreover, the utility coefficients generally show that course instructors highly place importance on a contingency plan to deal with an online education platform shutdown, followed by a phase of teaching, breakdown of teaching content, online support from teaching assistants, strengthening students’ active learning outside class and online instructional information. The multinomial logit estimates revealed that a contingency plan to deal with an online education platform shutdown in the case of a recorded lecture video ( $p<$ 0.0001) had the highest preference weight ( $\beta=$ 0.258; MP $=$ 0.6267) when compared with a recorded lecture audio $(\beta=$ $-$ 0.258; MP $=$ 0.3733). Other strongest preferences were observed for a breakdown of teaching content in the case of 45 minutes ( $\beta=$ 0.112; MP $=$ 0.3716; $p=$ 0.0225), followed by the phase of teaching in the case of providing online tutorials ( $\beta=$ 0.067; MP $=$ 0.5338; $p=$ 0.0181), and online support from teaching assistants in the case of providing online support/video tutoring ( $\beta=$ 0.061; MP $=$ 0.5310; $p=$ 0.0294) compared with their reference levels. Despite the preference for strengthening students’ active learning outside class in the case of providing homework/exercises $(\beta=$ 0.053; MP $=$ 0.5266), it was nearly significant at the 0.05 level ( $p=$ 0.0606). On the other hand, course instructors traded off the contingency plan to deal with online education platform shutdown in the case of recorded lecture audio, online support from teaching assistants in the case of providing consultation, and phase of teaching in the case of offline self-learning.

4. Conclusionss

The purpose of this study was to investigate effective teaching practices for delivering large-scale online education in higher education amid the COVID-19 pandemic; we believe that this is the first detailed study to utilize DCE to investigate university instructors’ preferences for current teaching strategies implemented in response to the COVID-19 pandemic to improve student learning concentration and collaboration. The study included 230 course instructors (lecturers) from the University of Ghana, with 51.74% females and 48.26% males.

The most important deciding factor in successfully delivering large-scale online education was a contingency plan for dealing with an online education platform shutdown in the case of a recorded lecture video. Other studies (Lapitan et al., 2021; Bao, 2020; Nerantzi, 2020; Newton et al., 2014; Smith, 2014) found that the usage of lecture videos has a positive impact on teaching and learning; recorded lecture videos provide flexibility and convenience for students and promote active learning; students can play back components or the entire lecture video and fully comprehend by just watching it whenever it is convenient for them, and they may repeat it anytime some concerns emerge while studying. Prerecorded lecture resources can help students perform better because they are more attentive and responsive in classes that use recorded lectures.

In the case of 45 minutes, other stronger preferences for breakdown teaching content were observed. Similarly, to our findings, Bao (2020) discovered that instructors should fairly split the teaching content into multiple discrete modules or different topics or use a modular teaching method, with each lasting approximately 20–25 minutes to guarantee a clear knowledge structure.

The study also found that phase of teaching in the case of providing online tutorials, and online support from teaching assistants in the case of providing online support/video tutoring were important instructional attributes. Instructors, as well as teaching assistants, should provide an online or video tutoring section for students to enhance their deep understanding of instructional materials. Although preference for strengthening students’ active learning outside class in the case of providing homework/exercises, and online instructional information in the case of slowing down speech, were generally not/less important, instructors should not overlook these attributes in online teaching. Instructors should provide students with exercises and homework/assignments; submit short or summary reports as determined by their reading of instructional content or course-specific literature before the next lecture and exercises. Furthermore, because online teaching restricts body language and facial expressions due to the difficulty of using these tools through screens, Instructors should appropriately slow down their speech or tones to assist students to capture important information throughout their delivery.

The study also highlighted the trade-offs course instructors would make regarding current online learning instructional strategies. Instructors were willing to trade off the contingency plan to deal with online education platform shutdown in the case of recorded lecture audio, online support from teaching assistants in the case of providing consultation, and phase of teaching in the case of offline self-learning. As a result of the high demand for online education in the wake of an epidemic or pandemic, it is recommended to prepare contingency plans to address possible problems, including bandwidth overload, or provide pre-recorded lecture videos instead of lecture audios in advance to address problems, such as traffic overload issues of online education platforms (such as Zoom, Telegram, Google meet, WhatsApp, SAKAI). Instructors and teaching assistants can help students remain on top of their academic work by providing video tutoring via online modules and email support after a normal class schedule. Using the recorded videos could also free instructors from part of their academic work and will allow them to concentrate their efforts on a more effective teaching strategy.

