The Challenges Faced in Technology-Driven Classes During COVID-19

In the wake of coronavirus pandemic, social distancing became a mandate that led to the transition from traditional classroom-based lectures to computer-based learning. This paper extensively deals with the ranking of the challenges faced by instructors and students. Primary data from 624 participants (399 students and 225 instructors) is collected through a questionnaire. To assign the ranking to the challenges, Technique of Order Preference Similarity to Ideal Solution (TOPSIS) is deployed. A contextual model is developed by using Interpretive Structural Model (ISM) technique that further provides recommendations for prioritizing the challenges that need to be addressed to mitigate the problems faced in online lectures in coronavirus situation. The number of variables is reduced to simplify the interpretation by exploratory factor analysis. The study also provides the basis to formulate the strategies for policymakers and administration after identifying which challenges need to be addressed first for mitigating all the other challenges.


INTRODUCTION
Education can be imparted through multiple platforms such as traditional face to face delivery, E-education and mixed methods.Today, universities are banking on online media completely for their distance learning programs that are completely executed through computers, laptops or mobile phones on platforms like Zoom, Google Meet, Cisco Webex etc.However, in traditional classroom which involves face to face delivery, the instructors sometimes would switch to online resources to illustrate the concepts graphically.However, this coronavirus pandemic propelled the instructors to switchover completely to the online mode of teaching as social distancing became a norm.
In India, the decision for lockdown which started from March 25 th 2020 was sudden.As a result of coronavirus pandemic, the academic calendar could not be followed as planned and students and teachers were both left in uncertainties and tribulations for the future.The course content was not completed and it required immediate measures to complete the content and continue the classes amidst lockdown.In this scenario, online classes were considered as the best possible alternative for the traditional classes as many of the past studies by Bernard et al. (2004), Jahng et al. (2007) have concluded that learning outcomes of traditional and online classes are similar.This platform was also the most feasible to launch, keeping in mind the constraints of social distancing presented by COVID-19 situation, while following the same time schedule as of traditional classes as the time to prepare and gather resources for online classes was negligible for both students and instructors.The students and instructors were present in different geographical locations, which also led to difficulties in conducting traditional classes.All these aspects contributed to a decision of organizing online classes.
Online classes encouraged students to attend the lectures as the content in video lectures incorporates interesting features such as the integration of audio and visual media that make the lectures interactive.Although online classes provided a lucrative alternative for traditional classes, they were ridden with many challenges.The immediate problems were lack of access to resources and infrastructure to conduct online classes, increased workload at home, mental stress and uncertainties due to Corona pandemic.Other problems were related to inadequate training and responsiveness of administrative authorities which limited the use of online platforms.Further, the lack of response by authorities led to anxiety in students especially who were about to receive their degree or certification.Some of the students and faculty members did not have access to the internet because of remote locations and incessant power failures.Motivation level of students was at a low ebb due to disconnect from friends, instructors and forced departure to their homes.This presented a unique challenge for the instructors as they had to improve the motivation levels of students as in the past studies by Deimann and Bastiaens (2010), high level of motivation in students is needed in order to make online classes successful.
It is necessary to understand these challenges as it would allow the government, administration and policy makers to provide solutions and facilities to instructors and students to make online classes successful and maximize the learning outcome.This paper aims to present a ranking for the challenges to the technology aided classes for students and instructors by using the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method.As the number of challenges identified are large, Exploratory Factor Analysis (EFA) is used in order to combine some of the challenges under latent variables facilitating easy interpretations and analysis.The paper also presents recommendations on how to mitigate the challenges by using an Interpretive Structural Model (ISM) based approach.The recommendations were further validated by students and instructors in order to select most effective one.
The interpretations of the study will allow to understand the various challenges faced during online classes from the perspective of students and instructors and enable effective strategy formulation in order to maximize the learning outcome.This can be accomplished by understanding the ranking of the challenges which identifies the impact created by the individual challenges.Further, by analyzing the contextual relationships between the challenges, the challenges having a higher influence or driving ability can be identified.The stakeholders involved in pedagogy can make efforts to mitigate the challenges which have a higher impact and have a greater power to drive other challenges.

