dissertation about online learning

Advertisement

The effects of online education on academic success: A meta-analysis study

• Published: 06 September 2021
• Volume 27 , pages 429–450, ( 2022 )

Cite this article

• Hakan Ulum   ORCID: orcid.org/0000-0002-1398-6935 1  

81k Accesses

30 Citations

10 Altmetric

Explore all metrics

The purpose of this study is to analyze the effect of online education, which has been extensively used on student achievement since the beginning of the pandemic. In line with this purpose, a meta-analysis of the related studies focusing on the effect of online education on students’ academic achievement in several countries between the years 2010 and 2021 was carried out. Furthermore, this study will provide a source to assist future studies with comparing the effect of online education on academic achievement before and after the pandemic. This meta-analysis study consists of 27 studies in total. The meta-analysis involves the studies conducted in the USA, Taiwan, Turkey, China, Philippines, Ireland, and Georgia. The studies included in the meta-analysis are experimental studies, and the total sample size is 1772. In the study, the funnel plot, Duval and Tweedie’s Trip and Fill Analysis, Orwin’s Safe N Analysis, and Egger’s Regression Test were utilized to determine the publication bias, which has been found to be quite low. Besides, Hedge’s g statistic was employed to measure the effect size for the difference between the means performed in accordance with the random effects model. The results of the study show that the effect size of online education on academic achievement is on a medium level. The heterogeneity test results of the meta-analysis study display that the effect size does not differ in terms of class level, country, online education approaches, and lecture moderators.

Avoid common mistakes on your manuscript.

1 Introduction

Information and communication technologies have become a powerful force in transforming the educational settings around the world. The pandemic has been an important factor in transferring traditional physical classrooms settings through adopting information and communication technologies and has also accelerated the transformation. The literature supports that learning environments connected to information and communication technologies highly satisfy students. Therefore, we need to keep interest in technology-based learning environments. Clearly, technology has had a huge impact on young people's online lives. This digital revolution can synergize the educational ambitions and interests of digitally addicted students. In essence, COVID-19 has provided us with an opportunity to embrace online learning as education systems have to keep up with the rapid emergence of new technologies.

Information and communication technologies that have an effect on all spheres of life are also actively included in the education field. With the recent developments, using technology in education has become inevitable due to personal and social reasons (Usta, 2011a ). Online education may be given as an example of using information and communication technologies as a consequence of the technological developments. Also, it is crystal clear that online learning is a popular way of obtaining instruction (Demiralay et al., 2016 ; Pillay et al., 2007 ), which is defined by Horton ( 2000 ) as a way of education that is performed through a web browser or an online application without requiring an extra software or a learning source. Furthermore, online learning is described as a way of utilizing the internet to obtain the related learning sources during the learning process, to interact with the content, the teacher, and other learners, as well as to get support throughout the learning process (Ally, 2004 ). Online learning has such benefits as learning independently at any time and place (Vrasidas & MsIsaac, 2000 ), granting facility (Poole, 2000 ), flexibility (Chizmar & Walbert, 1999 ), self-regulation skills (Usta, 2011b ), learning with collaboration, and opportunity to plan self-learning process.

Even though online education practices have not been comprehensive as it is now, internet and computers have been used in education as alternative learning tools in correlation with the advances in technology. The first distance education attempt in the world was initiated by the ‘Steno Courses’ announcement published in Boston newspaper in 1728. Furthermore, in the nineteenth century, Sweden University started the “Correspondence Composition Courses” for women, and University Correspondence College was afterwards founded for the correspondence courses in 1843 (Arat & Bakan, 2011 ). Recently, distance education has been performed through computers, assisted by the facilities of the internet technologies, and soon, it has evolved into a mobile education practice that is emanating from progress in the speed of internet connection, and the development of mobile devices.

With the emergence of pandemic (Covid-19), face to face education has almost been put to a halt, and online education has gained significant importance. The Microsoft management team declared to have 750 users involved in the online education activities on the 10 th March, just before the pandemic; however, on March 24, they informed that the number of users increased significantly, reaching the number of 138,698 users (OECD, 2020 ). This event supports the view that it is better to commonly use online education rather than using it as a traditional alternative educational tool when students do not have the opportunity to have a face to face education (Geostat, 2019 ). The period of Covid-19 pandemic has emerged as a sudden state of having limited opportunities. Face to face education has stopped in this period for a long time. The global spread of Covid-19 affected more than 850 million students all around the world, and it caused the suspension of face to face education. Different countries have proposed several solutions in order to maintain the education process during the pandemic. Schools have had to change their curriculum, and many countries supported the online education practices soon after the pandemic. In other words, traditional education gave its way to online education practices. At least 96 countries have been motivated to access online libraries, TV broadcasts, instructions, sources, video lectures, and online channels (UNESCO, 2020 ). In such a painful period, educational institutions went through online education practices by the help of huge companies such as Microsoft, Google, Zoom, Skype, FaceTime, and Slack. Thus, online education has been discussed in the education agenda more intensively than ever before.

Although online education approaches were not used as comprehensively as it has been used recently, it was utilized as an alternative learning approach in education for a long time in parallel with the development of technology, internet and computers. The academic achievement of the students is often aimed to be promoted by employing online education approaches. In this regard, academicians in various countries have conducted many studies on the evaluation of online education approaches and published the related results. However, the accumulation of scientific data on online education approaches creates difficulties in keeping, organizing and synthesizing the findings. In this research area, studies are being conducted at an increasing rate making it difficult for scientists to be aware of all the research outside of their ​​expertise. Another problem encountered in the related study area is that online education studies are repetitive. Studies often utilize slightly different methods, measures, and/or examples to avoid duplication. This erroneous approach makes it difficult to distinguish between significant differences in the related results. In other words, if there are significant differences in the results of the studies, it may be difficult to express what variety explains the differences in these results. One obvious solution to these problems is to systematically review the results of various studies and uncover the sources. One method of performing such systematic syntheses is the application of meta-analysis which is a methodological and statistical approach to draw conclusions from the literature. At this point, how effective online education applications are in increasing the academic success is an important detail. Has online education, which is likely to be encountered frequently in the continuing pandemic period, been successful in the last ten years? If successful, how much was the impact? Did different variables have an impact on this effect? Academics across the globe have carried out studies on the evaluation of online education platforms and publishing the related results (Chiao et al., 2018 ). It is quite important to evaluate the results of the studies that have been published up until now, and that will be published in the future. Has the online education been successful? If it has been, how big is the impact? Do the different variables affect this impact? What should we consider in the next coming online education practices? These questions have all motivated us to carry out this study. We have conducted a comprehensive meta-analysis study that tries to provide a discussion platform on how to develop efficient online programs for educators and policy makers by reviewing the related studies on online education, presenting the effect size, and revealing the effect of diverse variables on the general impact.

There have been many critical discussions and comprehensive studies on the differences between online and face to face learning; however, the focus of this paper is different in the sense that it clarifies the magnitude of the effect of online education and teaching process, and it represents what factors should be controlled to help increase the effect size. Indeed, the purpose here is to provide conscious decisions in the implementation of the online education process.

The general impact of online education on the academic achievement will be discovered in the study. Therefore, this will provide an opportunity to get a general overview of the online education which has been practiced and discussed intensively in the pandemic period. Moreover, the general impact of online education on academic achievement will be analyzed, considering different variables. In other words, the current study will allow to totally evaluate the study results from the related literature, and to analyze the results considering several cultures, lectures, and class levels. Considering all the related points, this study seeks to answer the following research questions:

What is the effect size of online education on academic achievement?

How do the effect sizes of online education on academic achievement change according to the moderator variable of the country?

How do the effect sizes of online education on academic achievement change according to the moderator variable of the class level?

How do the effect sizes of online education on academic achievement change according to the moderator variable of the lecture?

How do the effect sizes of online education on academic achievement change according to the moderator variable of the online education approaches?

This study aims at determining the effect size of online education, which has been highly used since the beginning of the pandemic, on students’ academic achievement in different courses by using a meta-analysis method. Meta-analysis is a synthesis method that enables gathering of several study results accurately and efficiently, and getting the total results in the end (Tsagris & Fragkos, 2018 ).

2.1 Selecting and coding the data (studies)

The required literature for the meta-analysis study was reviewed in July, 2020, and the follow-up review was conducted in September, 2020. The purpose of the follow-up review was to include the studies which were published in the conduction period of this study, and which met the related inclusion criteria. However, no study was encountered to be included in the follow-up review.

In order to access the studies in the meta-analysis, the databases of Web of Science, ERIC, and SCOPUS were reviewed by utilizing the keywords ‘online learning and online education’. Not every database has a search engine that grants access to the studies by writing the keywords, and this obstacle was considered to be an important problem to be overcome. Therefore, a platform that has a special design was utilized by the researcher. With this purpose, through the open access system of Cukurova University Library, detailed reviews were practiced using EBSCO Information Services (EBSCO) that allow reviewing the whole collection of research through a sole searching box. Since the fundamental variables of this study are online education and online learning, the literature was systematically reviewed in the related databases (Web of Science, ERIC, and SCOPUS) by referring to the keywords. Within this scope, 225 articles were accessed, and the studies were included in the coding key list formed by the researcher. The name of the researchers, the year, the database (Web of Science, ERIC, and SCOPUS), the sample group and size, the lectures that the academic achievement was tested in, the country that the study was conducted in, and the class levels were all included in this coding key.

The following criteria were identified to include 225 research studies which were coded based on the theoretical basis of the meta-analysis study: (1) The studies should be published in the refereed journals between the years 2020 and 2021, (2) The studies should be experimental studies that try to determine the effect of online education and online learning on academic achievement, (3) The values of the stated variables or the required statistics to calculate these values should be stated in the results of the studies, and (4) The sample group of the study should be at a primary education level. These criteria were also used as the exclusion criteria in the sense that the studies that do not meet the required criteria were not included in the present study.

After the inclusion criteria were determined, a systematic review process was conducted, following the year criterion of the study by means of EBSCO. Within this scope, 290,365 studies that analyze the effect of online education and online learning on academic achievement were accordingly accessed. The database (Web of Science, ERIC, and SCOPUS) was also used as a filter by analyzing the inclusion criteria. Hence, the number of the studies that were analyzed was 58,616. Afterwards, the keyword ‘primary education’ was used as the filter and the number of studies included in the study decreased to 3152. Lastly, the literature was reviewed by using the keyword ‘academic achievement’ and 225 studies were accessed. All the information of 225 articles was included in the coding key.

It is necessary for the coders to review the related studies accurately and control the validity, safety, and accuracy of the studies (Stewart & Kamins, 2001 ). Within this scope, the studies that were determined based on the variables used in this study were first reviewed by three researchers from primary education field, then the accessed studies were combined and processed in the coding key by the researcher. All these studies that were processed in the coding key were analyzed in accordance with the inclusion criteria by all the researchers in the meetings, and it was decided that 27 studies met the inclusion criteria (Atici & Polat, 2010 ; Carreon, 2018 ; Ceylan & Elitok Kesici, 2017 ; Chae & Shin, 2016 ; Chiang et al. 2014 ; Ercan, 2014 ; Ercan et al., 2016 ; Gwo-Jen et al., 2018 ; Hayes & Stewart, 2016 ; Hwang et al., 2012 ; Kert et al., 2017 ; Lai & Chen, 2010 ; Lai et al., 2015 ; Meyers et al., 2015 ; Ravenel et al., 2014 ; Sung et al., 2016 ; Wang & Chen, 2013 ; Yu, 2019 ; Yu & Chen, 2014 ; Yu & Pan, 2014 ; Yu et al., 2010 ; Zhong et al., 2017 ). The data from the studies meeting the inclusion criteria were independently processed in the second coding key by three researchers, and consensus meetings were arranged for further discussion. After the meetings, researchers came to an agreement that the data were coded accurately and precisely. Having identified the effect sizes and heterogeneity of the study, moderator variables that will show the differences between the effect sizes were determined. The data related to the determined moderator variables were added to the coding key by three researchers, and a new consensus meeting was arranged. After the meeting, researchers came to an agreement that moderator variables were coded accurately and precisely.

2.2 Study group

27 studies are included in the meta-analysis. The total sample size of the studies that are included in the analysis is 1772. The characteristics of the studies included are given in Table 1 .

2.3 Publication bias

Publication bias is the low capability of published studies on a research subject to represent all completed studies on the same subject (Card, 2011 ; Littell et al., 2008 ). Similarly, publication bias is the state of having a relationship between the probability of the publication of a study on a subject, and the effect size and significance that it produces. Within this scope, publication bias may occur when the researchers do not want to publish the study as a result of failing to obtain the expected results, or not being approved by the scientific journals, and consequently not being included in the study synthesis (Makowski et al., 2019 ). The high possibility of publication bias in a meta-analysis study negatively affects (Pecoraro, 2018 ) the accuracy of the combined effect size, causing the average effect size to be reported differently than it should be (Borenstein et al., 2009 ). For this reason, the possibility of publication bias in the included studies was tested before determining the effect sizes of the relationships between the stated variables. The possibility of publication bias of this meta-analysis study was analyzed by using the funnel plot, Orwin’s Safe N Analysis, Duval and Tweedie’s Trip and Fill Analysis, and Egger’s Regression Test.

2.4 Selecting the model

After determining the probability of publication bias of this meta-analysis study, the statistical model used to calculate the effect sizes was selected. The main approaches used in the effect size calculations according to the differentiation level of inter-study variance are fixed and random effects models (Pigott, 2012 ). Fixed effects model refers to the homogeneity of the characteristics of combined studies apart from the sample sizes, while random effects model refers to the parameter diversity between the studies (Cumming, 2012 ). While calculating the average effect size in the random effects model (Deeks et al., 2008 ) that is based on the assumption that effect predictions of different studies are only the result of a similar distribution, it is necessary to consider several situations such as the effect size apart from the sample error of combined studies, characteristics of the participants, duration, scope, and pattern of the study (Littell et al., 2008 ). While deciding the model in the meta-analysis study, the assumptions on the sample characteristics of the studies included in the analysis and the inferences that the researcher aims to make should be taken into consideration. The fact that the sample characteristics of the studies conducted in the field of social sciences are affected by various parameters shows that using random effects model is more appropriate in this sense. Besides, it is stated that the inferences made with the random effects model are beyond the studies included in the meta-analysis (Field, 2003 ; Field & Gillett, 2010 ). Therefore, using random effects model also contributes to the generalization of research data. The specified criteria for the statistical model selection show that according to the nature of the meta-analysis study, the model should be selected just before the analysis (Borenstein et al., 2007 ; Littell et al., 2008 ). Within this framework, it was decided to make use of the random effects model, considering that the students who are the samples of the studies included in the meta-analysis are from different countries and cultures, the sample characteristics of the studies differ, and the patterns and scopes of the studies vary as well.

2.5 Heterogeneity

Meta-analysis facilitates analyzing the research subject with different parameters by showing the level of diversity between the included studies. Within this frame, whether there is a heterogeneous distribution between the studies included in the study or not has been evaluated in the present study. The heterogeneity of the studies combined in this meta-analysis study has been determined through Q and I 2 tests. Q test evaluates the random distribution probability of the differences between the observed results (Deeks et al., 2008 ). Q value exceeding 2 value calculated according to the degree of freedom and significance, indicates the heterogeneity of the combined effect sizes (Card, 2011 ). I 2 test, which is the complementary of the Q test, shows the heterogeneity amount of the effect sizes (Cleophas & Zwinderman, 2017 ). I 2 value being higher than 75% is explained as high level of heterogeneity.

In case of encountering heterogeneity in the studies included in the meta-analysis, the reasons of heterogeneity can be analyzed by referring to the study characteristics. The study characteristics which may be related to the heterogeneity between the included studies can be interpreted through subgroup analysis or meta-regression analysis (Deeks et al., 2008 ). While determining the moderator variables, the sufficiency of the number of variables, the relationship between the moderators, and the condition to explain the differences between the results of the studies have all been considered in the present study. Within this scope, it was predicted in this meta-analysis study that the heterogeneity can be explained with the country, class level, and lecture moderator variables of the study in terms of the effect of online education, which has been highly used since the beginning of the pandemic, and it has an impact on the students’ academic achievement in different lectures. Some subgroups were evaluated and categorized together, considering that the number of effect sizes of the sub-dimensions of the specified variables is not sufficient to perform moderator analysis (e.g. the countries where the studies were conducted).

2.6 Interpreting the effect sizes

Effect size is a factor that shows how much the independent variable affects the dependent variable positively or negatively in each included study in the meta-analysis (Dinçer, 2014 ). While interpreting the effect sizes obtained from the meta-analysis, the classifications of Cohen et al. ( 2007 ) have been utilized. The case of differentiating the specified relationships of the situation of the country, class level, and school subject variables of the study has been identified through the Q test, degree of freedom, and p significance value Fig.  1 and 2 .

3 Findings and results

The purpose of this study is to determine the effect size of online education on academic achievement. Before determining the effect sizes in the study, the probability of publication bias of this meta-analysis study was analyzed by using the funnel plot, Orwin’s Safe N Analysis, Duval and Tweedie’s Trip and Fill Analysis, and Egger’s Regression Test.

When the funnel plots are examined, it is seen that the studies included in the analysis are distributed symmetrically on both sides of the combined effect size axis, and they are generally collected in the middle and lower sections. The probability of publication bias is low according to the plots. However, since the results of the funnel scatter plots may cause subjective interpretations, they have been supported by additional analyses (Littell et al., 2008 ). Therefore, in order to provide an extra proof for the probability of publication bias, it has been analyzed through Orwin’s Safe N Analysis, Duval and Tweedie’s Trip and Fill Analysis, and Egger’s Regression Test (Table 2 ).

Table 2 consists of the results of the rates of publication bias probability before counting the effect size of online education on academic achievement. According to the table, Orwin Safe N analysis results show that it is not necessary to add new studies to the meta-analysis in order for Hedges g to reach a value outside the range of ± 0.01. The Duval and Tweedie test shows that excluding the studies that negatively affect the symmetry of the funnel scatter plots for each meta-analysis or adding their exact symmetrical equivalents does not significantly differentiate the calculated effect size. The insignificance of the Egger tests results reveals that there is no publication bias in the meta-analysis study. The results of the analysis indicate the high internal validity of the effect sizes and the adequacy of representing the studies conducted on the relevant subject.

In this study, it was aimed to determine the effect size of online education on academic achievement after testing the publication bias. In line with the first purpose of the study, the forest graph regarding the effect size of online education on academic achievement is shown in Fig.  3 , and the statistics regarding the effect size are given in Table 3 .