The findings of this study may be used by governments, Higher education policymakers, and stakeholders in higher education to formulate strategies for a smooth transition from the traditional way of tutoring and learning to an online system since COVID-19 has shown us the need for online teaching and learning. This will also improve the enrolment of students in Higher Education since students will have time to work and school simultaneously. Furthermore, because this online teaching was deployed swiftly during the COVID-19 outbreak, students’ anxiety, despair, and tension must be addressed through a variety of methods so that students can participate in online learning actively and successfully.

Funding

For this research, authorship, and/or publishing of this work, the author(s) did not get any financial assistance.

Data/questionnaire availability

The data used to support the findings of this study as well as the questionnaire are available from the corresponding author upon request.

Footnotes

Acknowledgments

The authors would like to express their heartfelt thanks to all individuals who participated in this study: respondents and field enumerators. The authors are also thankful to the anonymous reviewers for their insightful comments.

Conflict of interest

The authors declare that there are no conflicts of interest.

References

Abidova

, & Guzacheva

(2020). Features of Medical Terminology Teaching in English Lessons. Bulletin of Science and Practice, 6, 434-437.

Adamowicz

W. L.

(1998). Introduction to Attribute-Based Stated Choice Methods. Washinton, U.S.A: Resource Valuation Branch Damage Assessment Center, NOAA, U.S Department.

Ankerson

K. S.

, & Pable

(2008). Interior design: Practical strategies for teaching and learning. New York, NY: Fairchild Books Inc.

Astri

L. Y.

(2017). Barrier factors that influence the satisfaction of e-learning: a literature study. Advanced Science Letters, 23, 3767-3771.

Bao

(2020). COVID-19 and online teaching in higher education: A case study of Peking University. Human Behavior and Emerging Technologies, 2, 113-115.

Bech

, & Gyrd-Hansen

(2005). Effects coding in discrete choice experiments. Health economics, 14, 1079-1083.

Bingimlas

K. A.

(2009). Barriers to the successful integration of ICT in teaching and learning environments: A review of the literature. Eurasia Journal of Mathematics, Science and Technology Education, 5, 235-245.

Chang,

C. L.

, & Fang,

(2020). E-Learning and online instructions for higher education during the 2019 novel coronavirus diseases (COVID-19) epidemic. Journal of Physics: Conference Series, 1574, 012166.

Elzainy

El Sadik

, & Al Abdulmonem

(2020). Experience of e-learning and online assessment during the COVID-19 pandemic at the College of Medicine, Qassim University. Journal of Taibah University Medical Sciences, 15, 456-462.

10.

Hauber

A. B.

Gonzalez

J. M.

Groothuis-Oudshoorn

C. G.

Prior

Marshall

D. A.

Cunningham

et al. (2016). Statistical Methods for the Analysis of Discrete Choice Experiments: A Report of the ISPOR Conjoint Analysis Good Research Practices Task Force. Value Health, 19, 300-315.

11.

Harapan

Itoh

Yufika

Winardi

Keam

, & Mudatsir

(2020). Coronavirus disease 2019 (COVID-19): A literature review. Journal of infection and public health, 13, 667-673.

12.

Hensher

D. A.

Rose

J. M.

, & Greene

W. H.

(2005). Applied Choice Analysis, A Primer. Cambridge: Cambridge University Press.

13.

Kohn

Maier

, & Thalmann

(2010). Knowledge Transfer with E-Learning Resources to Developing Countries: Barriers and Adaptive Solutions, in Breitner

M.H.

Lehner

Staff

& Winand

(Eds.), E-Learning Springer.

14.

Lancsar

, & Louviere

(2008). Conducting discrete choice experiments to inform healthcare decision making: a user’s guide. Pharmacoeconomics, 26, 661-677.

15.

Lapitan

L. D.

, Jr Tiangco

C. E.

Sumalinog

D. A. G.

Sabarillo

N. S.

, & Diaz

J. M.

(2021). An effective blended online teaching and learning strategy during the COVID-19 pandemic. Education for Chemical Engineers, 35, 116-131.