Traditional vs Online Classes
Online classrooms are inevitably blended with the education system and are considered as an alternative to traditional classrooms.The past studies by Picciano and Seaman (2009), Kim et al. (2011), andNacu et al. (2016) have supported the claim that the number of students enrolled in online classes is continuously increasing.Online classrooms gain leverage over traditional classrooms by providing several benefits such as flexibility over time, convenience, path, the pace of learning, and cost-effectiveness (Carnevale, 2000).The content in online classes is delivered through audio as well as visual media that allows the use of various forms of videos, animations, illustrations in online classrooms (Mayer, 2014), which positively influences student engagement (Junco et al., 2011).The combined use of media elements provides enhanced processing, visualization, and delivery of content and concepts, effective use of which can improve the attention level of students (Costley & Lange, 2020).However, there is a lower retention rate associated with online classes (Stover, 2005) which is attributed to the multiple features offered by online platforms such as screen sharing, audio settings, video display, aspect ratio of the screen etc. which can overwhelm the learners leading to distractions.
Past studies on online classes are focused more on the factors that are in tune with traditional lecture-based instructional models, identifying the reasons for lower grades and absence in the classes (Ho et al., 2014).In other studies, the attention is on technology adoption, continuation of use of technology, and learning outcomes (Panigrahi et al., 2018).Along with the use of technology the importance of feedback on students' performance has not been reduced.Feedback allows students to rectify their mistakes for a better understanding of the concepts and improve the learning outcomes.It also allows students to communicate effectively with peers and instructors encouraging social interaction.The role of feedback by the instructor is also considered highly important in the study by Gašević et al. (2015).The opportunities for receiving immediate feedback from the instructor is less in online classes (Faux & Black-Hughes, 2000).
Due to the deluge of material delivered at a fast pace, student tend to drop the courses taken online (Frankola, 2001;Oblender, 2002).As online classes have different nature compared to traditional, many students who are successful in traditional classroom model in terms of receiving higher grades, improved feedback, better retention etc. are not comfortable in online classes (Cheung & Kan, 2002) and it is a little tricky to understand the factors that can act as a predictor for success in online classes.This makes students more comfortable in attending the classes in the traditional format.Traditional classes allow the students to clarify their doubts immediately, however, in online classes due to lack of physical presence the assessment of non-verbal cues is missed, leading to postponement of doubt clarification.Further, the responsibility of self-assessment and autonomy fall on student's shoulder to expedite and improve the learning abilities.

Instructor's Competences in Online Classes
In order to make online classes successful, the role of instructor cannot be ignored.The instructor plays a key role in the delivery of online lectures, which involves balancing the pace, quality, and context of the lecture (Lange & Costley, 2020).According to their study, the various media factors to be taken care of are: pace of lectures, intelligibility of the learner, quality of classes, use of diverse forms of media, and congruency between lectures.Instructors can significantly influence the learning outcomes, interest, and involvement of students.The instructor should also aim to provide an experience in a class which student would not want to miss (Bender, 2003).The role of the instructor in online classes is considered as that of a planner providing clear instructions, model acting as an inspirational figure, coach encouraging students to learn, facilitator acting as a guide, and communicator to provide feedback and deliver lectures effectively (Heuer & King, 2004).Similarly, study by Hung and Chou (2015) has also described the role of instructor as course designer and organizer, discussion facilitator, social supporter, technological facilitator, and assessment designer.Shea and Bidjerano (2009) concluded that teacher's involvement in a class is an important factor in developing cognitive and social presence of students.A study by Asarta and Schmidt (2013) has identified that method of presentation is more important than content in online classes.Therefore, instructors should put an effort in making the online classes interactive and improve on the method rather than the content.
Several models about digital competences required by instructors to improve learning from online classes have been proposed.These models provide a framework of how digital technologies can be used to improve education and training.They also help to align the instructor's profile to the digital competencies required for online and distance learning programs.One of the popular frameworks is the European Framework for the Digital Competence of Educators (DigCompEdu), which organizes 22 competences in six areas which include professional engagement, digital resources, teaching and learning, assessment, empowering learners, and facilitating learner's digital competence.This model is directed to instructors at all levels of education including general and vocational learning.