The flow chart of the scanning and selection process of the studies

Funnel plot graphics representing the effect size of the effects of online education on academic success

Forest graph related to the effect size of online education on academic success

The square symbols in the forest graph in Fig.  3 represent the effect sizes, while the horizontal lines show the intervals in 95% confidence of the effect sizes, and the diamond symbol shows the overall effect size. When the forest graph is analyzed, it is seen that the lower and upper limits of the combined effect sizes are generally close to each other, and the study loads are similar. This similarity in terms of study loads indicates the similarity of the contribution of the combined studies to the overall effect size.

Figure  3 clearly represents that the study of Liu and others (Liu et al., 2018 ) has the lowest, and the study of Ercan and Bilen ( 2014 ) has the highest effect sizes. The forest graph shows that all the combined studies and the overall effect are positive. Furthermore, it is simply understood from the forest graph in Fig.  3 and the effect size statistics in Table 3 that the results of the meta-analysis study conducted with 27 studies and analyzing the effect of online education on academic achievement illustrate that this relationship is on average level (= 0.409).

After the analysis of the effect size in the study, whether the studies included in the analysis are distributed heterogeneously or not has also been analyzed. The heterogeneity of the combined studies was determined through the Q and I 2 tests. As a result of the heterogeneity test, Q statistical value was calculated as 29.576. With 26 degrees of freedom at 95% significance level in the chi-square table, the critical value is accepted as 38.885. The Q statistical value (29.576) counted in this study is lower than the critical value of 38.885. The I 2 value, which is the complementary of the Q statistics, is 12.100%. This value indicates that the accurate heterogeneity or the total variability that can be attributed to variability between the studies is 12%. Besides, p value is higher than (0.285) p = 0.05. All these values [Q (26) = 29.579, p = 0.285; I2 = 12.100] indicate that there is a homogeneous distribution between the effect sizes, and fixed effects model should be used to interpret these effect sizes. However, some researchers argue that even if the heterogeneity is low, it should be evaluated based on the random effects model (Borenstein et al., 2007 ). Therefore, this study gives information about both models. The heterogeneity of the combined studies has been attempted to be explained with the characteristics of the studies included in the analysis. In this context, the final purpose of the study is to determine the effect of the country, academic level, and year variables on the findings. Accordingly, the statistics regarding the comparison of the stated relations according to the countries where the studies were conducted are given in Table 4 .

As seen in Table 4 , the effect of online education on academic achievement does not differ significantly according to the countries where the studies were conducted in. Q test results indicate the heterogeneity of the relationships between the variables in terms of countries where the studies were conducted in. According to the table, the effect of online education on academic achievement was reported as the highest in other countries, and the lowest in the US. The statistics regarding the comparison of the stated relations according to the class levels are given in Table 5 .

As seen in Table 5 , the effect of online education on academic achievement does not differ according to the class level. However, the effect of online education on academic achievement is the highest in the 4 th class. The statistics regarding the comparison of the stated relations according to the class levels are given in Table 6 .

As seen in Table 6 , the effect of online education on academic achievement does not differ according to the school subjects included in the studies. However, the effect of online education on academic achievement is the highest in ICT subject.

The obtained effect size in the study was formed as a result of the findings attained from primary studies conducted in 7 different countries. In addition, these studies are the ones on different approaches to online education (online learning environments, social networks, blended learning, etc.). In this respect, the results may raise some questions about the validity and generalizability of the results of the study. However, the moderator analyzes, whether for the country variable or for the approaches covered by online education, did not create significant differences in terms of the effect sizes. If significant differences were to occur in terms of effect sizes, we could say that the comparisons we will make by comparing countries under the umbrella of online education would raise doubts in terms of generalizability. Moreover, no study has been found in the literature that is not based on a special approach or does not contain a specific technique conducted under the name of online education alone. For instance, one of the commonly used definitions is blended education which is defined as an educational model in which online education is combined with traditional education method (Colis & Moonen, 2001 ). Similarly, Rasmussen ( 2003 ) defines blended learning as “a distance education method that combines technology (high technology such as television, internet, or low technology such as voice e-mail, conferences) with traditional education and training.” Further, Kerres and Witt (2003) define blended learning as “combining face-to-face learning with technology-assisted learning.” As it is clearly observed, online education, which has a wider scope, includes many approaches.

As seen in Table 7 , the effect of online education on academic achievement does not differ according to online education approaches included in the studies. However, the effect of online education on academic achievement is the highest in Web Based Problem Solving Approach.

4 Conclusions and discussion

Considering the developments during the pandemics, it is thought that the diversity in online education applications as an interdisciplinary pragmatist field will increase, and the learning content and processes will be enriched with the integration of new technologies into online education processes. Another prediction is that more flexible and accessible learning opportunities will be created in online education processes, and in this way, lifelong learning processes will be strengthened. As a result, it is predicted that in the near future, online education and even digital learning with a newer name will turn into the main ground of education instead of being an alternative or having a support function in face-to-face learning. The lessons learned from the early period online learning experience, which was passed with rapid adaptation due to the Covid19 epidemic, will serve to develop this method all over the world, and in the near future, online learning will become the main learning structure through increasing its functionality with the contribution of new technologies and systems. If we look at it from this point of view, there is a necessity to strengthen online education.

In this study, the effect of online learning on academic achievement is at a moderate level. To increase this effect, the implementation of online learning requires support from teachers to prepare learning materials, to design learning appropriately, and to utilize various digital-based media such as websites, software technology and various other tools to support the effectiveness of online learning (Rolisca & Achadiyah, 2014 ). According to research conducted by Rahayu et al. ( 2017 ), it has been proven that the use of various types of software increases the effectiveness and quality of online learning. Implementation of online learning can affect students' ability to adapt to technological developments in that it makes students use various learning resources on the internet to access various types of information, and enables them to get used to performing inquiry learning and active learning (Hart et al., 2019 ; Prestiadi et al., 2019 ). In addition, there may be many reasons for the low level of effect in this study. The moderator variables examined in this study could be a guide in increasing the level of practical effect. However, the effect size did not differ significantly for all moderator variables. Different moderator analyzes can be evaluated in order to increase the level of impact of online education on academic success. If confounding variables that significantly change the effect level are detected, it can be spoken more precisely in order to increase this level. In addition to the technical and financial problems, the level of impact will increase if a few other difficulties are eliminated such as students, lack of interaction with the instructor, response time, and lack of traditional classroom socialization.

In addition, COVID-19 pandemic related social distancing has posed extreme difficulties for all stakeholders to get online as they have to work in time constraints and resource constraints. Adopting the online learning environment is not just a technical issue, it is a pedagogical and instructive challenge as well. Therefore, extensive preparation of teaching materials, curriculum, and assessment is vital in online education. Technology is the delivery tool and requires close cross-collaboration between teaching, content and technology teams (CoSN, 2020 ).

Online education applications have been used for many years. However, it has come to the fore more during the pandemic process. This result of necessity has brought with it the discussion of using online education instead of traditional education methods in the future. However, with this research, it has been revealed that online education applications are moderately effective. The use of online education instead of face-to-face education applications can only be possible with an increase in the level of success. This may have been possible with the experience and knowledge gained during the pandemic process. Therefore, the meta-analysis of experimental studies conducted in the coming years will guide us. In this context, experimental studies using online education applications should be analyzed well. It would be useful to identify variables that can change the level of impacts with different moderators. Moderator analyzes are valuable in meta-analysis studies (for example, the role of moderators in Karl Pearson's typhoid vaccine studies). In this context, each analysis study sheds light on future studies. In meta-analyses to be made about online education, it would be beneficial to go beyond the moderators determined in this study. Thus, the contribution of similar studies to the field will increase more.

The purpose of this study is to determine the effect of online education on academic achievement. In line with this purpose, the studies that analyze the effect of online education approaches on academic achievement have been included in the meta-analysis. The total sample size of the studies included in the meta-analysis is 1772. While the studies included in the meta-analysis were conducted in the US, Taiwan, Turkey, China, Philippines, Ireland, and Georgia, the studies carried out in Europe could not be reached. The reason may be attributed to that there may be more use of quantitative research methods from a positivist perspective in the countries with an American academic tradition. As a result of the study, it was found out that the effect size of online education on academic achievement (g = 0.409) was moderate. In the studies included in the present research, we found that online education approaches were more effective than traditional ones. However, contrary to the present study, the analysis of comparisons between online and traditional education in some studies shows that face-to-face traditional learning is still considered effective compared to online learning (Ahmad et al., 2016 ; Hamdani & Priatna, 2020 ; Wei & Chou, 2020 ). Online education has advantages and disadvantages. The advantages of online learning compared to face-to-face learning in the classroom is the flexibility of learning time in online learning, the learning time does not include a single program, and it can be shaped according to circumstances (Lai et al., 2019 ). The next advantage is the ease of collecting assignments for students, as these can be done without having to talk to the teacher. Despite this, online education has several weaknesses, such as students having difficulty in understanding the material, teachers' inability to control students, and students’ still having difficulty interacting with teachers in case of internet network cuts (Swan, 2007 ). According to Astuti et al ( 2019 ), face-to-face education method is still considered better by students than e-learning because it is easier to understand the material and easier to interact with teachers. The results of the study illustrated that the effect size (g = 0.409) of online education on academic achievement is of medium level. Therefore, the results of the moderator analysis showed that the effect of online education on academic achievement does not differ in terms of country, lecture, class level, and online education approaches variables. After analyzing the literature, several meta-analyses on online education were published (Bernard et al., 2004 ; Machtmes & Asher, 2000 ; Zhao et al., 2005 ). Typically, these meta-analyzes also include the studies of older generation technologies such as audio, video, or satellite transmission. One of the most comprehensive studies on online education was conducted by Bernard et al. ( 2004 ). In this study, 699 independent effect sizes of 232 studies published from 1985 to 2001 were analyzed, and face-to-face education was compared to online education, with respect to success criteria and attitudes of various learners from young children to adults. In this meta-analysis, an overall effect size close to zero was found for the students' achievement (g +  = 0.01).

In another meta-analysis study carried out by Zhao et al. ( 2005 ), 98 effect sizes were examined, including 51 studies on online education conducted between 1996 and 2002. According to the study of Bernard et al. ( 2004 ), this meta-analysis focuses on the activities done in online education lectures. As a result of the research, an overall effect size close to zero was found for online education utilizing more than one generation technology for students at different levels. However, the salient point of the meta-analysis study of Zhao et al. is that it takes the average of different types of results used in a study to calculate an overall effect size. This practice is problematic because the factors that develop one type of learner outcome (e.g. learner rehabilitation), particularly course characteristics and practices, may be quite different from those that develop another type of outcome (e.g. learner's achievement), and it may even cause damage to the latter outcome. While mixing the studies with different types of results, this implementation may obscure the relationship between practices and learning.

Some meta-analytical studies have focused on the effectiveness of the new generation distance learning courses accessed through the internet for specific student populations. For instance, Sitzmann and others (Sitzmann et al., 2006 ) reviewed 96 studies published from 1996 to 2005, comparing web-based education of job-related knowledge or skills with face-to-face one. The researchers found that web-based education in general was slightly more effective than face-to-face education, but it is insufficient in terms of applicability ("knowing how to apply"). In addition, Sitzmann et al. ( 2006 ) revealed that Internet-based education has a positive effect on theoretical knowledge in quasi-experimental studies; however, it positively affects face-to-face education in experimental studies performed by random assignment. This moderator analysis emphasizes the need to pay attention to the factors of designs of the studies included in the meta-analysis. The designs of the studies included in this meta-analysis study were ignored. This can be presented as a suggestion to the new studies that will be conducted.

Another meta-analysis study was conducted by Cavanaugh et al. ( 2004 ), in which they focused on online education. In this study on internet-based distance education programs for students under 12 years of age, the researchers combined 116 results from 14 studies published between 1999 and 2004 to calculate an overall effect that was not statistically different from zero. The moderator analysis carried out in this study showed that there was no significant factor affecting the students' success. This meta-analysis used multiple results of the same study, ignoring the fact that different results of the same student would not be independent from each other.

In conclusion, some meta-analytical studies analyzed the consequences of online education for a wide range of students (Bernard et al., 2004 ; Zhao et al., 2005 ), and the effect sizes were generally low in these studies. Furthermore, none of the large-scale meta-analyzes considered the moderators, database quality standards or class levels in the selection of the studies, while some of them just referred to the country and lecture moderators. Advances in internet-based learning tools, the pandemic process, and increasing popularity in different learning contexts have required a precise meta-analysis of students' learning outcomes through online learning. Previous meta-analysis studies were typically based on the studies, involving narrow range of confounding variables. In the present study, common but significant moderators such as class level and lectures during the pandemic process were discussed. For instance, the problems have been experienced especially in terms of eligibility of class levels in online education platforms during the pandemic process. It was found that there is a need to study and make suggestions on whether online education can meet the needs of teachers and students.

Besides, the main forms of online education in the past were to watch the open lectures of famous universities and educational videos of institutions. In addition, online education is mainly a classroom-based teaching implemented by teachers in their own schools during the pandemic period, which is an extension of the original school education. This meta-analysis study will stand as a source to compare the effect size of the online education forms of the past decade with what is done today, and what will be done in the future.

Lastly, the heterogeneity test results of the meta-analysis study display that the effect size does not differ in terms of class level, country, online education approaches, and lecture moderators.

*Studies included in meta-analysis

Ahmad, S., Sumardi, K., & Purnawan, P. (2016). Komparasi Peningkatan Hasil Belajar Antara Pembelajaran Menggunakan Sistem Pembelajaran Online Terpadu Dengan Pembelajaran Klasikal Pada Mata Kuliah Pneumatik Dan Hidrolik. Journal of Mechanical Engineering Education, 2 (2), 286–292.

Article   Google Scholar  

Ally, M. (2004). Foundations of educational theory for online learning. Theory and Practice of Online Learning, 2 , 15–44. Retrieved on the 11th of September, 2020 from https://eddl.tru.ca/wp-content/uploads/2018/12/01_Anderson_2008-Theory_and_Practice_of_Online_Learning.pdf

Arat, T., & Bakan, Ö. (2011). Uzaktan eğitim ve uygulamaları. Selçuk Üniversitesi Sosyal Bilimler Meslek Yüksek Okulu Dergisi , 14 (1–2), 363–374. https://doi.org/10.29249/selcuksbmyd.540741

Astuti, C. C., Sari, H. M. K., & Azizah, N. L. (2019). Perbandingan Efektifitas Proses Pembelajaran Menggunakan Metode E-Learning dan Konvensional. Proceedings of the ICECRS, 2 (1), 35–40.

*Atici, B., & Polat, O. C. (2010). Influence of the online learning environments and tools on the student achievement and opinions. Educational Research and Reviews, 5 (8), 455–464. Retrieved on the 11th of October, 2020 from https://academicjournals.org/journal/ERR/article-full-text-pdf/4C8DD044180.pdf

Bernard, R. M., Abrami, P. C., Lou, Y., Borokhovski, E., Wade, A., Wozney, L., et al. (2004). How does distance education compare with classroom instruction? A meta- analysis of the empirical literature. Review of Educational Research, 3 (74), 379–439. https://doi.org/10.3102/00346543074003379

Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis . Wiley.

Book   Google Scholar  

Borenstein, M., Hedges, L., & Rothstein, H. (2007). Meta-analysis: Fixed effect vs. random effects . UK: Wiley.

Card, N. A. (2011). Applied meta-analysis for social science research: Methodology in the social sciences . Guilford.

Google Scholar  

*Carreon, J. R. (2018 ). Facebook as integrated blended learning tool in technology and livelihood education exploratory. Retrieved on the 1st of October, 2020 from https://files.eric.ed.gov/fulltext/EJ1197714.pdf

Cavanaugh, C., Gillan, K. J., Kromrey, J., Hess, M., & Blomeyer, R. (2004). The effects of distance education on K-12 student outcomes: A meta-analysis. Learning Point Associates/North Central Regional Educational Laboratory (NCREL) . Retrieved on the 11th of September, 2020 from https://files.eric.ed.gov/fulltext/ED489533.pdf

*Ceylan, V. K., & Elitok Kesici, A. (2017). Effect of blended learning to academic achievement. Journal of Human Sciences, 14 (1), 308. https://doi.org/10.14687/jhs.v14i1.4141

*Chae, S. E., & Shin, J. H. (2016). Tutoring styles that encourage learner satisfaction, academic engagement, and achievement in an online environment. Interactive Learning Environments, 24(6), 1371–1385. https://doi.org/10.1080/10494820.2015.1009472

*Chiang, T. H. C., Yang, S. J. H., & Hwang, G. J. (2014). An augmented reality-based mobile learning system to improve students’ learning achievements and motivations in natural science inquiry activities. Educational Technology and Society, 17 (4), 352–365. Retrieved on the 11th of September, 2020 from https://www.researchgate.net/profile/Gwo_Jen_Hwang/publication/287529242_An_Augmented_Reality-based_Mobile_Learning_System_to_Improve_Students'_Learning_Achievements_and_Motivations_in_Natural_Science_Inquiry_Activities/links/57198c4808ae30c3f9f2c4ac.pdf

Chiao, H. M., Chen, Y. L., & Huang, W. H. (2018). Examining the usability of an online virtual tour-guiding platform for cultural tourism education. Journal of Hospitality, Leisure, Sport & Tourism Education, 23 (29–38), 1. https://doi.org/10.1016/j.jhlste.2018.05.002

Chizmar, J. F., & Walbert, M. S. (1999). Web-based learning environments guided by principles of good teaching practice. Journal of Economic Education, 30 (3), 248–264. https://doi.org/10.2307/1183061

Cleophas, T. J., & Zwinderman, A. H. (2017). Modern meta-analysis: Review and update of methodologies . Switzerland: Springer. https://doi.org/10.1007/978-3-319-55895-0

Cohen, L., Manion, L., & Morrison, K. (2007). Observation.  Research Methods in Education, 6 , 396–412. Retrieved on the 11th of September, 2020 from https://www.researchgate.net/profile/Nabil_Ashraf2/post/How_to_get_surface_potential_Vs_Voltage_curve_from_CV_and_GV_measurements_of_MOS_capacitor/attachment/5ac6033cb53d2f63c3c405b4/AS%3A612011817844736%401522926396219/download/Very+important_C-V+characterization+Lehigh+University+thesis.pdf

Colis, B., & Moonen, J. (2001). Flexible Learning in a Digital World: Experiences and Expectations. Open & Distance Learning Series . Stylus Publishing.