16.

Mailizar

Abdulsalam

, & Suci

(2020). Secondary schoolmathematics teachers’ views on e-learning implementation barriers during theCOVID-19 pandemic: The case of Indonesia. Eurasia Journal of Mathematics, Science & Technology Education, 16, 1-9.

17.

Mangham

L. J.

Hanson

, & McPake

(2009). How to do (or not to do). Designing a discrete choice experiment for application in a low-income country. Health policy and planning, 24, 151-158.

18.

Matheos

, & Cleveland-Innes

(2018). Blended learning: Enabling higher education reform. Revista Eletrônica de Educação, 12, 238-244.

19.

McFadden

(1978). Modeling the choice of residential location. In: Karlquist

Lundquist

. Snickars

, & Weibull

J. W.

, eds. Spatial interaction theory and planning models. Amsterdam: North-Holland Publishing Company.

20.

Naveenkumar

Nag

B. S.

Vijayaraghavan

, & Porteen

(2020). The Possible Risk of Reverse Zoonosis in COVID-19: An Epidemiological Driving Approach for the One Health Future Challenges: A Review. Asian Journal of Dairy and Food Research, 39, 173-179.

21.

Newton

Tucker

Dawson

, & Currie

(2014). Use of lecture capture in higher education-lessons from the trenches. TechTrends, 58, 32-45.

22.

Nerantzi

(2020). The use of peer instruction and flipped learning to support flexible blended learning during and after the COVID-19 Pandemic. International Journal of Management and Applied Research, 7, 184-195.

23.

Rajasulochana

Nyarko

Dash

, & Muraleedharan

V. R.

(2016). Expectant Mother’s preferences for Services in Public Hospitals of Tamil Nadu, India. Journal of Health Management, 18, 305-317.

24.

Schneider

(2021). Are there never too many choice options? The effect of increasing the number of choice options on learning with digital media. Human Behavior and Emerging Technologies, 3, 759-775.

25.

Schleicher

(2020). The impact of COVID-19 on education: Insights from education at a glance. OECD Publishing. https://www.oecd.org/education/education-at-a-glance/.

26.

Smith

D. K.

(2014). iTube, YouTube, WeTube: Social media videos in chemistry education and outreach. Journal of Chemical Education, 91, 1594-1599.

27.

Stork

Calandro

, & Gamage

(2014). The future of broadband in Africa. Info, 16, 76-93.

28.

Sundarasen

Chinna

Kamaludin

Nurunnabi

Baloch

G. M.

Khoshaim

H. B.

, & Sukayt

(2020). Psychological impact of COVID-19 and lockdown among university students in Malaysia: implications and policy recommendations. International Journal of Environmental Research and Public Health, 17, 6206-6206.

29.

Train

K. E.

(2009). Discrete choice methods with simulation. New York: Cambridge university press.

30.

Wang

Chen

Ratcliffe

Afzali

H. H. A.

Giles

, & Marshall

(2017). Adolescent values for immunization programs in Australia: A discrete choice experiment. PLoS ONE, 12, e0181073.

31.

Zaharah

Kirilova

G. I.

, & Windarti

(2020). Impact of corona virus outbreak towards teaching and learning activities in Indonesia. SALAM: Jurnal Sosial dan Budaya Syari, 7, 269-282.

32.

Zhelezniakova

E. A.

, & Lapteva

M. L.

(2021). Distance Learning Technologies in Teaching Russian as a Non-Native Language in Primary School. Russian Language Studies, 19, 21-33.

33.

Zhu

Wei

, & Niu

(2020). The novel coronavirus outbreak in Wuhan, China. sGlobal health research and policy, 5, 1-3.

COVID-19 and online teaching in higher education: A discrete choice experiment

Abstract

Keywords

1. Introduction

2. Methods

2.1 Survey

Table 1 Example of a DCE question used in the DCE survey

3.1 Respondents’ characteristics

Table 2 Demographic characteristics of the study population

4. Conclusionss

Funding

Data/questionnaire availability

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Example of a DCE question used in the DCE survey

Table 2
Demographic characteristics of the study population