Challenges Faced in Online Classes
From the viewpoint of students, the biggest challenge is lack of motivation and procrastination in attending the online classes.This is because of lack of push factor in online classes, which implies that students should be highly motivated to achieve the desired learning outcomes.Self-regulation is also required as there is a high chance to cheat in online classes (Lanier, 2006) and there are a limited number of opportunities for assessment through an online medium (Williams et al., 2012).Students should be able to self-regulate themselves as participation in online classes requires initiation by students themselves for clarifying the doubts and understanding the profound concepts for maximizing the learning output.Even in the past studies, the learning outcomes of the online classes are highly dependent on self-regulation (You & Kang, 2014;Henrie et al., 2015).Online classes require students to work and gather resources on their own to gain benefit from online classes.This is validated in the study of Diaz (2002) and Bell and Akroyd (2006) who concluded that students who perceive their academic accomplishments as their own work are more likely to succeed in online classes.The massive pre-requisites such as previous courses, academic or GPA cutoff etc. also make attending online classes challenging for the students.Wojciechowski and Palmer (2005) have identified several factors as predictors for success in online classes, including student's academic capabilities, motivation, and gender.Particularly the students having a higher grade point average (GPA) and academic acumen are more likely to succeed in online classes (Dupin-Bryant, 2004).Some of the studies also place demographic variables such as age (Murtaugh et al., 1999), gender (Willging & Johnson, 2009) and ethnicity (Stratton et al., 2007) for success in online classes.However, there are several variations obtained in the results of these studies, some conclude that older students don't perform well (Park & Choi, 2009) while others (Wojciechowski & Palmer, 2005) conclude they do.In some studies, better performance is by men (Kramarae, 2001) and in others by women (Willging & Johnson, 2009).However, there are variations in the predictors also as Lin et al. (2017) have identified that online learning outcomes from the class are not predicted by intrinsic or extrinsic motivation but the learning strategy, which includes pedagogy and learning ability.
There are innumerable hindrances in conducting online classes.Maguire (2005) reviewed thirteen studies and concluded that instructors faced numerous barriers.Muilenberg and Berge (2005) have identified several barriers to students' learning in online classes and classified them through factor analysis into eight categories-administrative issues, social interaction, academic skills, technical skills, learner motivation, time and support for studies, cost and access to the Internet, and technical problems.Haber and Mills (2008) have examined the barriers associated with the effectiveness of online classes with respect to Florida Community college faculty and concluded that several instructors face a lot of problems while conducting online classes.Gillette-Swan (2017) has identified several other problems such as inhibition in using technology, difficulty in peer interaction, and anxiety being out of comfort zone.Due to the involvement of several features and different interface in the platform, the use of various commands, procedures, and methods to execute the online classes make it necessary to train the participants to use the platform.Studies by Schmidt et al. (2016) and Rucker and Downey (2016) identified this challenge and also gave the specifications of how the training can be conducted.

Research gap
Although, there is an abundance of the past studies in which reference to certain challenges in executing the online classes is mentioned, less is known about the ranking of these challenges and how they are related to each other.The difference in perceived ranking of the major stakeholders in online classes-instructor and student has also not been discussed in the previous studies.This aspect is necessary to understand in order to provide effective solutions by understanding which challenge presents a bigger problem to the instructors and students so that efforts can be made to mitigate it.Another necessary step is to identify which challenge leads others, it will allow to analyze their perceived impact and identify which one to address first to generate maximum benefits.The spread of COVID-19 has also presented a number of challenges as it significantly affects the mental health creating problems such as anxiety, depression, stress, uncertainties, etc. (Galea et al., 2020).These issues should also be incorporated while analyzing the challenges faced in online classes.All of these facts present a need for a study that specifically deals with the challenges and provides recommendations for mitigating them.