CoSN. (2020). COVID-19 Response: Preparing to Take School Online. CoSN. (2020). COVID-19 Response: Preparing to Take School Online. Retrieved on the 3rd of September, 2021 from https://www.cosn.org/sites/default/files/COVID-19%20Member%20Exclusive_0.pdf

Cumming, G. (2012). Understanding new statistics: Effect sizes, confidence intervals, and meta-analysis. New York, USA: Routledge. https://doi.org/10.4324/9780203807002

Deeks, J. J., Higgins, J. P. T., & Altman, D. G. (2008). Analysing data and undertaking meta-analyses . In J. P. T. Higgins & S. Green (Eds.), Cochrane handbook for systematic reviews of interventions (pp. 243–296). Sussex: John Wiley & Sons. https://doi.org/10.1002/9780470712184.ch9

Demiralay, R., Bayır, E. A., & Gelibolu, M. F. (2016). Öğrencilerin bireysel yenilikçilik özellikleri ile çevrimiçi öğrenmeye hazır bulunuşlukları ilişkisinin incelenmesi. Eğitim ve Öğretim Araştırmaları Dergisi, 5 (1), 161–168. https://doi.org/10.23891/efdyyu.2017.10

Dinçer, S. (2014). Eğitim bilimlerinde uygulamalı meta-analiz. Pegem Atıf İndeksi, 2014(1), 1–133. https://doi.org/10.14527/pegem.001

*Durak, G., Cankaya, S., Yunkul, E., & Ozturk, G. (2017). The effects of a social learning network on students’ performances and attitudes. European Journal of Education Studies, 3 (3), 312–333. 10.5281/zenodo.292951

*Ercan, O. (2014). Effect of web assisted education supported by six thinking hats on students’ academic achievement in science and technology classes . European Journal of Educational Research, 3 (1), 9–23. https://doi.org/10.12973/eu-jer.3.1.9

Ercan, O., & Bilen, K. (2014). Effect of web assisted education supported by six thinking hats on students’ academic achievement in science and technology classes. European Journal of Educational Research, 3 (1), 9–23.

*Ercan, O., Bilen, K., & Ural, E. (2016). “Earth, sun and moon”: Computer assisted instruction in secondary school science - Achievement and attitudes. Issues in Educational Research, 26 (2), 206–224. https://doi.org/10.12973/eu-jer.3.1.9

Field, A. P. (2003). The problems in using fixed-effects models of meta-analysis on real-world data. Understanding Statistics, 2 (2), 105–124. https://doi.org/10.1207/s15328031us0202_02

Field, A. P., & Gillett, R. (2010). How to do a meta-analysis. British Journal of Mathematical and Statistical Psychology, 63 (3), 665–694. https://doi.org/10.1348/00071010x502733

Geostat. (2019). ‘Share of households with internet access’, National statistics office of Georgia . Retrieved on the 2nd September 2020 from https://www.geostat.ge/en/modules/categories/106/information-and-communication-technologies-usage-in-households

*Gwo-Jen, H., Nien-Ting, T., & Xiao-Ming, W. (2018). Creating interactive e-books through learning by design: The impacts of guided peer-feedback on students’ learning achievements and project outcomes in science courses. Journal of Educational Technology & Society., 21 (1), 25–36. Retrieved on the 2nd of October, 2020 https://ae-uploads.uoregon.edu/ISTE/ISTE2019/PROGRAM_SESSION_MODEL/HANDOUTS/112172923/CreatingInteractiveeBooksthroughLearningbyDesignArticle2018.pdf

Hamdani, A. R., & Priatna, A. (2020). Efektifitas implementasi pembelajaran daring (full online) dimasa pandemi Covid-19 pada jenjang Sekolah Dasar di Kabupaten Subang. Didaktik: Jurnal Ilmiah PGSD STKIP Subang, 6 (1), 1–9.

Hart, C. M., Berger, D., Jacob, B., Loeb, S., & Hill, M. (2019). Online learning, offline outcomes: Online course taking and high school student performance. Aera Open, 5(1).

*Hayes, J., & Stewart, I. (2016). Comparing the effects of derived relational training and computer coding on intellectual potential in school-age children. The British Journal of Educational Psychology, 86 (3), 397–411. https://doi.org/10.1111/bjep.12114

Horton, W. K. (2000). Designing web-based training: How to teach anyone anything anywhere anytime (Vol. 1). Wiley Publishing.

*Hwang, G. J., Wu, P. H., & Chen, C. C. (2012). An online game approach for improving students’ learning performance in web-based problem-solving activities. Computers and Education, 59 (4), 1246–1256. https://doi.org/10.1016/j.compedu.2012.05.009

*Kert, S. B., Köşkeroğlu Büyükimdat, M., Uzun, A., & Çayiroğlu, B. (2017). Comparing active game-playing scores and academic performances of elementary school students. Education 3–13, 45 (5), 532–542. https://doi.org/10.1080/03004279.2016.1140800

*Lai, A. F., & Chen, D. J. (2010). Web-based two-tier diagnostic test and remedial learning experiment. International Journal of Distance Education Technologies, 8 (1), 31–53. https://doi.org/10.4018/jdet.2010010103

*Lai, A. F., Lai, H. Y., Chuang W. H., & Wu, Z.H. (2015). Developing a mobile learning management system for outdoors nature science activities based on 5e learning cycle. Proceedings of the International Conference on e-Learning, ICEL. Proceedings of the International Association for Development of the Information Society (IADIS) International Conference on e-Learning (Las Palmas de Gran Canaria, Spain, July 21–24, 2015). Retrieved on the 14th November 2020 from https://files.eric.ed.gov/fulltext/ED562095.pdf

Lai, C. H., Lin, H. W., Lin, R. M., & Tho, P. D. (2019). Effect of peer interaction among online learning community on learning engagement and achievement. International Journal of Distance Education Technologies (IJDET), 17 (1), 66–77.

Littell, J. H., Corcoran, J., & Pillai, V. (2008). Systematic reviews and meta-analysis . Oxford University.

*Liu, K. P., Tai, S. J. D., & Liu, C. C. (2018). Enhancing language learning through creation: the effect of digital storytelling on student learning motivation and performance in a school English course. Educational Technology Research and Development, 66 (4), 913–935. https://doi.org/10.1007/s11423-018-9592-z

Machtmes, K., & Asher, J. W. (2000). A meta-analysis of the effectiveness of telecourses in distance education. American Journal of Distance Education, 14 (1), 27–46. https://doi.org/10.1080/08923640009527043

Makowski, D., Piraux, F., & Brun, F. (2019). From experimental network to meta-analysis: Methods and applications with R for agronomic and environmental sciences. Dordrecht: Springer. https://doi.org/10.1007/978-94-024_1696-1

* Meyers, C., Molefe, A., & Brandt, C. (2015). The Impact of the" Enhancing Missouri's Instructional Networked Teaching Strategies"(eMINTS) Program on Student Achievement, 21st-Century Skills, and Academic Engagement--Second-Year Results . Society for Research on Educational Effectiveness. Retrieved on the 14 th November, 2020 from https://files.eric.ed.gov/fulltext/ED562508.pdf

OECD. (2020). ‘A framework to guide an education response to the COVID-19 Pandemic of 2020 ’. https://doi.org/10.26524/royal.37.6

Pecoraro, V. (2018). Appraising evidence . In G. Biondi-Zoccai (Ed.), Diagnostic meta-analysis: A useful tool for clinical decision-making (pp. 99–114). Cham, Switzerland: Springer. https://doi.org/10.1007/978-3-319-78966-8_9

Pigott, T. (2012). Advances in meta-analysis . Springer.

Pillay, H. , Irving, K., & Tones, M. (2007). Validation of the diagnostic tool for assessing Tertiary students’ readiness for online learning. Higher Education Research & Development, 26 (2), 217–234. https://doi.org/10.1080/07294360701310821

Prestiadi, D., Zulkarnain, W., & Sumarsono, R. B. (2019). Visionary leadership in total quality management: efforts to improve the quality of education in the industrial revolution 4.0. In the 4th International Conference on Education and Management (COEMA 2019). Atlantis Press

Poole, D. M. (2000). Student participation in a discussion-oriented online course: a case study. Journal of Research on Computing in Education, 33 (2), 162–177. https://doi.org/10.1080/08886504.2000.10782307

Rahayu, F. S., Budiyanto, D., & Palyama, D. (2017). Analisis penerimaan e-learning menggunakan technology acceptance model (Tam)(Studi Kasus: Universitas Atma Jaya Yogyakarta). Jurnal Terapan Teknologi Informasi, 1 (2), 87–98.

Rasmussen, R. C. (2003). The quantity and quality of human interaction in a synchronous blended learning environment . Brigham Young University Press.

*Ravenel, J., T. Lambeth, D., & Spires, B. (2014). Effects of computer-based programs on mathematical achievement scores for fourth-grade students. i-manager’s Journal on School Educational Technology, 10 (1), 8–21. https://doi.org/10.26634/jsch.10.1.2830

Rolisca, R. U. C., & Achadiyah, B. N. (2014). Pengembangan media evaluasi pembelajaran dalam bentuk online berbasis e-learning menggunakan software wondershare quiz creator dalam mata pelajaran akuntansi SMA Brawijaya Smart School (BSS). Jurnal Pendidikan Akuntansi Indonesia, 12(2).

Sitzmann, T., Kraiger, K., Stewart, D., & Wisher, R. (2006). The comparative effective- ness of Web-based and classroom instruction: A meta-analysis . Personnel Psychology, 59 (3), 623–664. https://doi.org/10.1111/j.1744-6570.2006.00049.x

Stewart, D. W., & Kamins, M. A. (2001). Developing a coding scheme and coding study reports. In M. W. Lipsey & D. B. Wilson (Eds.), Practical meta­analysis: Applied social research methods series (Vol. 49, pp. 73–90). Sage.

Swan, K. (2007). Research on online learning. Journal of Asynchronous Learning Networks, 11 (1), 55–59.

*Sung, H. Y., Hwang, G. J., & Chang, Y. C. (2016). Development of a mobile learning system based on a collaborative problem-posing strategy. Interactive Learning Environments, 24 (3), 456–471. https://doi.org/10.1080/10494820.2013.867889

Tsagris, M., & Fragkos, K. C. (2018). Meta-analyses of clinical trials versus diagnostic test accuracy studies. In G. Biondi-Zoccai (Ed.), Diagnostic meta-analysis: A useful tool for clinical decision-making (pp. 31–42). Cham, Switzerland: Springer. https://doi.org/10.1007/978-3-319-78966-8_4

UNESCO. (2020, Match 13). COVID-19 educational disruption and response. Retrieved on the 14 th November 2020 from https://en.unesco.org/themes/education-emergencies/ coronavirus-school-closures

Usta, E. (2011a). The effect of web-based learning environments on attitudes of students regarding computer and internet. Procedia-Social and Behavioral Sciences, 28 (262–269), 1. https://doi.org/10.1016/j.sbspro.2011.11.051

Usta, E. (2011b). The examination of online self-regulated learning skills in web-based learning environments in terms of different variables. Turkish Online Journal of Educational Technology-TOJET, 10 (3), 278–286. Retrieved on the 14th November 2020 from https://files.eric.ed.gov/fulltext/EJ944994.pdf

Vrasidas, C. & MsIsaac, M. S. (2000). Principles of pedagogy and evaluation for web-based learning. Educational Media International, 37 (2), 105–111. https://doi.org/10.1080/095239800410405

*Wang, C. H., & Chen, C. P. (2013). Effects of facebook tutoring on learning english as a second language. Proceedings of the International Conference e-Learning 2013, (2009), 135–142. Retrieved on the 15th November 2020 from https://files.eric.ed.gov/fulltext/ED562299.pdf

Wei, H. C., & Chou, C. (2020). Online learning performance and satisfaction: Do perceptions and readiness matter? Distance Education, 41 (1), 48–69.

*Yu, F. Y. (2019). The learning potential of online student-constructed tests with citing peer-generated questions. Interactive Learning Environments, 27 (2), 226–241. https://doi.org/10.1080/10494820.2018.1458040

*Yu, F. Y., & Chen, Y. J. (2014). Effects of student-generated questions as the source of online drill-and-practice activities on learning . British Journal of Educational Technology, 45 (2), 316–329. https://doi.org/10.1111/bjet.12036

*Yu, F. Y., & Pan, K. J. (2014). The effects of student question-generation with online prompts on learning. Educational Technology and Society, 17 (3), 267–279. Retrieved on the 15th November 2020 from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.565.643&rep=rep1&type=pdf

*Yu, W. F., She, H. C., & Lee, Y. M. (2010). The effects of web-based/non-web-based problem-solving instruction and high/low achievement on students’ problem-solving ability and biology achievement. Innovations in Education and Teaching International, 47 (2), 187–199. https://doi.org/10.1080/14703291003718927

Zhao, Y., Lei, J., Yan, B, Lai, C., & Tan, S. (2005). A practical analysis of research on the effectiveness of distance education. Teachers College Record, 107 (8). https://doi.org/10.1111/j.1467-9620.2005.00544.x

*Zhong, B., Wang, Q., Chen, J., & Li, Y. (2017). Investigating the period of switching roles in pair programming in a primary school. Educational Technology and Society, 20 (3), 220–233. Retrieved on the 15th November 2020 from https://repository.nie.edu.sg/bitstream/10497/18946/1/ETS-20-3-220.pdf

Download references

Author information

Authors and affiliations.

Primary Education, Ministry of Turkish National Education, Mersin, Turkey

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Hakan Ulum .

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Ulum, H. The effects of online education on academic success: A meta-analysis study. Educ Inf Technol 27 , 429–450 (2022). https://doi.org/10.1007/s10639-021-10740-8

Download citation

Received : 06 December 2020

Accepted : 30 August 2021

Published : 06 September 2021

Issue Date : January 2022

DOI : https://doi.org/10.1007/s10639-021-10740-8

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Online education
• Student achievement
• Academic success
• Meta-analysis
• Find a journal
• Publish with us
• Track your research

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning: A social network analysis

Roles Data curation, Formal analysis, Methodology, Writing – review & editing

¶ ‡ JZ and YD are contributed equally to this work as first authors.

Affiliation School of Educational Information Technology, South China Normal University, Guangzhou, Guangdong, China

Roles Data curation, Formal analysis, Methodology, Writing – original draft

Affiliations School of Educational Information Technology, South China Normal University, Guangzhou, Guangdong, China, Hangzhou Zhongce Vocational School Qiantang, Hangzhou, Zhejiang, China

Roles Data curation, Writing – original draft

Roles Data curation

Roles Writing – original draft

Affiliation Faculty of Education, Shenzhen University, Shenzhen, Guangdong, China

Roles Conceptualization, Supervision, Writing – review & editing

* E-mail: [email protected] (JH); [email protected] (YZ)

• Junyi Zhang, 
• Yigang Ding, 
• Xinru Yang, 
• Jinping Zhong, 
• XinXin Qiu, 
• Zhishan Zou, 
• Yujie Xu, 
• Xiunan Jin, 
• Xiaomin Wu, 

• Published: August 23, 2022
• https://doi.org/10.1371/journal.pone.0273016
• Reader Comments

The COVID-19 outbreak brought online learning to the forefront of education. Scholars have conducted many studies on online learning during the pandemic, but only a few have performed quantitative comparative analyses of students’ online learning behavior before and after the outbreak. We collected review data from China’s massive open online course platform called icourse.163 and performed social network analysis on 15 courses to explore courses’ interaction characteristics before, during, and after the COVID-19 pan-demic. Specifically, we focused on the following aspects: (1) variations in the scale of online learning amid COVID-19; (2a) the characteristics of online learning interaction during the pandemic; (2b) the characteristics of online learning interaction after the pandemic; and (3) differences in the interaction characteristics of social science courses and natural science courses. Results revealed that only a small number of courses witnessed an uptick in online interaction, suggesting that the pandemic’s role in promoting the scale of courses was not significant. During the pandemic, online learning interaction became more frequent among course network members whose interaction scale increased. After the pandemic, although the scale of interaction declined, online learning interaction became more effective. The scale and level of interaction in Electrodynamics (a natural science course) and Economics (a social science course) both rose during the pan-demic. However, long after the pandemic, the Economics course sustained online interaction whereas interaction in the Electrodynamics course steadily declined. This discrepancy could be due to the unique characteristics of natural science courses and social science courses.

Citation: Zhang J, Ding Y, Yang X, Zhong J, Qiu X, Zou Z, et al. (2022) COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning: A social network analysis. PLoS ONE 17(8): e0273016. https://doi.org/10.1371/journal.pone.0273016

Editor: Heng Luo, Central China Normal University, CHINA

Received: April 20, 2022; Accepted: July 29, 2022; Published: August 23, 2022

Copyright: © 2022 Zhang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The data underlying the results presented in the study were downloaded from https://www.icourse163.org/ and are now shared fully on Github ( https://github.com/zjyzhangjunyi/dataset-from-icourse163-for-SNA ). These data have no private information and can be used for academic research free of charge.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

1. Introduction

The development of the mobile internet has spurred rapid advances in online learning, offering novel prospects for teaching and learning and a learning experience completely different from traditional instruction. Online learning harnesses the advantages of network technology and multimedia technology to transcend the boundaries of conventional education [ 1 ]. Online courses have become a popular learning mode owing to their flexibility and openness. During online learning, teachers and students are in different physical locations but interact in multiple ways (e.g., via online forum discussions and asynchronous group discussions). An analysis of online learning therefore calls for attention to students’ participation. Alqurashi [ 2 ] defined interaction in online learning as the process of constructing meaningful information and thought exchanges between more than two people; such interaction typically occurs between teachers and learners, learners and learners, and the course content and learners.

Massive open online courses (MOOCs), a 21st-century teaching mode, have greatly influenced global education. Data released by China’s Ministry of Education in 2020 show that the country ranks first globally in the number and scale of higher education MOOCs. The COVID-19 outbreak has further propelled this learning mode, with universities being urged to leverage MOOCs and other online resource platforms to respond to government’s “School’s Out, But Class’s On” policy [ 3 ]. Besides MOOCs, to reduce in-person gatherings and curb the spread of COVID-19, various online learning methods have since become ubiquitous [ 4 ]. Though Lederman asserted that the COVID-19 outbreak has positioned online learning technologies as the best way for teachers and students to obtain satisfactory learning experiences [ 5 ], it remains unclear whether the COVID-19 pandemic has encouraged interaction in online learning, as interactions between students and others play key roles in academic performance and largely determine the quality of learning experiences [ 6 ]. Similarly, it is also unclear what impact the COVID-19 pandemic has had on the scale of online learning.

Social constructivism paints learning as a social phenomenon. As such, analyzing the social structures or patterns that emerge during the learning process can shed light on learning-based interaction [ 7 ]. Social network analysis helps to explain how a social network, rooted in interactions between learners and their peers, guides individuals’ behavior, emotions, and outcomes. This analytical approach is especially useful for evaluating interactive relationships between network members [ 8 ]. Mohammed cited social network analysis (SNA) as a method that can provide timely information about students, learning communities and interactive networks. SNA has been applied in numerous fields, including education, to identify the number and characteristics of interelement relationships. For example, Lee et al. also used SNA to explore the effects of blogs on peer relationships [ 7 ]. Therefore, adopting SNA to examine interactions in online learning communities during the COVID-19 pandemic can uncover potential issues with this online learning model.