METHOD
The challenges faced by the students and instructors in switching over to traditional to online platform during COVID-19 were determined by consulting them through semi-structured personal interviews.The participants were asked about the challenges faced during the online classes, who were contacted by using stratified random sampling from the list of total students and instructors in the universities.The universities chosen for the study were from India and were distributed evenly in the NIRF rankings provided by Government of India.Rigorous interview of 40 minutes was conducted in which participants were asked to elaborate the challenges they faced during the online classes and determined contextual relationships of the challenges i.e. which challenge influenced others or by which it was driven.The interviews were discontinued when the challenges observed began to get repeated.In total, the sample for this step consisted of 50 students and 25 instructors.Table 1 shows the list of challenges identified by taking the union of challenges reported in the interview.

Measures
In order to identify the ranking, the challenges faced in online classes were considered as present in Table 1.A questionnaire consisting of two sections was developed in which participants had to rate the impact of each challenge faced during the online classes on a Likert scale of 1-5; 1 having the least impact and 5 having the most.In order to gauge the views of participants on the ways of mitigating the challenges in Table 1, the recommendations were mentioned in the questionnaire to be filled by the respondents.Some of the recommendations were identified while conducting the interviews, others were identified through an open-ended question.The additional responses obtained through this field was included in the analysis.

Procedure
After identifying the challenges, a questionnaire was developed and tested on a pilot sample for its reliability by using Cronbach alpha criteria yielding a value of 0.869, which is more than the recommended value of 0.7 (Taber, 2018).To measure the content validity, the method prescribed by Lynn (1986) and Waltz and Bausell (2005) is used where a group of four experts was to rate two item sections present in the questionnaire-impact of challenges and identifying recommendations, on the scale of 1-4.The following ratings were mentioned for the purpose of testing the validity of the sections: 1-not relevant, 2-somewhat relevant, 3-quite relevant, 4-highly relevant.Each expert rated the questionnaire section in the relevant zone (3 and 4).These results indicate that the questionnaire is valid as well as reliable, therefore it can be used in the main study where participants were asked to fill the questionnaire in a quiet environment in order to minimize the distractions.
For the study, the following data analysis methods are used: The ISM technique is used to determine structural relationships amongst the variables and provides an explanation of how the variables are related to each other.ISM has been used in researches to transform poorly developed models to well defined models (Sage, 1977).This technique provides a structure to the variables by presenting them in a graphical model with upper level variables having dependence on lower level variables.For the study, after the data collection, ISM and MICMAC analysis were used in order to analyze the contextual relationship between challenges (Sage, 1977) and to provide realistic picture of variables involved (Attri et al., 2013).
• Exploratory Factor Analysis (EFA) When a large number of variables are encountered in a research design, it is necessary to identify the underlying constructs.This identification allows to investigate few variables combined in a form of a latent construct compared to the original variables.EFA is used in the study by using principal component analysis to make interpretations easier to understand by combining several challenges under latent variables.These latent variables are classified in accordance to the composition of the challenges within these latent parameters.
• Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) TOPSIS is one of the most used multi-criteria decision making technique i.e. identifying the best alternative within the possible constraints.TOPSIS has been considered as a simple, intuitive and logical way of ordering objects.The classical TOPSIS method aims to assign priority to the challenges and contributes to research by providing a method to define the ranking of challenges.TOPSIS method was proposed by Huang and Yun (1981), and it has been adopted in several past studies in order to rank the challenges or barriers (Ahmadi et al., 2013;Wanke et al., 2016;Singh et al., 2017).This method is suitable in the current study because of the involvement of quantitative data.TOPSIS is a seven-step process: Step 1: Establish a decision matrix consisting of the alternatives and criteria on which alternative are evaluated.In this study, the criteria are the ratings on the Likert scale (from 1 to 5) and the alternatives are the 25 challenges listed in Table 1.
Step 2: Calculate the normalized decision matrix.
Step 3: Calculate the weighted normalized decision matrix i.e. multiply the weights associated to each alternative in the normalized matrix.
Step 4: Determine positive and negative ideal solution.
Step 5: Calculate the separation distance i.e. deviation of each alternative from ideal positive and negative solution.
Step 6: Calculate the Closeness Coefficient i.e. relative closeness of each alternative to ideal solution.
Step 7: Rank the alternatives on basis of value of Closeness Coefficient (CC), the alternative having CC close to 1 is ranked higher.

Participants
For the pilot study, the number of participants was 25 (15 students; 10 instructors).In the main study, a total of 624 participants were present.The number of students participated was 399(Age:15 to 28 years; Gender:270 males, 129 females) and number of instructors was 225(Age:26 to 58 years; Gender:60 males, 165 females).The participants were chosen randomly from 14 different academic institutes in the same manner as mentioned in Method section.The participants were from different departments such as Humanities, Economics, Management, Social Sciences, Computer science, Electrical and Mechanical engineering, etc. pursuing/teaching different degree programs such as PhD, BTech, BA, MA, BBA, MBA, MBBS, Secondary School Education, etc.The participants had already completed a part of the registered courses before the lockdown in a traditional face to face classroom and completed the rest of the course through online mode with the same time schedule, instructor and course content.

RESULTS
Before the exploratory factor analysis, outliers were removed and KMO test was conducted.KMO test confirms whether the sample collected for the study was adequate to run factor analysis or not.
The minimum KMO index for the individual items was 0.87 for directly connecting with teacher.The overall KMO index obtained was 0.93 which is classified as good (Hutcheson & Sofroniou, 1999) and is more than the sufficient value of 0.5.Therefore, factor analysis could be used to group the challenges under latent variables.The number of latent variables was chosen on the basis of eigenvalue criteria.Using principal component analysis with varimax rotation, the number of variables having an eigenvalue greater than 1 was chosen.The number of latent variables that satisfies this criterion was five.The following Figure 1 depicts the grouping of challenges under latent variables and results of exploratory factor analysis.
The results are valid as Root Mean Square Error of Approximation (RMSEA) value is 0.046 which is close to 0 (MacCallum et al., 1996), Tucker-Lewis Index is 0.952, which is greater than or equal to 0.95 as specified by Hu and Bentler (1999).The latent variables observed were reported in the Table 2.
To generate the interpretive structural model, contextual relationships were identified among the challenges as described in the method section.The following Figure 2 shows the structural selfinteraction matrix, where the symbols mean: V-Challenge i will lead to challenge j A-Challenge j will lead to challenge i X-Challenge i and j will lead each other O-Challenge i and j are not linked i and j refer to serial numbers of challenges The final reachability matrix, shown in Figure 3 is computed by using the following rules: If V is present in any cell (i,j)-the value in cell (i,j) is 1 and the value in cell (j,i) is 0. If A is present in any cell (i,j)-the value in cell (i,j) is 0 and the value in cell (j,i) is 1.If X is present in any cell (i,j)-the value in cell (i,j) is 1 and the value in cell (j,i) is 1.If O is present in any cell (i,j)-the value in cell (i,j) is 0 and the value in cell (j,i) is 0.  For diagonal entries i.e. for cell where i=j, value is 1.