Taking China’s icourse.163 MOOC platform as an example, we chose 15 courses with a large number of participants for SNA, focusing on learners’ interaction characteristics before, during, and after the COVID-19 outbreak. We visually assessed changes in the scale of network interaction before, during, and after the outbreak along with the characteristics of interaction in Gephi. Examining students’ interactions in different courses revealed distinct interactive network characteristics, the pandemic’s impact on online courses, and relevant suggestions. Findings are expected to promote effective interaction and deep learning among students in addition to serving as a reference for the development of other online learning communities.

2. Literature review and research questions

Interaction is deemed as central to the educational experience and is a major focus of research on online learning. Moore began to study the problem of interaction in distance education as early as 1989. He defined three core types of interaction: student–teacher, student–content, and student–student [ 9 ]. Lear et al. [ 10 ] described an interactivity/ community-process model of distance education: they specifically discussed the relationships between interactivity, community awareness, and engaging learners and found interactivity and community awareness to be correlated with learner engagement. Zulfikar et al. [ 11 ] suggested that discussions initiated by the students encourage more students’ engagement than discussions initiated by the instructors. It is most important to afford learners opportunities to interact purposefully with teachers, and improving the quality of learner interaction is crucial to fostering profound learning [ 12 ]. Interaction is an important way for learners to communicate and share information, and a key factor in the quality of online learning [ 13 ].

Timely feedback is the main component of online learning interaction. Woo and Reeves discovered that students often become frustrated when they fail to receive prompt feedback [ 14 ]. Shelley et al. conducted a three-year study of graduate and undergraduate students’ satisfaction with online learning at universities and found that interaction with educators and students is the main factor affecting satisfaction [ 15 ]. Teachers therefore need to provide students with scoring justification, support, and constructive criticism during online learning. Some researchers examined online learning during the COVID-19 pandemic. They found that most students preferred face-to-face learning rather than online learning due to obstacles faced online, such as a lack of motivation, limited teacher-student interaction, and a sense of isolation when learning in different times and spaces [ 16 , 17 ]. However, it can be reduced by enhancing the online interaction between teachers and students [ 18 ].

Research showed that interactions contributed to maintaining students’ motivation to continue learning [ 19 ]. Baber argued that interaction played a key role in students’ academic performance and influenced the quality of the online learning experience [ 20 ]. Hodges et al. maintained that well-designed online instruction can lead to unique teaching experiences [ 21 ]. Banna et al. mentioned that using discussion boards, chat sessions, blogs, wikis, and other tools could promote student interaction and improve participation in online courses [ 22 ]. During the COVID-19 pandemic, Mahmood proposed a series of teaching strategies suitable for distance learning to improve its effectiveness [ 23 ]. Lapitan et al. devised an online strategy to ease the transition from traditional face-to-face instruction to online learning [ 24 ]. The preceding discussion suggests that online learning goes beyond simply providing learning resources; teachers should ideally design real-life activities to give learners more opportunities to participate.

As mentioned, COVID-19 has driven many scholars to explore the online learning environment. However, most have ignored the uniqueness of online learning during this time and have rarely compared pre- and post-pandemic online learning interaction. Taking China’s icourse.163 MOOC platform as an example, we chose 15 courses with a large number of participants for SNA, centering on student interaction before and after the pandemic. Gephi was used to visually analyze changes in the scale and characteristics of network interaction. The following questions were of particular interest:

• (1) Can the COVID-19 pandemic promote the expansion of online learning?
• (2a) What are the characteristics of online learning interaction during the pandemic?
• (2b) What are the characteristics of online learning interaction after the pandemic?
• (3) How do interaction characteristics differ between social science courses and natural science courses?

3. Methodology

3.1 research context.

We selected several courses with a large number of participants and extensive online interaction among hundreds of courses on the icourse.163 MOOC platform. These courses had been offered on the platform for at least three semesters, covering three periods (i.e., before, during, and after the COVID-19 outbreak). To eliminate the effects of shifts in irrelevant variables (e.g., course teaching activities), we chose several courses with similar teaching activities and compared them on multiple dimensions. All course content was taught online. The teachers of each course posted discussion threads related to learning topics; students were expected to reply via comments. Learners could exchange ideas freely in their responses in addition to asking questions and sharing their learning experiences. Teachers could answer students’ questions as well. Conversations in the comment area could partly compensate for a relative absence of online classroom interaction. Teacher–student interaction is conducive to the formation of a social network structure and enabled us to examine teachers’ and students’ learning behavior through SNA. The comment areas in these courses were intended for learners to construct knowledge via reciprocal communication. Meanwhile, by answering students’ questions, teachers could encourage them to reflect on their learning progress. These courses’ successive terms also spanned several phases of COVID-19, allowing us to ascertain the pandemic’s impact on online learning.

3.2 Data collection and preprocessing

To avoid interference from invalid or unclear data, the following criteria were applied to select representative courses: (1) generality (i.e., public courses and professional courses were chosen from different schools across China); (2) time validity (i.e., courses were held before during, and after the pandemic); and (3) notability (i.e., each course had at least 2,000 participants). We ultimately chose 15 courses across the social sciences and natural sciences (see Table 1 ). The coding is used to represent the course name.

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0273016.t001

To discern courses’ evolution during the pandemic, we gathered data on three terms before, during, and after the COVID-19 outbreak in addition to obtaining data from two terms completed well before the pandemic and long after. Our final dataset comprised five sets of interactive data. Finally, we collected about 120,000 comments for SNA. Because each course had a different start time—in line with fluctuations in the number of confirmed COVID-19 cases in China and the opening dates of most colleges and universities—we divided our sample into five phases: well before the pandemic (Phase I); before the pandemic (Phase Ⅱ); during the pandemic (Phase Ⅲ); after the pandemic (Phase Ⅳ); and long after the pandemic (Phase Ⅴ). We sought to preserve consistent time spans to balance the amount of data in each period ( Fig 1 ).

https://doi.org/10.1371/journal.pone.0273016.g001

3.3 Instrumentation

Participants’ comments and “thumbs-up” behavior data were converted into a network structure and compared using social network analysis (SNA). Network analysis, according to M’Chirgui, is an effective tool for clarifying network relationships by employing sophisticated techniques [ 25 ]. Specifically, SNA can help explain the underlying relationships among team members and provide a better understanding of their internal processes. Yang and Tang used SNA to discuss the relationship between team structure and team performance [ 26 ]. Golbeck argued that SNA could improve the understanding of students’ learning processes and reveal learners’ and teachers’ role dynamics [ 27 ].

To analyze Question (1), the number of nodes and diameter in the generated network were deemed as indicators of changes in network size. Social networks are typically represented as graphs with nodes and degrees, and node count indicates the sample size [ 15 ]. Wellman et al. proposed that the larger the network scale, the greater the number of network members providing emotional support, goods, services, and companionship [ 28 ]. Jan’s study measured the network size by counting the nodes which represented students, lecturers, and tutors [ 29 ]. Similarly, network nodes in the present study indicated how many learners and teachers participated in the course, with more nodes indicating more participants. Furthermore, we investigated the network diameter, a structural feature of social networks, which is a common metric for measuring network size in SNA [ 30 ]. The network diameter refers to the longest path between any two nodes in the network. There has been evidence that a larger network diameter leads to greater spread of behavior [ 31 ]. Likewise, Gašević et al. found that larger networks were more likely to spread innovative ideas about educational technology when analyzing MOOC-related research citations [ 32 ]. Therefore, we employed node count and network diameter to measure the network’s spatial size and further explore the expansion characteristic of online courses. Brief introduction of these indicators can be summarized in Table 2 .

https://doi.org/10.1371/journal.pone.0273016.t002

To address Question (2), a list of interactive analysis metrics in SNA were introduced to scrutinize learners’ interaction characteristics in online learning during and after the pandemic, as shown below:

• (1) The average degree reflects the density of the network by calculating the average number of connections for each node. As Rong and Xu suggested, the average degree of a network indicates how active its participants are [ 33 ]. According to Hu, a higher average degree implies that more students are interacting directly with each other in a learning context [ 34 ]. The present study inherited the concept of the average degree from these previous studies: the higher the average degree, the more frequent the interaction between individuals in the network.
• (2) Essentially, a weighted average degree in a network is calculated by multiplying each degree by its respective weight, and then taking the average. Bydžovská took the strength of the relationship into account when determining the weighted average degree [ 35 ]. By calculating friendship’s weighted value, Maroulis assessed peer achievement within a small-school reform [ 36 ]. Accordingly, we considered the number of interactions as the weight of the degree, with a higher average degree indicating more active interaction among learners.
• (3) Network density is the ratio between actual connections and potential connections in a network. The more connections group members have with each other, the higher the network density. In SNA, network density is similar to group cohesion, i.e., a network of more strong relationships is more cohesive [ 37 ]. Network density also reflects how much all members are connected together [ 38 ]. Therefore, we adopted network density to indicate the closeness among network members. Higher network density indicates more frequent interaction and closer communication among students.
• (4) Clustering coefficient describes local network attributes and indicates that two nodes in the network could be connected through adjacent nodes. The clustering coefficient measures users’ tendency to gather (cluster) with others in the network: the higher the clustering coefficient, the more frequently users communicate with other group members. We regarded this indicator as a reflection of the cohesiveness of the group [ 39 ].
• (5) In a network, the average path length is the average number of steps along the shortest paths between any two nodes. Oliveres has observed that when an average path length is small, the route from one node to another is shorter when graphed [ 40 ]. This is especially true in educational settings where students tend to become closer friends. So we consider that the smaller the average path length, the greater the possibility of interaction between individuals in the network.
• (6) A network with a large number of nodes, but whose average path length is surprisingly small, is known as the small-world effect [ 41 ]. A higher clustering coefficient and shorter average path length are important indicators of a small-world network: a shorter average path length enables the network to spread information faster and more accurately; a higher clustering coefficient can promote frequent knowledge exchange within the group while boosting the timeliness and accuracy of knowledge dissemination [ 42 ]. Brief introduction of these indicators can be summarized in Table 3 .

https://doi.org/10.1371/journal.pone.0273016.t003

To analyze Question 3, we used the concept of closeness centrality, which determines how close a vertex is to others in the network. As Opsahl et al. explained, closeness centrality reveals how closely actors are coupled with their entire social network [ 43 ]. In order to analyze social network-based engineering education, Putnik et al. examined closeness centrality and found that it was significantly correlated with grades [ 38 ]. We used closeness centrality to measure the position of an individual in the network. Brief introduction of these indicators can be summarized in Table 4 .

https://doi.org/10.1371/journal.pone.0273016.t004

3.4 Ethics statement

This study was approved by the Academic Committee Office (ACO) of South China Normal University ( http://fzghb.scnu.edu.cn/ ), Guangzhou, China. Research data were collected from the open platform and analyzed anonymously. There are thus no privacy issues involved in this study.

4.1 COVID-19’s role in promoting the scale of online courses was not as important as expected

As shown in Fig 2 , the number of course participants and nodes are closely correlated with the pandemic’s trajectory. Because the number of participants in each course varied widely, we normalized the number of participants and nodes to more conveniently visualize course trends. Fig 2 depicts changes in the chosen courses’ number of participants and nodes before the pandemic (Phase II), during the pandemic (Phase III), and after the pandemic (Phase IV). The number of participants in most courses during the pandemic exceeded those before and after the pandemic. But the number of people who participate in interaction in some courses did not increase.

https://doi.org/10.1371/journal.pone.0273016.g002

In order to better analyze the trend of interaction scale in online courses before, during, and after the pandemic, the selected courses were categorized according to their scale change. When the number of participants increased (decreased) beyond 20% (statistical experience) and the diameter also increased (decreased), the course scale was determined to have increased (decreased); otherwise, no significant change was identified in the course’s interaction scale. Courses were subsequently divided into three categories: increased interaction scale, decreased interaction scale, and no significant change. Results appear in Table 5 .

https://doi.org/10.1371/journal.pone.0273016.t005

From before the pandemic until it broke out, the interaction scale of five courses increased, accounting for 33.3% of the full sample; one course’s interaction scale declined, accounting for 6.7%. The interaction scale of nine courses decreased, accounting for 60%. The pandemic’s role in promoting online courses thus was not as important as anticipated, and most courses’ interaction scale did not change significantly throughout.

No courses displayed growing interaction scale after the pandemic: the interaction scale of nine courses fell, accounting for 60%; and the interaction scale of six courses did not shift significantly, accounting for 40%. Courses with an increased scale of interaction during the pandemic did not maintain an upward trend. On the contrary, the improvement in the pandemic caused learners’ enthusiasm for online learning to wane. We next analyzed several interaction metrics to further explore course interaction during different pandemic periods.

4.2 Characteristics of online learning interaction amid COVID-19

4.2.1 during the covid-19 pandemic, online learning interaction in some courses became more active..

Changes in course indicators with the growing interaction scale during the pandemic are presented in Fig 3 , including SS5, SS6, NS1, NS3, and NS8. The horizontal ordinate indicates the number of courses, with red color representing the rise of the indicator value on the vertical ordinate and blue representing the decline.

https://doi.org/10.1371/journal.pone.0273016.g003

Specifically: (1) The average degree and weighted average degree of the five course networks demonstrated an upward trend. The emergence of the pandemic promoted students’ enthusiasm; learners were more active in the interactive network. (2) Fig 3 shows that 3 courses had increased network density and 2 courses had decreased. The higher the network density, the more communication within the team. Even though the pandemic accelerated the interaction scale and frequency, the tightness between learners in some courses did not improve. (3) The clustering coefficient of social science courses rose whereas the clustering coefficient and small-world property of natural science courses fell. The higher the clustering coefficient and the small-world property, the better the relationship between adjacent nodes and the higher the cohesion [ 39 ]. (4) Most courses’ average path length increased as the interaction scale increased. However, when the average path length grew, adverse effects could manifest: communication between learners might be limited to a small group without multi-directional interaction.

When the pandemic emerged, the only declining network scale belonged to a natural science course (NS2). The change in each course index is pictured in Fig 4 . The abscissa indicates the size of the value, with larger values to the right. The red dot indicates the index value before the pandemic; the blue dot indicates its value during the pandemic. If the blue dot is to the right of the red dot, then the value of the index increased; otherwise, the index value declined. Only the weighted average degree of the course network increased. The average degree, network density decreased, indicating that network members were not active and that learners’ interaction degree and communication frequency lessened. Despite reduced learner interaction, the average path length was small and the connectivity between learners was adequate.

https://doi.org/10.1371/journal.pone.0273016.g004

4.2.2 After the COVID-19 pandemic, the scale decreased rapidly, but most course interaction was more effective.

Fig 5 shows the changes in various courses’ interaction indicators after the pandemic, including SS1, SS2, SS3, SS6, SS7, NS2, NS3, NS7, and NS8.

https://doi.org/10.1371/journal.pone.0273016.g005

Specifically: (1) The average degree and weighted average degree of most course networks decreased. The scope and intensity of interaction among network members declined rapidly, as did learners’ enthusiasm for communication. (2) The network density of seven courses also fell, indicating weaker connections between learners in most courses. (3) In addition, the clustering coefficient and small-world property of most course networks decreased, suggesting little possibility of small groups in the network. The scope of interaction between learners was not limited to a specific space, and the interaction objects had no significant tendencies. (4) Although the scale of course interaction became smaller in this phase, the average path length of members’ social networks shortened in nine courses. Its shorter average path length would expedite the spread of information within the network as well as communication and sharing among network members.

Fig 6 displays the evolution of course interaction indicators without significant changes in interaction scale after the pandemic, including SS4, SS5, NS1, NS4, NS5, and NS6.

https://doi.org/10.1371/journal.pone.0273016.g006

Specifically: (1) Some course members’ social networks exhibited an increase in the average and weighted average. In these cases, even though the course network’s scale did not continue to increase, communication among network members rose and interaction became more frequent and deeper than before. (2) Network density and average path length are indicators of social network density. The greater the network density, the denser the social network; the shorter the average path length, the more concentrated the communication among network members. However, at this phase, the average path length and network density in most courses had increased. Yet the network density remained small despite having risen ( Table 6 ). Even with more frequent learner interaction, connections remained distant and the social network was comparatively sparse.

https://doi.org/10.1371/journal.pone.0273016.t006

In summary, the scale of interaction did not change significantly overall. Nonetheless, some course members’ frequency and extent of interaction increased, and the relationships between network members became closer as well. In the study, we found it interesting that the interaction scale of Economics (a social science course) course and Electrodynamics (a natural science course) course expanded rapidly during the pandemic and retained their interaction scale thereafter. We next assessed these two courses to determine whether their level of interaction persisted after the pandemic.

4.3 Analyses of natural science courses and social science courses

4.3.1 analyses of the interaction characteristics of economics and electrodynamics..

Economics and Electrodynamics are social science courses and natural science courses, respectively. Members’ interaction within these courses was similar: the interaction scale increased significantly when COVID-19 broke out (Phase Ⅲ), and no significant changes emerged after the pandemic (Phase Ⅴ). We hence focused on course interaction long after the outbreak (Phase V) and compared changes across multiple indicators, as listed in Table 7 .

https://doi.org/10.1371/journal.pone.0273016.t007

As the pandemic continued to improve, the number of participants and the diameter long after the outbreak (Phase V) each declined for Economics compared with after the pandemic (Phase IV). The interaction scale decreased, but the interaction between learners was much deeper. Specifically: (1) The weighted average degree, network density, clustering coefficient, and small-world property each reflected upward trends. The pandemic therefore exerted a strong impact on this course. Interaction was well maintained even after the pandemic. The smaller network scale promoted members’ interaction and communication. (2) Compared with after the pandemic (Phase IV), members’ network density increased significantly, showing that relationships between learners were closer and that cohesion was improving. (3) At the same time, as the clustering coefficient and small-world property grew, network members demonstrated strong small-group characteristics: the communication between them was deepening and their enthusiasm for interaction was higher. (4) Long after the COVID-19 outbreak (Phase V), the average path length was reduced compared with previous terms, knowledge flowed more quickly among network members, and the degree of interaction gradually deepened.

The average degree, weighted average degree, network density, clustering coefficient, and small-world property of Electrodynamics all decreased long after the COVID-19 outbreak (Phase V) and were lower than during the outbreak (Phase Ⅲ). The level of learner interaction therefore gradually declined long after the outbreak (Phase V), and connections between learners were no longer active. Although the pandemic increased course members’ extent of interaction, this rise was merely temporary: students’ enthusiasm for learning waned rapidly and their interaction decreased after the pandemic (Phase IV). To further analyze the interaction characteristics of course members in Economics and Electrodynamics, we evaluated the closeness centrality of their social networks, as shown in section 4.3.2.