After analyzing the antecedent and precedent set from the matrix represented in Figure 3, the following map (shown in Figure 4) for the influence (marked on y axis) which provides a measure of driving power of challenge and dependence (marked on x axis) which provides an estimate of dependence on other challenges is obtained by calculating the sum of individual rows and columns respectively.The serial number of challenges as present in Table 1 is marked on the map by plotting the values obtained for influence and dependence as mentioned in Figure 3.The results indicate a good fit as the challenges are aligned equally with the dependence and influence axis.From this step, the challenges are classified in six levels by combining the challenges present as clusters.The final model is depicted in Figure 5.To generate the ranking of the challenges in three categories-students, instructor and overall, TOPSIS technique was used, consisting of seven methods as prescribed by Huang and Yun (1981).The initial decision matrix consists of 25 identified challenges (as rows) and number of people (as columns) who rated the particular scale item on 1-5 on Likert scale.The following Figure 6 shows the initial decision matrix obtained by collecting the number of responses corresponding to a particular rating mentioned on the Likert scale in the questionnaire.A total of three tables are constructed considering responses from students, instructors, and overall category consisting of both students and instructors.
TOPSIS requires input weights for each criterion, which in this case are responses 1 to 5 obtained through section 1 of the questionnaire.The weights for all of the three cases were generated by using entropy method as used in past studies such as Zou et al. (2006).To generate the weights, the matrix in Figure 6 was first normalized, i.e. dividing each entry of matrix with the sum of particular column as prescribed in the study of Zou et al. (2006).Then the following equations were used: where Rij is the entry in the matrix and i, j ranges from the first entry of the table to the last entry.K is the natural logarithm of normalized entry of the particular matrix element being considered as per the values of i and j: For the overall case, the weights for response observed were 1-0.20795, 2-0.12165, 3-0.08014, 4-0.16186, 5-0.42838; for students-1-0.25221,2-0.13783, 3-0.12995, 4-0.12656, 5-0.35343; for faculty-1-0.08786,2-0.08638, 3-0.13554, 4-0.26301, 5-0.42718.The sum of weights in each of the categories is 1.Although the weights are in similar ranges for all three categories; slight difference is observed due to normalization of the variation in data points.
In step 2 of TOPSIS, table in Figure 6 is normalized by dividing the square root of the sum of squares in each column for three categories.The normalized matrix is then multiplied with the associated weights.For example, in the overall category data, for item scale 1, 0.20795 is multiplied with each entry in that column.The following Figure 7 show the results after multiplying the weights.After computing weighted normalized matrix, the maximum and minimum ideal solutions are identified and distance from each of these alternatives is calculated by the following equations: The normalized distances d+ and d-are calculated by taking the square root of the sum of squares of these values by using: Finally, the closeness coefficient (CC) is calculated by using the value d-/(d+ + d-).When the closeness coefficient (CC) is close to 1, it is more comparable to the ideal solution and is ranked higher.The following Tables 3, 4 and, 5 depict the ranking of the challenges in different categories.
Figure 8 shows the closeness coefficient for the challenges.The value of closeness coefficient is represented on the x-axis and challenges are presented on the y-axis.The values are shown for three categories-students, instructors, and overall.The difference in perception of the challenges can be easily gauged by comparing the observed value of closeness coefficient among the categories.