4.3.2 Analysis of the closeness centrality of Economics and Electrodynamics.

The change in the closeness centrality of social networks in Economics was small, and no sharp upward trend appeared during the pandemic outbreak, as shown in Fig 7 . The emergence of COVID-19 apparently fostered learners’ interaction in Economics albeit without a significant impact. The closeness centrality changed in Electrodynamics varied from that of Economics: upon the COVID-19 outbreak, closeness centrality was significantly different from other semesters. Communication between learners was closer and interaction was more effective. Electrodynamics course members’ social network proximity decreased rapidly after the pandemic. Learners’ communication lessened. In general, Economics course showed better interaction before the outbreak and was less affected by the pandemic; Electrodynamics course was more affected by the pandemic and showed different interaction characteristics at different periods of the pandemic.

(Note: "****" indicates the significant distinction in closeness centrality between the two periods, otherwise no significant distinction).

https://doi.org/10.1371/journal.pone.0273016.g007

5. Discussion

We referred to discussion forums from several courses on the icourse.163 MOOC platform to compare online learning before, during, and after the COVID-19 pandemic via SNA and to delineate the pandemic’s effects on online courses. Only 33.3% of courses in our sample increased in terms of interaction during the pandemic; the scale of interaction did not rise in any courses thereafter. When the courses scale rose, the scope and frequency of interaction showed upward trends during the pandemic; and the clustering coefficient of natural science courses and social science courses differed: the coefficient for social science courses tended to rise whereas that for natural science courses generally declined. When the pandemic broke out, the interaction scale of a single natural science course decreased along with its interaction scope and frequency. The amount of interaction in most courses shrank rapidly during the pandemic and network members were not as active as they had been before. However, after the pandemic, some courses saw declining interaction but greater communication between members; interaction also became more frequent and deeper than before.

5.1 During the COVID-19 pandemic, the scale of interaction increased in only a few courses

The pandemic outbreak led to a rapid increase in the number of participants in most courses; however, the change in network scale was not significant. The scale of online interaction expanded swiftly in only a few courses; in others, the scale either did not change significantly or displayed a downward trend. After the pandemic, the interaction scale in most courses decreased quickly; the same pattern applied to communication between network members. Learners’ enthusiasm for online interaction reduced as the circumstances of the pandemic improved—potentially because, during the pandemic, China’s Ministry of Education declared “School’s Out, But Class’s On” policy. Major colleges and universities were encouraged to use the Internet and informational resources to provide learning support, hence the sudden increase in the number of participants and interaction in online courses [ 46 ]. After the pandemic, students’ enthusiasm for online learning gradually weakened, presumably due to easing of the pandemic [ 47 ]. More activities also transitioned from online to offline, which tempered learners’ online discussion. Research has shown that long-term online learning can even bore students [ 48 ].

Most courses’ interaction scale decreased significantly after the pandemic. First, teachers and students occupied separate spaces during the outbreak, had few opportunities for mutual cooperation and friendship, and lacked a sense of belonging [ 49 ]. Students’ enthusiasm for learning dissipated over time [ 50 ]. Second, some teachers were especially concerned about adapting in-person instructional materials for digital platforms; their pedagogical methods were ineffective, and they did not provide learning activities germane to student interaction [ 51 ]. Third, although teachers and students in remote areas were actively engaged in online learning, some students could not continue to participate in distance learning due to inadequate technology later in the outbreak [ 52 ].

5.2 Characteristics of online learning interaction during and after the COVID-19 pandemic

5.2.1 during the covid-19 pandemic, online interaction in most courses did not change significantly..

The interaction scale of only a few courses increased during the pandemic. The interaction scope and frequency of these courses climbed as well. Yet even as the degree of network interaction rose, course network density did not expand in all cases. The pandemic sparked a surge in the number of online learners and a rapid increase in network scale, but students found it difficult to interact with all learners. Yau pointed out that a greater network scale did not enrich the range of interaction between individuals; rather, the number of individuals who could interact directly was limited [ 53 ]. The internet facilitates interpersonal communication. However, not everyone has the time or ability to establish close ties with others [ 54 ].

In addition, social science courses and natural science courses in our sample revealed disparate trends in this regard: the clustering coefficient of social science courses increased and that of natural science courses decreased. Social science courses usually employ learning approaches distinct from those in natural science courses [ 55 ]. Social science courses emphasize critical and innovative thinking along with personal expression [ 56 ]. Natural science courses focus on practical skills, methods, and principles [ 57 ]. Therefore, the content of social science courses can spur large-scale discussion among learners. Some course evaluations indicated that the course content design was suboptimal as well: teachers paid close attention to knowledge transmission and much less to piquing students’ interest in learning. In addition, the thread topics that teachers posted were scarcely diversified and teachers’ questions lacked openness. These attributes could not spark active discussion among learners.

5.2.2 Online learning interaction declined after the COVID-19 pandemic.

Most courses’ interaction scale and intensity decreased rapidly after the pandemic, but some did not change. Courses with a larger network scale did not continue to expand after the outbreak, and students’ enthusiasm for learning paled. The pandemic’s reduced severity also influenced the number of participants in online courses. Meanwhile, restored school order moved many learning activities from virtual to in-person spaces. Face-to-face learning has gradually replaced online learning, resulting in lower enrollment and less interaction in online courses. Prolonged online courses could have also led students to feel lonely and to lack a sense of belonging [ 58 ].

The scale of interaction in some courses did not change substantially after the pandemic yet learners’ connections became tighter. We hence recommend that teachers seize pandemic-related opportunities to design suitable activities. Additionally, instructors should promote student-teacher and student-student interaction, encourage students to actively participate online, and generally intensify the impact of online learning.

5.3 What are the characteristics of interaction in social science courses and natural science courses?

The level of interaction in Economics (a social science course) was significantly higher than that in Electrodynamics (a natural science course), and the small-world property in Economics increased as well. To boost online courses’ learning-related impacts, teachers can divide groups of learners based on the clustering coefficient and the average path length. Small groups of students may benefit teachers in several ways: to participate actively in activities intended to expand students’ knowledge, and to serve as key actors in these small groups. Cultivating students’ keenness to participate in class activities and self-management can also help teachers guide learner interaction and foster deep knowledge construction.

As evidenced by comments posted in the Electrodynamics course, we observed less interaction between students. Teachers also rarely urged students to contribute to conversations. These trends may have arisen because teachers and students were in different spaces. Teachers might have struggled to discern students’ interaction status. Teachers could also have failed to intervene in time, to design online learning activities that piqued learners’ interest, and to employ sound interactive theme planning and guidance. Teachers are often active in traditional classroom settings. Their roles are comparatively weakened online, such that they possess less control over instruction [ 59 ]. Online instruction also requires a stronger hand in learning: teachers should play a leading role in regulating network members’ interactive communication [ 60 ]. Teachers can guide learners to participate, help learners establish social networks, and heighten students’ interest in learning [ 61 ]. Teachers should attend to core members in online learning while also considering edge members; by doing so, all network members can be driven to share their knowledge and become more engaged. Finally, teachers and assistant teachers should help learners develop knowledge, exchange topic-related ideas, pose relevant questions during course discussions, and craft activities that enable learners to interact online [ 62 ]. These tactics can improve the effectiveness of online learning.

As described, network members displayed distinct interaction behavior in Economics and Electrodynamics courses. First, these courses varied in their difficulty: the social science course seemed easier to understand and focused on divergent thinking. Learners were often willing to express their views in comments and to ponder others’ perspectives [ 63 ]. The natural science course seemed more demanding and was oriented around logical thinking and skills [ 64 ]. Second, courses’ content differed. In general, social science courses favor the acquisition of declarative knowledge and creative knowledge compared with natural science courses. Social science courses also entertain open questions [ 65 ]. Natural science courses revolve around principle knowledge, strategic knowledge, and transfer knowledge [ 66 ]. Problems in these courses are normally more complicated than those in social science courses. Third, the indicators affecting students’ attitudes toward learning were unique. Guo et al. discovered that “teacher feedback” most strongly influenced students’ attitudes towards learning social science courses but had less impact on students in natural science courses [ 67 ]. Therefore, learners in social science courses likely expect more feedback from teachers and greater interaction with others.

6. Conclusion and future work

Our findings show that the network interaction scale of some online courses expanded during the COVID-19 pandemic. The network scale of most courses did not change significantly, demonstrating that the pandemic did not notably alter the scale of course interaction. Online learning interaction among course network members whose interaction scale increased also became more frequent during the pandemic. Once the outbreak was under control, although the scale of interaction declined, the level and scope of some courses’ interactive networks continued to rise; interaction was thus particularly effective in these cases. Overall, the pandemic appeared to have a relatively positive impact on online learning interaction. We considered a pair of courses in detail and found that Economics (a social science course) fared much better than Electrodynamics (a natural science course) in classroom interaction; learners were more willing to partake in-class activities, perhaps due to these courses’ unique characteristics. Brint et al. also came to similar conclusions [ 57 ].

This study was intended to be rigorous. Even so, several constraints can be addressed in future work. The first limitation involves our sample: we focused on a select set of courses hosted on China’s icourse.163 MOOC platform. Future studies should involve an expansive collection of courses to provide a more holistic understanding of how the pandemic has influenced online interaction. Second, we only explored the interactive relationship between learners and did not analyze interactive content. More in-depth content analysis should be carried out in subsequent research. All in all, the emergence of COVID-19 has provided a new path for online learning and has reshaped the distance learning landscape. To cope with associated challenges, educational practitioners will need to continue innovating in online instructional design, strengthen related pedagogy, optimize online learning conditions, and bolster teachers’ and students’ competence in online learning.

• View Article
• Google Scholar
• PubMed/NCBI
• 30. Serrat O. Social network analysis. Knowledge solutions: Springer; 2017. p. 39–43. https://doi.org/10.1007/978-981-10-0983-9_9
• 33. Rong Y, Xu E, editors. Strategies for the Management of the Government Affairs Microblogs in China Based on the SNA of Fifty Government Affairs Microblogs in Beijing. 14th International Conference on Service Systems and Service Management 2017.
• 34. Hu X, Chu S, editors. A comparison on using social media in a professional experience course. International Conference on Social Media and Society; 2013.
• 35. Bydžovská H. A Comparative Analysis of Techniques for Predicting Student Performance. Proceedings of the 9th International Conference on Educational Data Mining; Raleigh, NC, USA: International Educational Data Mining Society2016. p. 306–311.
• 40. Olivares D, Adesope O, Hundhausen C, et al., editors. Using social network analysis to measure the effect of learning analytics in computing education. 19th IEEE International Conference on Advanced Learning Technologies 2019.
• 41. Travers J, Milgram S. An experimental study of the small world problem. Social Networks: Elsevier; 1977. p. 179–197. https://doi.org/10.1016/B978-0-12-442450-0.50018–3
• 43. Okamoto K, Chen W, Li X-Y, editors. Ranking of closeness centrality for large-scale social networks. International workshop on frontiers in algorithmics; 2008; Springer, Berlin, Heidelberg: Springer.
• 47. Ding Y, Yang X, Zheng Y, editors. COVID-19’s Effects on the Scope, Effectiveness, and Roles of Teachers in Online Learning Based on Social Network Analysis: A Case Study. International Conference on Blended Learning; 2021: Springer.
• 64. Boys C, Brennan J., Henkel M., Kirkland J., Kogan M., Youl P. Higher Education and Preparation for Work. Jessica Kingsley Publishers. 1988. https://doi.org/10.1080/03075079612331381467

Open Access Theses and Dissertations

Thursday, April 18, 8:20am (EDT): Searching is temporarily offline. We apologize for the inconvenience and are working to bring searching back up as quickly as possible.

Advanced research and scholarship. Theses and dissertations, free to find, free to use.

Advanced search options

Browse by author name (“Author name starts with…”).

Find ETDs with:

Written in any language English Portuguese French German Spanish Swedish Lithuanian Dutch Italian Chinese Finnish Greek Published in any country US or Canada Argentina Australia Austria Belgium Bolivia Brazil Canada Chile China Colombia Czech Republic Denmark Estonia Finland France Germany Greece Hong Kong Hungary Iceland India Indonesia Ireland Italy Japan Latvia Lithuania Malaysia Mexico Netherlands New Zealand Norway Peru Portugal Russia Singapore South Africa South Korea Spain Sweden Switzerland Taiwan Thailand UK US Earliest date Latest date

Sorted by Relevance Author University Date

Only ETDs with Creative Commons licenses

Results per page: 30 60 100

October 3, 2022. OATD is dealing with a number of misbehaved crawlers and robots, and is currently taking some steps to minimize their impact on the system. This may require you to click through some security screen. Our apologies for any inconvenience.

Recent Additions

See all of this week’s new additions.

About OATD.org

OATD.org aims to be the best possible resource for finding open access graduate theses and dissertations published around the world. Metadata (information about the theses) comes from over 1100 colleges, universities, and research institutions . OATD currently indexes 7,173,081 theses and dissertations.

About OATD (our FAQ) .

Visual OATD.org

We’re happy to present several data visualizations to give an overall sense of the OATD.org collection by county of publication, language, and field of study.

You may also want to consult these sites to search for other theses:

• Google Scholar
• NDLTD , the Networked Digital Library of Theses and Dissertations. NDLTD provides information and a search engine for electronic theses and dissertations (ETDs), whether they are open access or not.
• Proquest Theses and Dissertations (PQDT), a database of dissertations and theses, whether they were published electronically or in print, and mostly available for purchase. Access to PQDT may be limited; consult your local library for access information.

RESEARCH ON ONLINE LEARNING

• February 2019
• Online Learning 11(1)

• University of Illinois Springfield

Discover the world's research

• 25+ million members
• 160+ million publication pages
• 2.3+ billion citations
• Jose Norman

• Adlia Alfiriani

• Henry Praherdhiono

• Guanqun Zhao

• Chei Sian Lee
• Alaa Jaber Zeyab

• Amani M Albaqshi
• Fatima A Alabdullaziz

• Internet High Educ

• Judith I. Dziuban
• Charles D. Dziuban
• Recruit researchers
• Join for free
• Login Email Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google Welcome back! Please log in. Email · Hint Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google No account? Sign up

• SUGGESTED TOPICS
• The Magazine
• Newsletters
• Managing Yourself
• Managing Teams
• Work-life Balance
• The Big Idea
• Data & Visuals
• Reading Lists
• Case Selections
• HBR Learning
• Topic Feeds
• Account Settings
• Email Preferences

The Case for College in the Era of Online Learning

• Robert Walker

In-person education provides valuable experiences, opportunities, and skills that can’t be replicated online.

Does pursuing a college education still make sense in the age of online learning and AI — when we have access to information for free via the internet? The problem with this question is that it frames college as an information gatekeeper, misunderstanding much of its value. For many, higher education institutions offer more than that: a transformative journey where students can network and develop transferable soft skills that require teamwork and repetition. You can maximize your college experience by leaning into these opportunities while simultaneously staying up to date with the latest technological trends. By being agile, networking vigorously, cultivating problem-solving skills, and seeking learning opportunities in the real world while in school, you can prepare yourself for a successful career.

Today, we have access to more information than ever before. YouTube and TikTok can provide us with in-depth learning opportunities for free — from professional development tips to AI tutorials . In more recent years, large language models like ChatGPT and Gemini have shown they can answer almost any question that comes to mind with an increasing level of accuracy .

• RW Robert Walker is the director of high school admissions at University of Advancing Technology. Walker has over 12 years of in-depth experience in recruitment and technology, has a genuine passion helping others achieve their educations dream,s and holds advanced degrees in technology leadership and cyber security.  

Partner Center

Cookie Acknowledgement

This website uses cookies to collect information to improve your browsing experience. Please review our Privacy Statement for more information.

• Administration
• Toggle Search
• Find People

Show simple item record

PianoMentor: A Technological Framework to Teach and Practice Piano/Keyboard Online via Machine Learning with an Embedded Practice Lesson Generator

Files in this item

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Elsevier - PMC COVID-19 Collection

A systematic review of research on online teaching and learning from 2009 to 2018

Associated data.

Systematic reviews were conducted in the nineties and early 2000's on online learning research. However, there is no review examining the broader aspect of research themes in online learning in the last decade. This systematic review addresses this gap by examining 619 research articles on online learning published in twelve journals in the last decade. These studies were examined for publication trends and patterns, research themes, research methods, and research settings and compared with the research themes from the previous decades. While there has been a slight decrease in the number of studies on online learning in 2015 and 2016, it has then continued to increase in 2017 and 2018. The majority of the studies were quantitative in nature and were examined in higher education. Online learning research was categorized into twelve themes and a framework across learner, course and instructor, and organizational levels was developed. Online learner characteristics and online engagement were examined in a high number of studies and were consistent with three of the prior systematic reviews. However, there is still a need for more research on organization level topics such as leadership, policy, and management and access, culture, equity, inclusion, and ethics and also on online instructor characteristics.

• • Twelve online learning research themes were identified in 2009–2018.
• • A framework with learner, course and instructor, and organizational levels was used.
• • Online learner characteristics and engagement were the mostly examined themes.
• • The majority of the studies used quantitative research methods and in higher education.
• • There is a need for more research on organization level topics.

1. Introduction

Online learning has been on the increase in the last two decades. In the United States, though higher education enrollment has declined, online learning enrollment in public institutions has continued to increase ( Allen & Seaman, 2017 ), and so has the research on online learning. There have been review studies conducted on specific areas on online learning such as innovations in online learning strategies ( Davis et al., 2018 ), empirical MOOC literature ( Liyanagunawardena et al., 2013 ; Veletsianos & Shepherdson, 2016 ; Zhu et al., 2018 ), quality in online education ( Esfijani, 2018 ), accessibility in online higher education ( Lee, 2017 ), synchronous online learning ( Martin et al., 2017 ), K-12 preparation for online teaching ( Moore-Adams et al., 2016 ), polychronicity in online learning ( Capdeferro et al., 2014 ), meaningful learning research in elearning and online learning environments ( Tsai, Shen, & Chiang, 2013 ), problem-based learning in elearning and online learning environments ( Tsai & Chiang, 2013 ), asynchronous online discussions ( Thomas, 2013 ), self-regulated learning in online learning environments ( Tsai, Shen, & Fan, 2013 ), game-based learning in online learning environments ( Tsai & Fan, 2013 ), and online course dropout ( Lee & Choi, 2011 ). While there have been review studies conducted on specific online learning topics, very few studies have been conducted on the broader aspect of online learning examining research themes.

2. Systematic Reviews of Distance Education and Online Learning Research

Distance education has evolved from offline to online settings with the access to internet and COVID-19 has made online learning the common delivery method across the world. Tallent-Runnels et al. (2006) reviewed research late 1990's to early 2000's, Berge and Mrozowski (2001) reviewed research 1990 to 1999, and Zawacki-Richter et al. (2009) reviewed research in 2000–2008 on distance education and online learning. Table 1 shows the research themes from previous systematic reviews on online learning research. There are some themes that re-occur in the various reviews, and there are also new themes that emerge. Though there have been reviews conducted in the nineties and early 2000's, there is no review examining the broader aspect of research themes in online learning in the last decade. Hence, the need for this systematic review which informs the research themes in online learning from 2009 to 2018. In the following sections, we review these systematic review studies in detail.