DISCUSSION
As the first latent variable in the EFA model consists of the challenges related to disturbance in execution, it is named as Executional barriers.The second latent variable comprises of the challenges which have an effect on social contact, hence called Hiatus between people.The third latent variable is named Infrastructure Limitations, as it is related to problems associated with infrastructure.The fourth latent variable comprises of technology and platform-based challenges, therefore are listed as Technological Inhibitions.The fifth latent variable constituted various challenges relating to State of mind, hence named so.
From Figure 4, the challenges which have a high influence and low dependence (present in the top left corner) are called driver variables which constitute of mental stress due to Corona, uncertainties leading to loss in motivation, increased workload due to home activities, lack of infrastructure, power failure, time lag in speaking, use of mobiles during classes, lack of interaction with peers, lack of internet access, inhibition in using technology, insufficient training and lack of technical skills.If we compare these driver variables with EFA model, these challenges form a part of the latent variables constituting of technological inhibition and infrastructure barrier.
The challenges which have a low influence and high dependence (present in the lower right corner of Figure 4) are termed as dependent variables and they comprise of lack of instruction, interruptions in online classrooms, lack of understanding of platform, lack of access to resources, distractions/ disturbances in voice presentation, less confidence for online classes, lower quality of material delivered online, procrastination, lack of immediate feedback, directly connecting with teacher, delay in timing of lecture, teachers ignoring doubts and lack of personal motivation.If we compare these dependent variables with EFA model, these challenges form a part of the latent variables constituting of hiatus between people and state of mind.
The variables having low dependence and low influence are called autonomous variables and those having high dependence and high influence are linkage variables.The model doesn't contain any of these variables which is evident from Figure 4 as there are no entries in the lower left and upper right corner.From these details, a contextual model is obtained by combining specific clusters.The six clusters identified are arranged in six levels, lower levels formed by driver variables and upper levels formed by dependent variables.
In Figure 5; the challenges in the lower levels of the model drive the challenges in the upper level as these challenges have more influence and less dependence.The challenges present in Level 1 have cumulative effect on all the upper levels i.e.Level 2, 3, 4, 5, and 6, similarly, all the lower levels challenges have an effect on upper level challenges.Hence, while addressing the challenges, priority should be given to the lower levels including Level 1 that consist of the challenges-Mental stress due to corona, Uncertainties leading to loss in motivation, Increased workload due to home Hence, policymakers should focus on challenges located in lowest level first, to effectively abate all the challenges.The upper levels of the model can be dealt at a later stage.In all, the first three levels can be considered as drivers and last three levels can be considered as driven challenges and if the challenges present in driver levels are effectively countered the problems can be resolved quickly.
It was predicted that the recommendations which affect the lower level would be perceived as the most useful in dealing with the challenges, which is observed in the recommendation section of Regarding the ranking of the challenges, the highest overall challenge is presented by lack of internet access, increased workload and the least is by delays in lecture and lack of information.For students, the highest perceived challenge was lack of internet access and lack of personal motivation, whereas for instructor it was direct connect with teachers and interaction with peers.The lowestranked challenges for students were inhibition using technology and lack of instruction as students are more comfortable using technology as compared to instructors.Further lack of instructions can be easily countered by contacting the resource person or peers; for the instructor it was less confidence and lack of instructions as instructors are taking the course for a long time and don't feel much of a difference when the medium of delivery changes.Major differences are explained further.Direct connect with teacher and peers are the challenges which affect the instructor more than the students, because instructors are habituated to deliver the classes in traditional format, interacting with students and peers in order to make the course content better, online classes limit this interaction.Instructors are heavily conscious about the quality of lecture delivered in online classes and they are not ready to compromise with the quality of material delivered whereas most students are not familiar with the subject at the time it is taught, hence lower quality of material is not much of a challenge for them.Hence, lower quality content is perceived as a bigger challenge by the instructor than students.When everyone is at home, there is an increased workload of managing lectures as well as doing daily chores of home, this presents a big challenge to the instructor as they have to manage home chores along with professional tasks.Instructors do not use technology frequently, hence explaining the inhibition to use technology, insufficient training, and lack of understanding of operations is perceived as a greater challenge by the instructor.Lack of resources and instructions affect students more, as instructors have access to resource material and have good command over the subjects they teach and don't need resources as much as students need.Mental stress and uncertainties have more effect on students as they are immature as compared to instructors.Moreover, students' lives are at a progressive stage and every decision can impact their career.There is a tendency of students to use mobile phones during class which distracts them and is a big challenge from the student's viewpoint as compared to the instructor.Lack of infrastructure and power failures are perceived as a bigger challenge for students as uncertainties are more from student's perspective due to situations at their home and instructors with administration's support can reduce these challenges for them by backup facilities.Delays in lecture also affect students more than instructors as they are anxious about the content, delivery style and learning outcomes of the online lectures, especially in COVID-19 situation.
Some of the challenges are equally rated by students and instructors.Procrastination has been equally ranked by both students and instructors as it is based on personal behavior not category.Interruptions in classes are also ranked equal as interruptions break the pace of lecture and instructor and students both are affected alike.