Comparison of online learning research themes from previous studies.

2.1. Distance education research themes, 1990 to 1999 ( Berge & Mrozowski, 2001 )

Berge and Mrozowski (2001) reviewed 890 research articles and dissertation abstracts on distance education from 1990 to 1999. The four distance education journals chosen by the authors to represent distance education included, American Journal of Distance Education, Distance Education, Open Learning, and the Journal of Distance Education. This review overlapped in the dates of the Tallent-Runnels et al. (2006) study. Berge and Mrozowski (2001) categorized the articles according to Sherry's (1996) ten themes of research issues in distance education: redefining roles of instructor and students, technologies used, issues of design, strategies to stimulate learning, learner characteristics and support, issues related to operating and policies and administration, access and equity, and costs and benefits.

In the Berge and Mrozowski (2001) study, more than 100 studies focused on each of the three themes: (1) design issues, (2) learner characteristics, and (3) strategies to increase interactivity and active learning. By design issues, the authors focused on instructional systems design and focused on topics such as content requirement, technical constraints, interactivity, and feedback. The next theme, strategies to increase interactivity and active learning, were closely related to design issues and focused on students’ modes of learning. Learner characteristics focused on accommodating various learning styles through customized instructional theory. Less than 50 studies focused on the three least examined themes: (1) cost-benefit tradeoffs, (2) equity and accessibility, and (3) learner support. Cost-benefit trade-offs focused on the implementation costs of distance education based on school characteristics. Equity and accessibility focused on the equity of access to distance education systems. Learner support included topics such as teacher to teacher support as well as teacher to student support.

2.2. Online learning research themes, 1993 to 2004 ( Tallent-Runnels et al., 2006 )

Tallent-Runnels et al. (2006) reviewed research on online instruction from 1993 to 2004. They reviewed 76 articles focused on online learning by searching five databases, ERIC, PsycINFO, ContentFirst, Education Abstracts, and WilsonSelect. Tallent-Runnels et al. (2006) categorized research into four themes, (1) course environment, (2) learners' outcomes, (3) learners’ characteristics, and (4) institutional and administrative factors. The first theme that the authors describe as course environment ( n  = 41, 53.9%) is an overarching theme that includes classroom culture, structural assistance, success factors, online interaction, and evaluation.

Tallent-Runnels et al. (2006) for their second theme found that studies focused on questions involving the process of teaching and learning and methods to explore cognitive and affective learner outcomes ( n  = 29, 38.2%). The authors stated that they found the research designs flawed and lacked rigor. However, the literature comparing traditional and online classrooms found both delivery systems to be adequate. Another research theme focused on learners’ characteristics ( n  = 12, 15.8%) and the synergy of learners, design of the online course, and system of delivery. Research findings revealed that online learners were mainly non-traditional, Caucasian, had different learning styles, and were highly motivated to learn. The final theme that they reported was institutional and administrative factors (n  = 13, 17.1%) on online learning. Their findings revealed that there was a lack of scholarly research in this area and most institutions did not have formal policies in place for course development as well as faculty and student support in training and evaluation. Their research confirmed that when universities offered online courses, it improved student enrollment numbers.

2.3. Distance education research themes 2000 to 2008 ( Zawacki-Richter et al., 2009 )

Zawacki-Richter et al. (2009) reviewed 695 articles on distance education from 2000 to 2008 using the Delphi method for consensus in identifying areas and classified the literature from five prominent journals. The five journals selected due to their wide scope in research in distance education included Open Learning, Distance Education, American Journal of Distance Education, the Journal of Distance Education, and the International Review of Research in Open and Distributed Learning. The reviewers examined the main focus of research and identified gaps in distance education research in this review.

Zawacki-Richter et al. (2009) classified the studies into macro, meso and micro levels focusing on 15 areas of research. The five areas of the macro-level addressed: (1) access, equity and ethics to deliver distance education for developing nations and the role of various technologies to narrow the digital divide, (2) teaching and learning drivers, markets, and professional development in the global context, (3) distance delivery systems and institutional partnerships and programs and impact of hybrid modes of delivery, (4) theoretical frameworks and models for instruction, knowledge building, and learner interactions in distance education practice, and (5) the types of preferred research methodologies. The meso-level focused on seven areas that involve: (1) management and organization for sustaining distance education programs, (2) examining financial aspects of developing and implementing online programs, (3) the challenges and benefits of new technologies for teaching and learning, (4) incentives to innovate, (5) professional development and support for faculty, (6) learner support services, and (7) issues involving quality standards and the impact on student enrollment and retention. The micro-level focused on three areas: (1) instructional design and pedagogical approaches, (2) culturally appropriate materials, interaction, communication, and collaboration among a community of learners, and (3) focus on characteristics of adult learners, socio-economic backgrounds, learning preferences, and dispositions.

The top three research themes in this review by Zawacki-Richter et al. (2009) were interaction and communities of learning ( n  = 122, 17.6%), instructional design ( n  = 121, 17.4%) and learner characteristics ( n  = 113, 16.3%). The lowest number of studies (less than 3%) were found in studies examining the following research themes, management and organization ( n  = 18), research methods in DE and knowledge transfer ( n  = 13), globalization of education and cross-cultural aspects ( n  = 13), innovation and change ( n  = 13), and costs and benefits ( n  = 12).

2.4. Online learning research themes

These three systematic reviews provide a broad understanding of distance education and online learning research themes from 1990 to 2008. However, there is an increase in the number of research studies on online learning in this decade and there is a need to identify recent research themes examined. Based on the previous systematic reviews ( Berge & Mrozowski, 2001 ; Hung, 2012 ; Tallent-Runnels et al., 2006 ; Zawacki-Richter et al., 2009 ), online learning research in this study is grouped into twelve different research themes which include Learner characteristics, Instructor characteristics, Course or program design and development, Course Facilitation, Engagement, Course Assessment, Course Technologies, Access, Culture, Equity, Inclusion, and Ethics, Leadership, Policy and Management, Instructor and Learner Support, and Learner Outcomes. Table 2 below describes each of the research themes and using these themes, a framework is derived in Fig. 1 .

Research themes in online learning.

Online learning research themes framework.

The collection of research themes is presented as a framework in Fig. 1 . The themes are organized by domain or level to underscore the nested relationship that exists. As evidenced by the assortment of themes, research can focus on any domain of delivery or associated context. The “Learner” domain captures characteristics and outcomes related to learners and their interaction within the courses. The “Course and Instructor” domain captures elements about the broader design of the course and facilitation by the instructor, and the “Organizational” domain acknowledges the contextual influences on the course. It is important to note as well that due to the nesting, research themes can cross domains. For example, the broader cultural context may be studied as it pertains to course design and development, and institutional support can include both learner support and instructor support. Likewise, engagement research can involve instructors as well as learners.

In this introduction section, we have reviewed three systematic reviews on online learning research ( Berge & Mrozowski, 2001 ; Tallent-Runnels et al., 2006 ; Zawacki-Richter et al., 2009 ). Based on these reviews and other research, we have derived twelve themes to develop an online learning research framework which is nested in three levels: learner, course and instructor, and organization.

2.5. Purpose of this research

In two out of the three previous reviews, design, learner characteristics and interaction were examined in the highest number of studies. On the other hand, cost-benefit tradeoffs, equity and accessibility, institutional and administrative factors, and globalization and cross-cultural aspects were examined in the least number of studies. One explanation for this may be that it is a function of nesting, noting that studies falling in the Organizational and Course levels may encompass several courses or many more participants within courses. However, while some research themes re-occur, there are also variations in some themes across time, suggesting the importance of research themes rise and fall over time. Thus, a critical examination of the trends in themes is helpful for understanding where research is needed most. Also, since there is no recent study examining online learning research themes in the last decade, this study strives to address that gap by focusing on recent research themes found in the literature, and also reviewing research methods and settings. Notably, one goal is to also compare findings from this decade to the previous review studies. Overall, the purpose of this study is to examine publication trends in online learning research taking place during the last ten years and compare it with the previous themes identified in other review studies. Due to the continued growth of online learning research into new contexts and among new researchers, we also examine the research methods and settings found in the studies of this review.

The following research questions are addressed in this study.

• 1. What percentage of the population of articles published in the journals reviewed from 2009 to 2018 were related to online learning and empirical?
• 2. What is the frequency of online learning research themes in the empirical online learning articles of journals reviewed from 2009 to 2018?
• 3. What is the frequency of research methods and settings that researchers employed in the empirical online learning articles of the journals reviewed from 2009 to 2018?

This five-step systematic review process described in the U.S. Department of Education, Institute of Education Sciences, What Works Clearinghouse Procedures and Standards Handbook, Version 4.0 ( 2017 ) was used in this systematic review: (a) developing the review protocol, (b) identifying relevant literature, (c) screening studies, (d) reviewing articles, and (e) reporting findings.

3.1. Data sources and search strategies

The Education Research Complete database was searched using the keywords below for published articles between the years 2009 and 2018 using both the Title and Keyword function for the following search terms.

“online learning" OR "online teaching" OR "online program" OR "online course" OR “online education”

3.2. Inclusion/exclusion criteria

The initial search of online learning research among journals in the database resulted in more than 3000 possible articles. Therefore, we limited our search to select journals that focus on publishing peer-reviewed online learning and educational research. Our aim was to capture the journals that published the most articles in online learning. However, we also wanted to incorporate the concept of rigor, so we used expert perception to identify 12 peer-reviewed journals that publish high-quality online learning research. Dissertations and conference proceedings were excluded. To be included in this systematic review, each study had to meet the screening criteria as described in Table 3 . A research study was excluded if it did not meet all of the criteria to be included.

Inclusion/Exclusion criteria.

3.3. Process flow selection of articles

Fig. 2 shows the process flow involved in the selection of articles. The search in the database Education Research Complete yielded an initial sample of 3332 articles. Targeting the 12 journals removed 2579 articles. After reviewing the abstracts, we removed 134 articles based on the inclusion/exclusion criteria. The final sample, consisting of 619 articles, was entered into the computer software MAXQDA ( VERBI Software, 2019 ) for coding.

Flowchart of online learning research selection.

3.4. Developing review protocol

A review protocol was designed as a codebook in MAXQDA ( VERBI Software, 2019 ) by the three researchers. The codebook was developed based on findings from the previous review studies and from the initial screening of the articles in this review. The codebook included 12 research themes listed earlier in Table 2 (Learner characteristics, Instructor characteristics, Course or program design and development, Course Facilitation, Engagement, Course Assessment, Course Technologies, Access, Culture, Equity, Inclusion, and Ethics, Leadership, Policy and Management, Instructor and Learner Support, and Learner Outcomes), four research settings (higher education, continuing education, K-12, corporate/military), and three research designs (quantitative, qualitative and mixed methods). Fig. 3 below is a screenshot of MAXQDA used for the coding process.

Codebook from MAXQDA.

3.5. Data coding

Research articles were coded by two researchers in MAXQDA. Two researchers independently coded 10% of the articles and then discussed and updated the coding framework. The second author who was a doctoral student coded the remaining studies. The researchers met bi-weekly to address coding questions that emerged. After the first phase of coding, we found that more than 100 studies fell into each of the categories of Learner Characteristics or Engagement, so we decided to pursue a second phase of coding and reexamine the two themes. Learner Characteristics were classified into the subthemes of Academic, Affective, Motivational, Self-regulation, Cognitive, and Demographic Characteristics. Engagement was classified into the subthemes of Collaborating, Communication, Community, Involvement, Interaction, Participation, and Presence.

3.6. Data analysis

Frequency tables were generated for each of the variables so that outliers could be examined and narrative data could be collapsed into categories. Once cleaned and collapsed into a reasonable number of categories, descriptive statistics were used to describe each of the coded elements. We first present the frequencies of publications related to online learning in the 12 journals. The total number of articles for each journal (collectively, the population) was hand-counted from journal websites, excluding editorials and book reviews. The publication trend of online learning research was also depicted from 2009 to 2018. Then, the descriptive information of the 12 themes, including the subthemes of Learner Characteristics and Engagement were provided. Finally, research themes by research settings and methodology were elaborated.

4.1. Publication trends on online learning

Publication patterns of the 619 articles reviewed from the 12 journals are presented in Table 4 . International Review of Research in Open and Distributed Learning had the highest number of publications in this review. Overall, about 8% of the articles appearing in these twelve journals consisted of online learning publications; however, several journals had concentrations of online learning articles totaling more than 20%.

Empirical online learning research articles by journal, 2009–2018.

Note . Journal's Total Article count excludes reviews and editorials.

The publication trend of online learning research is depicted in Fig. 4 . When disaggregated by year, the total frequency of publications shows an increasing trend. Online learning articles increased throughout the decade and hit a relative maximum in 2014. The greatest number of online learning articles ( n  = 86) occurred most recently, in 2018.

Online learning publication trends by year.

4.2. Online learning research themes that appeared in the selected articles

The publications were categorized into the twelve research themes identified in Fig. 1 . The frequency counts and percentages of the research themes are provided in Table 5 below. A majority of the research is categorized into the Learner domain. The fewest number of articles appears in the Organization domain.

Research themes in the online learning publications from 2009 to 2018.

The specific themes of Engagement ( n  = 179, 28.92%) and Learner Characteristics ( n  = 134, 21.65%) were most often examined in publications. These two themes were further coded to identify sub-themes, which are described in the next two sections. Publications focusing on Instructor Characteristics ( n  = 21, 3.39%) were least common in the dataset.

4.2.1. Research on engagement

The largest number of studies was on engagement in online learning, which in the online learning literature is referred to and examined through different terms. Hence, we explore this category in more detail. In this review, we categorized the articles into seven different sub-themes as examined through different lenses including presence, interaction, community, participation, collaboration, involvement, and communication. We use the term “involvement” as one of the terms since researchers sometimes broadly used the term engagement to describe their work without further description. Table 6 below provides the description, frequency, and percentages of the various studies related to engagement.

Research sub-themes on engagement.

In the sections below, we provide several examples of the different engagement sub-themes that were studied within the larger engagement theme.

Presence. This sub-theme was the most researched in engagement. With the development of the community of inquiry framework most of the studies in this subtheme examined social presence ( Akcaoglu & Lee, 2016 ; Phirangee & Malec, 2017 ; Wei et al., 2012 ), teaching presence ( Orcutt & Dringus, 2017 ; Preisman, 2014 ; Wisneski et al., 2015 ) and cognitive presence ( Archibald, 2010 ; Olesova et al., 2016 ).

Interaction . This was the second most studied theme under engagement. Researchers examined increasing interpersonal interactions ( Cung et al., 2018 ), learner-learner interactions ( Phirangee, 2016 ; Shackelford & Maxwell, 2012 ; Tawfik et al., 2018 ), peer-peer interaction ( Comer et al., 2014 ), learner-instructor interaction ( Kuo et al., 2014 ), learner-content interaction ( Zimmerman, 2012 ), interaction through peer mentoring ( Ruane & Koku, 2014 ), interaction and community building ( Thormann & Fidalgo, 2014 ), and interaction in discussions ( Ruane & Lee, 2016 ; Tibi, 2018 ).

Community. Researchers examined building community in online courses ( Berry, 2017 ), supporting a sense of community ( Jiang, 2017 ), building an online learning community of practice ( Cho, 2016 ), building an academic community ( Glazer & Wanstreet, 2011 ; Nye, 2015 ; Overbaugh & Nickel, 2011 ), and examining connectedness and rapport in an online community ( Bolliger & Inan, 2012 ; Murphy & Rodríguez-Manzanares, 2012 ; Slagter van Tryon & Bishop, 2012 ).

Participation. Researchers examined engagement through participation in a number of studies. Some of the topics include, participation patterns in online discussion ( Marbouti & Wise, 2016 ; Wise et al., 2012 ), participation in MOOCs ( Ahn et al., 2013 ; Saadatmand & Kumpulainen, 2014 ), features that influence students’ online participation ( Rye & Støkken, 2012 ) and active participation.

Collaboration. Researchers examined engagement through collaborative learning. Specific studies focused on cross-cultural collaboration ( Kumi-Yeboah, 2018 ; Yang et al., 2014 ), how virtual teams collaborate ( Verstegen et al., 2018 ), types of collaboration teams ( Wicks et al., 2015 ), tools for collaboration ( Boling et al., 2014 ), and support for collaboration ( Kopp et al., 2012 ).

Involvement. Researchers examined engaging learners through involvement in various learning activities ( Cundell & Sheepy, 2018 ), student engagement through various measures ( Dixson, 2015 ), how instructors included engagement to involve students in learning ( O'Shea et al., 2015 ), different strategies to engage the learner ( Amador & Mederer, 2013 ), and designed emotionally engaging online environments ( Koseoglu & Doering, 2011 ).

Communication. Researchers examined communication in online learning in studies using social network analysis ( Ergün & Usluel, 2016 ), using informal communication tools such as Facebook for class discussion ( Kent, 2013 ), and using various modes of communication ( Cunningham et al., 2010 ; Rowe, 2016 ). Studies have also focused on both asynchronous and synchronous aspects of communication ( Swaggerty & Broemmel, 2017 ; Yamagata-Lynch, 2014 ).

4.2.2. Research on learner characteristics

The second largest theme was learner characteristics. In this review, we explore this further to identify several aspects of learner characteristics. In this review, we categorized the learner characteristics into self-regulation characteristics, motivational characteristics, academic characteristics, affective characteristics, cognitive characteristics, and demographic characteristics. Table 7 provides the number of studies and percentages examining the various learner characteristics.

Research sub-themes on learner characteristics.

Online learning has elements that are different from the traditional face-to-face classroom and so the characteristics of the online learners are also different. Yukselturk and Top (2013) categorized online learner profile into ten aspects: gender, age, work status, self-efficacy, online readiness, self-regulation, participation in discussion list, participation in chat sessions, satisfaction, and achievement. Their categorization shows that there are differences in online learner characteristics in these aspects when compared to learners in other settings. Some of the other aspects such as participation and achievement as discussed by Yukselturk and Top (2013) are discussed in different research themes in this study. The sections below provide examples of the learner characteristics sub-themes that were studied.

Self-regulation. Several researchers have examined self-regulation in online learning. They found that successful online learners are academically motivated ( Artino & Stephens, 2009 ), have academic self-efficacy ( Cho & Shen, 2013 ), have grit and intention to succeed ( Wang & Baker, 2018 ), have time management and elaboration strategies ( Broadbent, 2017 ), set goals and revisit course content ( Kizilcec et al., 2017 ), and persist ( Glazer & Murphy, 2015 ). Researchers found a positive relationship between learner's self-regulation and interaction ( Delen et al., 2014 ) and self-regulation and communication and collaboration ( Barnard et al., 2009 ).

Motivation. Researchers focused on motivation of online learners including different motivation levels of online learners ( Li & Tsai, 2017 ), what motivated online learners ( Chaiprasurt & Esichaikul, 2013 ), differences in motivation of online learners ( Hartnett et al., 2011 ), and motivation when compared to face to face learners ( Paechter & Maier, 2010 ). Harnett et al. (2011) found that online learner motivation was complex, multifaceted, and sensitive to situational conditions.

Academic. Several researchers have focused on academic aspects for online learner characteristics. Readiness for online learning has been examined as an academic factor by several researchers ( Buzdar et al., 2016 ; Dray et al., 2011 ; Wladis & Samuels, 2016 ; Yu, 2018 ) specifically focusing on creating and validating measures to examine online learner readiness including examining students emotional intelligence as a measure of student readiness for online learning. Researchers have also examined other academic factors such as academic standing ( Bradford & Wyatt, 2010 ), course level factors ( Wladis et al., 2014 ) and academic skills in online courses ( Shea & Bidjerano, 2014 ).

Affective. Anderson and Bourke (2013) describe affective characteristics through which learners express feelings or emotions. Several research studies focused on the affective characteristics of online learners. Learner satisfaction for online learning has been examined by several researchers ( Cole et al., 2014 ; Dziuban et al., 2015 ; Kuo et al., 2013 ; Lee, 2014a ) along with examining student emotions towards online assessment ( Kim et al., 2014 ).

Cognitive. Researchers have also examined cognitive aspects of learner characteristics including meta-cognitive skills, cognitive variables, higher-order thinking, cognitive density, and critical thinking ( Chen & Wu, 2012 ; Lee, 2014b ). Lee (2014b) examined the relationship between cognitive presence density and higher-order thinking skills. Chen and Wu (2012) examined the relationship between cognitive and motivational variables in an online system for secondary physical education.

Demographic. Researchers have examined various demographic factors in online learning. Several researchers have examined gender differences in online learning ( Bayeck et al., 2018 ; Lowes et al., 2016 ; Yukselturk & Bulut, 2009 ), ethnicity, age ( Ke & Kwak, 2013 ), and minority status ( Yeboah & Smith, 2016 ) of online learners.

4.2.3. Less frequently studied research themes

While engagement and learner characteristics were studied the most, other themes were less often studied in the literature and are presented here, according to size, with general descriptions of the types of research examined for each.

Evaluation and Quality Assurance. There were 38 studies (6.14%) published in the theme of evaluation and quality assurance. Some of the studies in this theme focused on course quality standards, using quality matters to evaluate quality, using the CIPP model for evaluation, online learning system evaluation, and course and program evaluations.

Course Technologies. There were 35 studies (5.65%) published in the course technologies theme. Some of the studies examined specific technologies such as Edmodo, YouTube, Web 2.0 tools, wikis, Twitter, WebCT, Screencasts, and Web conferencing systems in the online learning context.

Course Facilitation. There were 34 studies (5.49%) published in the course facilitation theme. Some of the studies in this theme examined facilitation strategies and methods, experiences of online facilitators, and online teaching methods.

Institutional Support. There were 33 studies (5.33%) published in the institutional support theme which included support for both the instructor and learner. Some of the studies on instructor support focused on training new online instructors, mentoring programs for faculty, professional development resources for faculty, online adjunct faculty training, and institutional support for online instructors. Studies on learner support focused on learning resources for online students, cognitive and social support for online learners, and help systems for online learner support.

Learner Outcome. There were 32 studies (5.17%) published in the learner outcome theme. Some of the studies that were examined in this theme focused on online learner enrollment, completion, learner dropout, retention, and learner success.

Course Assessment. There were 30 studies (4.85%) published in the course assessment theme. Some of the studies in the course assessment theme examined online exams, peer assessment and peer feedback, proctoring in online exams, and alternative assessments such as eportfolio.

Access, Culture, Equity, Inclusion, and Ethics. There were 29 studies (4.68%) published in the access, culture, equity, inclusion, and ethics theme. Some of the studies in this theme examined online learning across cultures, multi-cultural effectiveness, multi-access, and cultural diversity in online learning.

Leadership, Policy, and Management. There were 27 studies (4.36%) published in the leadership, policy, and management theme. Some of the studies on leadership, policy, and management focused on online learning leaders, stakeholders, strategies for online learning leadership, resource requirements, university policies for online course policies, governance, course ownership, and faculty incentives for online teaching.

Course Design and Development. There were 27 studies (4.36%) published in the course design and development theme. Some of the studies examined in this theme focused on design elements, design issues, design process, design competencies, design considerations, and instructional design in online courses.

Instructor Characteristics. There were 21 studies (3.39%) published in the instructor characteristics theme. Some of the studies in this theme were on motivation and experiences of online instructors, ability to perform online teaching duties, roles of online instructors, and adjunct versus full-time online instructors.

4.3. Research settings and methodology used in the studies

The research methods used in the studies were classified into quantitative, qualitative, and mixed methods ( Harwell, 2012 , pp. 147–163). The research setting was categorized into higher education, continuing education, K-12, and corporate/military. As shown in Table A in the appendix, the vast majority of the publications used higher education as the research setting ( n  = 509, 67.6%). Table B in the appendix shows that approximately half of the studies adopted the quantitative method ( n  = 324, 43.03%), followed by the qualitative method ( n  = 200, 26.56%). Mixed methods account for the smallest portion ( n  = 95, 12.62%).

Table A shows that the patterns of the four research settings were approximately consistent across the 12 themes except for the theme of Leaner Outcome and Institutional Support. Continuing education had a higher relative frequency in Learner Outcome (0.28) and K-12 had a higher relative frequency in Institutional Support (0.33) compared to the frequencies they had in the total themes (0.09 and 0.08 respectively). Table B in the appendix shows that the distribution of the three methods were not consistent across the 12 themes. While quantitative studies and qualitative studies were roughly evenly distributed in Engagement, they had a large discrepancy in Learner Characteristics. There were 100 quantitative studies; however, only 18 qualitative studies published in the theme of Learner Characteristics.

In summary, around 8% of the articles published in the 12 journals focus on online learning. Online learning publications showed a tendency of increase on the whole in the past decade, albeit fluctuated, with the greatest number occurring in 2018. Among the 12 research themes related to online learning, the themes of Engagement and Learner Characteristics were studied the most and the theme of Instructor Characteristics was studied the least. Most studies were conducted in the higher education setting and approximately half of the studies used the quantitative method. Looking at the 12 themes by setting and method, we found that the patterns of the themes by setting or by method were not consistent across the 12 themes.

The quality of our findings was ensured by scientific and thorough searches and coding consistency. The selection of the 12 journals provides evidence of the representativeness and quality of primary studies. In the coding process, any difficulties and questions were resolved by consultations with the research team at bi-weekly meetings, which ensures the intra-rater and interrater reliability of coding. All these approaches guarantee the transparency and replicability of the process and the quality of our results.

5. Discussion

This review enabled us to identify the online learning research themes examined from 2009 to 2018. In the section below, we review the most studied research themes, engagement and learner characteristics along with implications, limitations, and directions for future research.

5.1. Most studied research themes

Three out of the four systematic reviews informing the design of the present study found that online learner characteristics and online engagement were examined in a high number of studies. In this review, about half of the studies reviewed (50.57%) focused on online learner characteristics or online engagement. This shows the continued importance of these two themes. In the Tallent-Runnels et al.’s (2006) study, the learner characteristics theme was identified as least studied for which they state that researchers are beginning to investigate learner characteristics in the early days of online learning.

One of the differences found in this review is that course design and development was examined in the least number of studies in this review compared to two prior systematic reviews ( Berge & Mrozowski, 2001 ; Zawacki-Richter et al., 2009 ). Zawacki-Richter et al. did not use a keyword search but reviewed all the articles in five different distance education journals. Berge and Mrozowski (2001) included a research theme called design issues to include all aspects of instructional systems design in distance education journals. In our study, in addition to course design and development, we also had focused themes on learner outcomes, course facilitation, course assessment and course evaluation. These are all instructional design focused topics and since we had multiple themes focusing on instructional design topics, the course design and development category might have resulted in fewer studies. There is still a need for more studies to focus on online course design and development.

5.2. Least frequently studied research themes

Three out of the four systematic reviews discussed in the opening of this study found management and organization factors to be least studied. In this review, Leadership, Policy, and Management was studied among 4.36% of the studies and Access, Culture, Equity, Inclusion, and Ethics was studied among 4.68% of the studies in the organizational level. The theme on Equity and accessibility was also found to be the least studied theme in the Berge and Mrozowski (2001) study. In addition, instructor characteristics was the least examined research theme among the twelve themes studied in this review. Only 3.39% of the studies were on instructor characteristics. While there were some studies examining instructor motivation and experiences, instructor ability to teach online, online instructor roles, and adjunct versus full-time online instructors, there is still a need to examine topics focused on instructors and online teaching. This theme was not included in the prior reviews as the focus was more on the learner and the course but not on the instructor. While it is helpful to see research evolving on instructor focused topics, there is still a need for more research on the online instructor.

5.3. Comparing research themes from current study to previous studies

The research themes from this review were compared with research themes from previous systematic reviews, which targeted prior decades. Table 8 shows the comparison.

Comparison of most and least studied online learning research themes from current to previous reviews.

L = Learner, C=Course O=Organization.

5.4. Need for more studies on organizational level themes of online learning

In this review there is a greater concentration of studies focused on Learner domain topics, and reduced attention to broader more encompassing research themes that fall into the Course and Organization domains. There is a need for organizational level topics such as Access, Culture, Equity, Inclusion and Ethics, and Leadership, Policy and Management to be researched on within the context of online learning. Examination of access, culture, equity, inclusion and ethics is very important to support diverse online learners, particularly with the rapid expansion of online learning across all educational levels. This was also least studied based on Berge and Mrozowski (2001) systematic review.

The topics on leadership, policy and management were least studied both in this review and also in the Tallent-Runnels et al. (2006) and Zawacki-Richter et al. (2009) study. Tallent-Runnels categorized institutional and administrative aspects into institutional policies, institutional support, and enrollment effects. While we included support as a separate category, in this study leadership, policy and management were combined. There is still a need for research on leadership of those who manage online learning, policies for online education, and managing online programs. In the Zawacki-Richter et al. (2009) study, only a few studies examined management and organization focused topics. They also found management and organization to be strongly correlated with costs and benefits. In our study, costs and benefits were collectively included as an aspect of management and organization and not as a theme by itself. These studies will provide research-based evidence for online education administrators.

6. Limitations

As with any systematic review, there are limitations to the scope of the review. The search is limited to twelve journals in the field that typically include research on online learning. These manuscripts were identified by searching the Education Research Complete database which focuses on education students, professionals, and policymakers. Other discipline-specific journals as well as dissertations and proceedings were not included due to the volume of articles. Also, the search was performed using five search terms “online learning" OR "online teaching" OR "online program" OR "online course" OR “online education” in title and keyword. If authors did not include these terms, their respective work may have been excluded from this review even if it focused on online learning. While these terms are commonly used in North America, it may not be commonly used in other parts of the world. Additional studies may exist outside this scope.

The search strategy also affected how we presented results and introduced limitations regarding generalization. We identified that only 8% of the articles published in these journals were related to online learning; however, given the use of search terms to identify articles within select journals it was not feasible to identify the total number of research-based articles in the population. Furthermore, our review focused on the topics and general methods of research and did not systematically consider the quality of the published research. Lastly, some journals may have preferences for publishing studies on a particular topic or that use a particular method (e.g., quantitative methods), which introduces possible selection and publication biases which may skew the interpretation of results due to over/under representation. Future studies are recommended to include more journals to minimize the selection bias and obtain a more representative sample.

Certain limitations can be attributed to the coding process. Overall, the coding process for this review worked well for most articles, as each tended to have an individual or dominant focus as described in the abstracts, though several did mention other categories which likely were simultaneously considered to a lesser degree. However, in some cases, a dominant theme was not as apparent and an effort to create mutually exclusive groups for clearer interpretation the coders were occasionally forced to choose between two categories. To facilitate this coding, the full-texts were used to identify a study focus through a consensus seeking discussion among all authors. Likewise, some studies focused on topics that we have associated with a particular domain, but the design of the study may have promoted an aggregated examination or integrated factors from multiple domains (e.g., engagement). Due to our reliance on author descriptions, the impact of construct validity is likely a concern that requires additional exploration. Our final grouping of codes may not have aligned with the original author's description in the abstract. Additionally, coding of broader constructs which disproportionately occur in the Learner domain, such as learner outcomes, learner characteristics, and engagement, likely introduced bias towards these codes when considering studies that involved multiple domains. Additional refinement to explore the intersection of domains within studies is needed.

7. Implications and future research

One of the strengths of this review is the research categories we have identified. We hope these categories will support future researchers and identify areas and levels of need for future research. Overall, there is some agreement on research themes on online learning research among previous reviews and this one, at the same time there are some contradicting findings. We hope the most-researched themes and least-researched themes provide authors a direction on the importance of research and areas of need to focus on.

The leading themes found in this review is online engagement research. However, presentation of this research was inconsistent, and often lacked specificity. This is not unique to online environments, but the nuances of defining engagement in an online environment are unique and therefore need further investigation and clarification. This review points to seven distinct classifications of online engagement. Further research on engagement should indicate which type of engagement is sought. This level of specificity is necessary to establish instruments for measuring engagement and ultimately testing frameworks for classifying engagement and promoting it in online environments. Also, it might be of importance to examine the relationship between these seven sub-themes of engagement.

Additionally, this review highlights growing attention to learner characteristics, which constitutes a shift in focus away from instructional characteristics and course design. Although this is consistent with the focus on engagement, the role of the instructor, and course design with respect to these outcomes remains important. Results of the learner characteristics and engagement research paired with course design will have important ramifications for the use of teaching and learning professionals who support instruction. Additionally, the review also points to a concentration of research in the area of higher education. With an immediate and growing emphasis on online learning in K-12 and corporate settings, there is a critical need for further investigation in these settings.

Lastly, because the present review did not focus on the overall effect of interventions, opportunities exist for dedicated meta-analyses. Particular attention to research on engagement and learner characteristics as well as how these vary by study design and outcomes would be logical additions to the research literature.

8. Conclusion

This systematic review builds upon three previous reviews which tackled the topic of online learning between 1990 and 2010 by extending the timeframe to consider the most recent set of published research. Covering the most recent decade, our review of 619 articles from 12 leading online learning journal points to a more concentrated focus on the learner domain including engagement and learner characteristics, with more limited attention to topics pertaining to the classroom or organizational level. The review highlights an opportunity for the field to clarify terminology concerning online learning research, particularly in the areas of learner outcomes where there is a tendency to classify research more generally (e.g., engagement). Using this sample of published literature, we provide a possible taxonomy for categorizing this research using subcategories. The field could benefit from a broader conversation about how these categories can shape a comprehensive framework for online learning research. Such efforts will enable the field to effectively prioritize research aims over time and synthesize effects.

Credit author statement

Florence Martin: Conceptualization; Writing - original draft, Writing - review & editing Preparation, Supervision, Project administration. Ting Sun: Methodology, Formal analysis, Writing - original draft, Writing - review & editing. Carl Westine: Methodology, Formal analysis, Writing - original draft, Writing - review & editing, Supervision

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

1 Includes articles that are cited in this manuscript and also included in the systematic review. The entire list of 619 articles used in the systematic review can be obtained by emailing the authors.*

Appendix B Supplementary data to this article can be found online at https://doi.org/10.1016/j.compedu.2020.104009 .

Appendix A. 

Research Themes by the Settings in the Online Learning Publications

Research Themes by the Methodology in the Online Learning Publications

Appendix B. Supplementary data

The following are the Supplementary data to this article:

References 1

• Ahn J., Butler B.S., Alam A., Webster S.A. Learner participation and engagement in open online courses: Insights from the Peer 2 Peer University. MERLOT Journal of Online Learning and Teaching. 2013; 9 (2):160–171. * [ Google Scholar ]
• Akcaoglu M., Lee E. Increasing social presence in online learning through small group discussions. International Review of Research in Open and Distance Learning. 2016; 17 (3) * [ Google Scholar ]
• Allen I.E., Seaman J. Babson survey research group; 2017. Digital compass learning: Distance education enrollment Report 2017. [ Google Scholar ]
• Amador J.A., Mederer H. Migrating successful student engagement strategies online: Opportunities and challenges using jigsaw groups and problem-based learning. Journal of Online Learning and Teaching. 2013; 9 (1):89. * [ Google Scholar ]
• Anderson L.W., Bourke S.F. Routledge; 2013. Assessing affective characteristics in the schools. [ Google Scholar ]
• Archibald D. Fostering the development of cognitive presence: Initial findings using the community of inquiry survey instrument. The Internet and Higher Education. 2010; 13 (1–2):73–74. * [ Google Scholar ]
• Artino A.R., Jr., Stephens J.M. Academic motivation and self-regulation: A comparative analysis of undergraduate and graduate students learning online. The Internet and Higher Education. 2009; 12 (3–4):146–151. [ Google Scholar ]
• Barnard L., Lan W.Y., To Y.M., Paton V.O., Lai S.L. Measuring self-regulation in online and blended learning environments. Internet and Higher Education. 2009; 12 (1):1–6. * [ Google Scholar ]
• Bayeck R.Y., Hristova A., Jablokow K.W., Bonafini F. Exploring the relevance of single‐gender group formation: What we learn from a massive open online course (MOOC) British Journal of Educational Technology. 2018; 49 (1):88–100. * [ Google Scholar ]
• Berge Z., Mrozowski S. Review of research in distance education, 1990 to 1999. American Journal of Distance Education. 2001; 15 (3):5–19. doi: 10.1080/08923640109527090. [ CrossRef ] [ Google Scholar ]
• Berry S. Building community in online doctoral classrooms: Instructor practices that support community. Online Learning. 2017; 21 (2):n2. * [ Google Scholar ]
• Boling E.C., Holan E., Horbatt B., Hough M., Jean-Louis J., Khurana C., Spiezio C. Using online tools for communication and collaboration: Understanding educators' experiences in an online course. The Internet and Higher Education. 2014; 23 :48–55. * [ Google Scholar ]
• Bolliger D.U., Inan F.A. Development and validation of the online student connectedness survey (OSCS) International Review of Research in Open and Distance Learning. 2012; 13 (3):41–65. * [ Google Scholar ]
• Bradford G., Wyatt S. Online learning and student satisfaction: Academic standing, ethnicity and their influence on facilitated learning, engagement, and information fluency. The Internet and Higher Education. 2010; 13 (3):108–114. * [ Google Scholar ]
• Broadbent J. Comparing online and blended learner's self-regulated learning strategies and academic performance. The Internet and Higher Education. 2017; 33 :24–32. [ Google Scholar ]
• Buzdar M., Ali A., Tariq R. Emotional intelligence as a determinant of readiness for online learning. International Review of Research in Open and Distance Learning. 2016; 17 (1) * [ Google Scholar ]
• Capdeferro N., Romero M., Barberà E. Polychronicity: Review of the literature and a new configuration for the study of this hidden dimension of online learning. Distance Education. 2014; 35 (3):294–310. [ Google Scholar ]
• Chaiprasurt C., Esichaikul V. Enhancing motivation in online courses with mobile communication tool support: A comparative study. International Review of Research in Open and Distance Learning. 2013; 14 (3):377–401. [ Google Scholar ]
• Chen C.H., Wu I.C. The interplay between cognitive and motivational variables in a supportive online learning system for secondary physical education. Computers & Education. 2012; 58 (1):542–550. * [ Google Scholar ]
• Cho H. Under co-construction: An online community of practice for bilingual pre-service teachers. Computers & Education. 2016; 92 :76–89. * [ Google Scholar ]
• Cho M.H., Shen D. Self-regulation in online learning. Distance Education. 2013; 34 (3):290–301. [ Google Scholar ]
• Cole M.T., Shelley D.J., Swartz L.B. Online instruction, e-learning, and student satisfaction: A three-year study. International Review of Research in Open and Distance Learning. 2014; 15 (6) * [ Google Scholar ]
• Comer D.K., Clark C.R., Canelas D.A. Writing to learn and learning to write across the disciplines: Peer-to-peer writing in introductory-level MOOCs. International Review of Research in Open and Distance Learning. 2014; 15 (5):26–82. * [ Google Scholar ]
• Cundell A., Sheepy E. Student perceptions of the most effective and engaging online learning activities in a blended graduate seminar. Online Learning. 2018; 22 (3):87–102. * [ Google Scholar ]
• Cung B., Xu D., Eichhorn S. Increasing interpersonal interactions in an online course: Does increased instructor email activity and voluntary meeting time in a physical classroom facilitate student learning? Online Learning. 2018; 22 (3):193–215. [ Google Scholar ]
• Cunningham U.M., Fägersten K.B., Holmsten E. Can you hear me, Hanoi?" Compensatory mechanisms employed in synchronous net-based English language learning. International Review of Research in Open and Distance Learning. 2010; 11 (1):161–177. [ Google Scholar ]
• Davis D., Chen G., Hauff C., Houben G.J. Activating learning at scale: A review of innovations in online learning strategies. Computers & Education. 2018; 125 :327–344. [ Google Scholar ]
• Delen E., Liew J., Willson V. Effects of interactivity and instructional scaffolding on learning: Self-regulation in online video-based environments. Computers & Education. 2014; 78 :312–320. [ Google Scholar ]
• Dixson M.D. Measuring student engagement in the online course: The Online Student Engagement scale (OSE) Online Learning. 2015; 19 (4):n4. * [ Google Scholar ]
• Dray B.J., Lowenthal P.R., Miszkiewicz M.J., Ruiz‐Primo M.A., Marczynski K. Developing an instrument to assess student readiness for online learning: A validation study. Distance Education. 2011; 32 (1):29–47. * [ Google Scholar ]
• Dziuban C., Moskal P., Thompson J., Kramer L., DeCantis G., Hermsdorfer A. Student satisfaction with online learning: Is it a psychological contract? Online Learning. 2015; 19 (2):n2. * [ Google Scholar ]
• Ergün E., Usluel Y.K. An analysis of density and degree-centrality according to the social networking structure formed in an online learning environment. Journal of Educational Technology & Society. 2016; 19 (4):34–46. * [ Google Scholar ]
• Esfijani A. Measuring quality in online education: A meta-synthesis. American Journal of Distance Education. 2018; 32 (1):57–73. [ Google Scholar ]
• Glazer H.R., Murphy J.A. Optimizing success: A model for persistence in online education. American Journal of Distance Education. 2015; 29 (2):135–144. [ Google Scholar ]
• Glazer H.R., Wanstreet C.E. Connection to the academic community: Perceptions of students in online education. Quarterly Review of Distance Education. 2011; 12 (1):55. * [ Google Scholar ]
• Hartnett M., George A.S., Dron J. Examining motivation in online distance learning environments: Complex, multifaceted and situation-dependent. International Review of Research in Open and Distance Learning. 2011; 12 (6):20–38. [ Google Scholar ]
• Harwell M.R. 2012. Research design in qualitative/quantitative/mixed methods. Section III. Opportunities and challenges in designing and conducting inquiry. [ Google Scholar ]
• Hung J.L. Trends of e‐learning research from 2000 to 2008: Use of text mining and bibliometrics. British Journal of Educational Technology. 2012; 43 (1):5–16. [ Google Scholar ]
• Jiang W. Interdependence of roles, role rotation, and sense of community in an online course. Distance Education. 2017; 38 (1):84–105. [ Google Scholar ]
• Ke F., Kwak D. Online learning across ethnicity and age: A study on learning interaction participation, perception, and learning satisfaction. Computers & Education. 2013; 61 :43–51. [ Google Scholar ]
• Kent M. Changing the conversation: Facebook as a venue for online class discussion in higher education. MERLOT Journal of Online Learning and Teaching. 2013; 9 (4):546–565. * [ Google Scholar ]
• Kim C., Park S.W., Cozart J. Affective and motivational factors of learning in online mathematics courses. British Journal of Educational Technology. 2014; 45 (1):171–185. [ Google Scholar ]
• Kizilcec R.F., Pérez-Sanagustín M., Maldonado J.J. Self-regulated learning strategies predict learner behavior and goal attainment in Massive Open Online Courses. Computers & Education. 2017; 104 :18–33. [ Google Scholar ]
• Kopp B., Matteucci M.C., Tomasetto C. E-tutorial support for collaborative online learning: An explorative study on experienced and inexperienced e-tutors. Computers & Education. 2012; 58 (1):12–20. [ Google Scholar ]
• Koseoglu S., Doering A. Understanding complex ecologies: An investigation of student experiences in adventure learning programs. Distance Education. 2011; 32 (3):339–355. * [ Google Scholar ]
• Kumi-Yeboah A. Designing a cross-cultural collaborative online learning framework for online instructors. Online Learning. 2018; 22 (4):181–201. * [ Google Scholar ]
• Kuo Y.C., Walker A.E., Belland B.R., Schroder K.E. A predictive study of student satisfaction in online education programs. International Review of Research in Open and Distance Learning. 2013; 14 (1):16–39. * [ Google Scholar ]
• Kuo Y.C., Walker A.E., Schroder K.E., Belland B.R. Interaction, Internet self-efficacy, and self-regulated learning as predictors of student satisfaction in online education courses. Internet and Higher Education. 2014; 20 :35–50. * [ Google Scholar ]
• Lee J. An exploratory study of effective online learning: Assessing satisfaction levels of graduate students of mathematics education associated with human and design factors of an online course. International Review of Research in Open and Distance Learning. 2014; 15 (1) [ Google Scholar ]
• Lee S.M. The relationships between higher order thinking skills, cognitive density, and social presence in online learning. The Internet and Higher Education. 2014; 21 :41–52. * [ Google Scholar ]
• Lee K. Rethinking the accessibility of online higher education: A historical review. The Internet and Higher Education. 2017; 33 :15–23. [ Google Scholar ]
• Lee Y., Choi J. A review of online course dropout research: Implications for practice and future research. Educational Technology Research & Development. 2011; 59 (5):593–618. [ Google Scholar ]
• Li L.Y., Tsai C.C. Accessing online learning material: Quantitative behavior patterns and their effects on motivation and learning performance. Computers & Education. 2017; 114 :286–297. [ Google Scholar ]
• Liyanagunawardena T., Adams A., Williams S. MOOCs: A systematic study of the published literature 2008-2012. International Review of Research in Open and Distance Learning. 2013; 14 (3):202–227. [ Google Scholar ]
• Lowes S., Lin P., Kinghorn B.R. Gender differences in online high school courses. Online Learning. 2016; 20 (4):100–117. [ Google Scholar ]
• Marbouti F., Wise A.F. Starburst: A new graphical interface to support purposeful attention to others' posts in online discussions. Educational Technology Research & Development. 2016; 64 (1):87–113. * [ Google Scholar ]
• Martin F., Ahlgrim-Delzell L., Budhrani K. Systematic review of two decades (1995 to 2014) of research on synchronous online learning. American Journal of Distance Education. 2017; 31 (1):3–19. [ Google Scholar ]
• Moore-Adams B.L., Jones W.M., Cohen J. Learning to teach online: A systematic review of the literature on K-12 teacher preparation for teaching online. Distance Education. 2016; 37 (3):333–348. [ Google Scholar ]
• Murphy E., Rodríguez-Manzanares M.A. Rapport in distance education. International Review of Research in Open and Distance Learning. 2012; 13 (1):167–190. * [ Google Scholar ]
• Nye A. Building an online academic learning community among undergraduate students. Distance Education. 2015; 36 (1):115–128. * [ Google Scholar ]
• Olesova L., Slavin M., Lim J. Exploring the effect of scripted roles on cognitive presence in asynchronous online discussions. Online Learning. 2016; 20 (4):34–53. * [ Google Scholar ]
• Orcutt J.M., Dringus L.P. Beyond being there: Practices that establish presence, engage students and influence intellectual curiosity in a structured online learning environment. Online Learning. 2017; 21 (3):15–35. * [ Google Scholar ]
• Overbaugh R.C., Nickel C.E. A comparison of student satisfaction and value of academic community between blended and online sections of a university-level educational foundations course. The Internet and Higher Education. 2011; 14 (3):164–174. * [ Google Scholar ]
• O'Shea S., Stone C., Delahunty J. “I ‘feel’like I am at university even though I am online.” Exploring how students narrate their engagement with higher education institutions in an online learning environment. Distance Education. 2015; 36 (1):41–58. * [ Google Scholar ]
• Paechter M., Maier B. Online or face-to-face? Students' experiences and preferences in e-learning. Internet and Higher Education. 2010; 13 (4):292–297. [ Google Scholar ]
• Phirangee K. Students' perceptions of learner-learner interactions that weaken a sense of community in an online learning environment. Online Learning. 2016; 20 (4):13–33. * [ Google Scholar ]
• Phirangee K., Malec A. Othering in online learning: An examination of social presence, identity, and sense of community. Distance Education. 2017; 38 (2):160–172. * [ Google Scholar ]
• Preisman K.A. Teaching presence in online education: From the instructor's point of view. Online Learning. 2014; 18 (3):n3. * [ Google Scholar ]
• Rowe M. Developing graduate attributes in an open online course. British Journal of Educational Technology. 2016; 47 (5):873–882. * [ Google Scholar ]
• Ruane R., Koku E.F. Social network analysis of undergraduate education student interaction in online peer mentoring settings. Journal of Online Learning and Teaching. 2014; 10 (4):577–589. * [ Google Scholar ]
• Ruane R., Lee V.J. Analysis of discussion board interaction in an online peer mentoring site. Online Learning. 2016; 20 (4):79–99. * [ Google Scholar ]
• Rye S.A., Støkken A.M. The implications of the local context in global virtual education. International Review of Research in Open and Distance Learning. 2012; 13 (1):191–206. * [ Google Scholar ]
• Saadatmand M., Kumpulainen K. Participants' perceptions of learning and networking in connectivist MOOCs. Journal of Online Learning and Teaching. 2014; 10 (1):16. * [ Google Scholar ]
• Shackelford J.L., Maxwell M. Sense of community in graduate online education: Contribution of learner to learner interaction. International Review of Research in Open and Distance Learning. 2012; 13 (4):228–249. * [ Google Scholar ]
• Shea P., Bidjerano T. Does online learning impede degree completion? A national study of community college students. Computers & Education. 2014; 75 :103–111. * [ Google Scholar ]
• Sherry L. Issues in distance learning. International Journal of Educational Telecommunications. 1996; 1 (4):337–365. [ Google Scholar ]
• Slagter van Tryon P.J., Bishop M.J. Evaluating social connectedness online: The design and development of the social perceptions in learning contexts instrument. Distance Education. 2012; 33 (3):347–364. * [ Google Scholar ]
• Swaggerty E.A., Broemmel A.D. Authenticity, relevance, and connectedness: Graduate students' learning preferences and experiences in an online reading education course. The Internet and Higher Education. 2017; 32 :80–86. * [ Google Scholar ]
• Tallent-Runnels M.K., Thomas J.A., Lan W.Y., Cooper S., Ahern T.C., Shaw S.M., Liu X. Teaching courses online: A review of the research. Review of Educational Research. 2006; 76 (1):93–135. doi: 10.3102/00346543076001093. [ CrossRef ] [ Google Scholar ]
• Tawfik A.A., Giabbanelli P.J., Hogan M., Msilu F., Gill A., York C.S. Effects of success v failure cases on learner-learner interaction. Computers & Education. 2018; 118 :120–132. [ Google Scholar ]
• Thomas J. Exploring the use of asynchronous online discussion in health care education: A literature review. Computers & Education. 2013; 69 :199–215. [ Google Scholar ]
• Thormann J., Fidalgo P. Guidelines for online course moderation and community building from a student's perspective. Journal of Online Learning and Teaching. 2014; 10 (3):374–388. * [ Google Scholar ]
• Tibi M.H. Computer science students' attitudes towards the use of structured and unstructured discussion forums in fully online courses. Online Learning. 2018; 22 (1):93–106. * [ Google Scholar ]
• Tsai C.W., Chiang Y.C. Research trends in problem‐based learning (pbl) research in e‐learning and online education environments: A review of publications in SSCI‐indexed journals from 2004 to 2012. British Journal of Educational Technology. 2013; 44 (6):E185–E190. [ Google Scholar ]
• Tsai C.W., Fan Y.T. Research trends in game‐based learning research in online learning environments: A review of studies published in SSCI‐indexed journals from 2003 to 2012. British Journal of Educational Technology. 2013; 44 (5):E115–E119. [ Google Scholar ]
• Tsai C.W., Shen P.D., Chiang Y.C. Research trends in meaningful learning research on e‐learning and online education environments: A review of studies published in SSCI‐indexed journals from 2003 to 2012. British Journal of Educational Technology. 2013; 44 (6):E179–E184. [ Google Scholar ]
• Tsai C.W., Shen P.D., Fan Y.T. Research trends in self‐regulated learning research in online learning environments: A review of studies published in selected journals from 2003 to 2012. British Journal of Educational Technology. 2013; 44 (5):E107–E110. [ Google Scholar ]
• U.S. Department of Education, Institute of Education Sciences . InstituteofEducationSciences; Washington,DC: 2017. What Works Clearinghouse procedures and standards handbook, version3.0. https://ies.ed.gov/ncee/wwc/Docs/referenceresources/wwc_procedures_v3_0_standards_handbook.pdf Retrievedfrom. [ Google Scholar ]
• Veletsianos G., Shepherdson P. A systematic analysis and synthesis of the empirical MOOC literature published in 2013–2015. International Review of Research in Open and Distance Learning. 2016; 17 (2) [ Google Scholar ]
• VERBI Software . 2019. MAXQDA 2020 online manual. Retrieved from maxqda. Com/help-max20/welcome [ Google Scholar ]
• Verstegen D., Dailey-Hebert A., Fonteijn H., Clarebout G., Spruijt A. How do virtual teams collaborate in online learning tasks in a MOOC? International Review of Research in Open and Distance Learning. 2018; 19 (4) * [ Google Scholar ]
• Wang Y., Baker R. Grit and intention: Why do learners complete MOOCs? International Review of Research in Open and Distance Learning. 2018; 19 (3) * [ Google Scholar ]
• Wei C.W., Chen N.S., Kinshuk A model for social presence in online classrooms. Educational Technology Research & Development. 2012; 60 (3):529–545. * [ Google Scholar ]
• Wicks D., Craft B.B., Lee D., Lumpe A., Henrikson R., Baliram N., Wicks K. An evaluation of low versus high collaboration in online learning. Online Learning. 2015; 19 (4):n4. * [ Google Scholar ]
• Wise A.F., Perera N., Hsiao Y.T., Speer J., Marbouti F. Microanalytic case studies of individual participation patterns in an asynchronous online discussion in an undergraduate blended course. The Internet and Higher Education. 2012; 15 (2):108–117. * [ Google Scholar ]
• Wisneski J.E., Ozogul G., Bichelmeyer B.A. Does teaching presence transfer between MBA teaching environments? A comparative investigation of instructional design practices associated with teaching presence. The Internet and Higher Education. 2015; 25 :18–27. * [ Google Scholar ]
• Wladis C., Hachey A.C., Conway K. An investigation of course-level factors as predictors of online STEM course outcomes. Computers & Education. 2014; 77 :145–150. * [ Google Scholar ]
• Wladis C., Samuels J. Do online readiness surveys do what they claim? Validity, reliability, and subsequent student enrollment decisions. Computers & Education. 2016; 98 :39–56. [ Google Scholar ]
• Yamagata-Lynch L.C. Blending online asynchronous and synchronous learning. International Review of Research in Open and Distance Learning. 2014; 15 (2) * [ Google Scholar ]
• Yang J., Kinshuk, Yu H., Chen S.J., Huang R. Strategies for smooth and effective cross-cultural online collaborative learning. Journal of Educational Technology & Society. 2014; 17 (3):208–221. * [ Google Scholar ]
• Yeboah A.K., Smith P. Relationships between minority students online learning experiences and academic performance. Online Learning. 2016; 20 (4):n4. * [ Google Scholar ]
• Yu T. Examining construct validity of the student online learning readiness (SOLR) instrument using confirmatory factor analysis. Online Learning. 2018; 22 (4):277–288. * [ Google Scholar ]
• Yukselturk E., Bulut S. Gender differences in self-regulated online learning environment. Educational Technology & Society. 2009; 12 (3):12–22. [ Google Scholar ]
• Yukselturk E., Top E. Exploring the link among entry characteristics, participation behaviors and course outcomes of online learners: An examination of learner profile using cluster analysis. British Journal of Educational Technology. 2013; 44 (5):716–728. [ Google Scholar ]
• Zawacki-Richter O., Backer E., Vogt S. Review of distance education research (2000 to 2008): Analysis of research areas, methods, and authorship patterns. International Review of Research in Open and Distance Learning. 2009; 10 (6):30. doi: 10.19173/irrodl.v10i6.741. [ CrossRef ] [ Google Scholar ]
• Zhu M., Sari A., Lee M.M. A systematic review of research methods and topics of the empirical MOOC literature (2014–2016) The Internet and Higher Education. 2018; 37 :31–39. [ Google Scholar ]
• Zimmerman T.D. Exploring learner to content interaction as a success factor in online courses. International Review of Research in Open and Distance Learning. 2012; 13 (4):152–165. [ Google Scholar ]