CONCLUSION
The sudden shift from traditional classes to the online mode of instructions caught the instructors and students unaware.They did not get sufficient time for preparation and were not able to gather the appropriate resources amidst the lockdown.Online classes pose multiple challenges such as overwhelming features provided by platforms, lack of internet access, inhibitions while using technology, lack of social interaction, etc.Combined with these challenges were the problems associated with coronavirus pandemic such as mental stress and anxiety resulting from the uncertainties.Hence, there were a number of challenges for both students and instructors while participating in online classes.
The ranking of the challenges by TOPSIS method presents the viewpoint of students and instructor for problems in executing the online classes.These rankings allow to understand which of the challenge has greater impact.Meticulous efforts can be made to assuage the challenge.
To interpret how the challenges affect each other and which challenges have a greater role to play in the given scenario, an ISM model is generated.This model provides an approach to help in deciding which challenges need to be addressed first to mitigate all the challenges effectively.The challenges which are present in Level I of the ISM significantly influence the other challenges, hence they should be mitigated first.Lack of access to internet ranked as the most significant challenge affecting all the other challenges as it is present in Level 1 of ISM model, should be addressed first.Although lack of personal motivation is considered a big challenge in the overall ranking, its effect is due to other challenges, which is evident as it lies in the top level of ISM.Hence, it should not be the prime challenge to mitigate.It is of utmost importance for the universities to make the online classes successful, in order to do so they must provide proper infrastructure, backup power and internet facilities.
The main focus for any policymaker should be to rectify the strategy so it reduces the impact of challenges in the first three level of ISM.This can be accomplished by the recommendations: conscious effort should be made to focus in the class, administration should provide backup services, cordial environment must be maintained and formal training should be provided.

LIMITATIONS
Although, the study has considered the impact of 25 challenges, more challenges can be encountered at later stages.The study does not deal with open learning systems as COVID-19 has not affected this particular style of study as compared to instructor based traditional style of teaching.Further, this study does not consider the personal competences of instructors and students as during the COVID-19 the students registered in particular course continued to learn under the same instructor for the remaining part of the course, only the medium of delivery of lectures changed.The effect of personal competences and learning abilities will have a role if either the students or instructors changed during the course execution or the mode of course offering was different.
Sangeeta Sharma (PhD) has been working as faculty of communication, creativity, and advertising for the last 25 years in the Department of Humanities and social sciences at BITS, Pilani, and had been head of department and coordinator of centre for women studies.She is an active and lively person who participates in sports and related activities.Dr. Sharma is an academician of par excellence for the last two and half decades.She has excelled in all aspects of academics such as classroom delivery, corporate connect and training, writing of books and research papers, paper presentation in national and international conferences, editorial work and organizing academic events, etc. Arpan Bumb is a student at Birla Institute of Technology and Science, Pilani, and has a strong academic background.He consistently is in the top 10 percentile of the class and has topped in ten courses at the undergraduate level.He is interested in marketing mix, market research, and consumer behavior and has certification in them.

Figure 4 .
Figure 4. Influence and dependence map

Figure
Figure 5. Final Graph Interpretive Structural Model (ISM)

Table 3 . Overall ranking for the challenges
Lack/problems of infrastructure to participate in online class, and Power failures, Level 2 consisting of Lack of internet access or connectivity issues, Time lag in speaking, Use of mobiles and other technologies during online class, and Lack of interaction with peers, and Level 3 having the challenges Inhibition in using technology, Insufficient training to use the system, and Lack of immediate feedback.The challenges in Level 1 are related to effect on mental health and availability of infrastructure which have direct influence on the challenges present is all levels.For example, mental stress can lead a person to use mobile phones during class to release the anxiety.Similarly, power failure can lead to lack of internet access, time lag in speaking and lack of interaction.