Blended Learning in Algeria: Assessing Students’ Satisfaction and Future Preferences Using SEM and Sentiment Analysis

Abstract

Given the still existing restrictions of COVID-19, blended learning is undoubtedly becoming a better-fitting strategy for higher education institutions in underprivileged countries. Acknowledging the current changes in higher education, this study aims to investigate the elements that influence students’ satisfaction and their future preferences regarding blended learning in Algeria. A total of 782 questionnaires were collected from different Algerian universities. A structural Equation Modeling (SEM) analysis was conducted to investigate the relationship among the latent variables of the proposed theoretical model. Moreover, an unsupervised sentiment analysis approach was applied to analyze the qualitative data received in the form of feedback from the participants. The results show that students’ perceived ease of use and perceived usefulness of blended learning had a significant positive impact on their satisfaction. Similarly, satisfaction had a positive influence on students’ future preferences regarding blended learning. In turn, students’ perceived ease of use and usefulness had an indirect effect on their future preferences, mediated by satisfaction. Additionally, qualitative data echoed students’ eagerness to adopt more advanced learning technologies and what obstacles currently stand in their way. The contribution of this study is to reflect the current situation of blended learning adoption in developing countries and to support future curriculum planning and development. It can also help teachers, students, and policymakers to make better decisions and recommendations for an improved and more sustainable learning and teaching environment in the future.

Introduction

The development of educational technology is influenced not solely by technological progress, but also by politics and economics (Altmann, 2017). Recently, however, health is becoming a key influencing factor that drives and hastens the evolution of learning technology. Since December 2019, a new disease known as COVID-19 has triggered a pandemic worldwide, causing changes in many aspects of our social life. This pandemic had a significant impact on higher education. Many higher education institutions shifted to online learning to alleviate the negative effects of closing buildings and limiting face-to-face interactions as a strategy to restrain the spread of the virus. Undoubtedly, educational resources and knowledge are now more readily available to students, accelerating the learning process. Nevertheless, it is undeniable that the sudden transition to online learning has caused considerable difficulties (Abu Talib et al., 2021). Despite the exaggerated promises and predictions of early research, online learning is still far from attaining its full potential of being the primary educational method of the future.

In developing countries, the challenges of online learning are more eminent as they are unprepared for the rapid migration to online learning due to incompetence, inequality, and lack of access to the necessary technologies. Moreover, teachers and students have not been able to quickly acclimate to a completely new distance learning strategy. As the world continues to adjust to the COVID-19 crisis, the debate on post-pandemic pedagogy has begun, and a blended learning approach will most likely become a viable technique for mediating the transition from face-to-face to online learning, especially for underprivileged countries.

Blended learning is gaining popularity in the higher education community because it offers the benefits of both face-to-face and online learning modalities. For instance, blended learning offers flexibility in accessing content and information (Huang, 2021). It can also enhance students’ self-direction and self-reflection (Bamoallem & Altarteer, 2022), and increase their curiosity and engagement (Garrison & Vaughan, 2007).

Algerian higher education, like that of many other developing countries, has been struggling with various obstacles to meet the demands of globalization and the information society since long before the outbreak of COVID-19. In addition, the pandemic may have ushered in a lasting change in the way higher education is managed in Algeria. Recently, the Minister of Higher Education and Scientific Research stated that blended learning is a potential permanent strategy for Algerian higher education institutions.

The discussion of the effectiveness of the blended learning experience in Algerian higher education during COVID-19 has not been given the necessary attention in the literature. Whereas most research related to blended learning has been conducted in Western and Asian nations (e.g., (Clark & Post, 2021; Huang, 2021; Jones & Sharma, 2021)), this study focuses on Algeria, a North African country where the diffusion of the Internet and the adoption of online learning and teaching has been rather gradual.

The main objective of this paper is to investigate students’ acceptance and satisfaction with blended learning. More specifically, we examine factors that influence students’ satisfaction, and the impact of these factors on students’ future preferences regarding blended learning. We paid particular attention to the exceptional situation of Algerian higher education institutions during COVID-19 and the implementation of blended learning, which was primarily driven by the government’s initiative to promote the inclusion of information and Communication Technologies (ICTs) in education. This work is an important step toward our long-term objective of developing a more resilient blended learning design that is fit for underrepresented countries.

The next section of this paper presents a larger picture of our context in which blended learning and the use of ICTs in Algerian higher education are addressed. The underlying theoretical foundation of this study is also presented in the second section. In the third section, we describe our research methodology, including the profile of the target population, the data collection process, and the analysis techniques. In the fourth section, we present our findings, followed by a discussion and comparison with the existing literature in the fifth section. We conclude the paper in the sixth section by addressing the limitations of this study and drawing implications for future research.

Literature Review

Blended Learning

Blended learning has been a popular concept in higher education over the past few decades, associated with the hype surrounding online learning. Implementing a blended learning solution entails selecting from an ever-expanding variety of modalities and tools (Inoue, 2010). It involves combining different forms of delivery while concentrating on optimizing the achievement of learning objectives and the cost of learning programs during a specific timeframe (Bamoallem & Altarteer, 2022). More precisely, blended learning is a combination of asynchronous online content (often text or video-based) and synchronous conventional face-to-face learning (Clark & Post, 2021). The blending can be “robust” or “weak” depending on the extent and ferocity with which online learning technologies are used. Recently, the notion of blended learning has gained prominence in higher education. The term is now increasingly recognized as referring to a combination of learning and teaching activities that include active face-to-face instruction in lieu of traditional lectures, as well as dynamic online activities and information that enable learning/teaching at any time and from any location (Jones & Sharma, 2021).

Students’ intrinsic learning needs should be addressed in both online and face-to-face classrooms for a blended learning mode to thrive. This implies that autonomy, competence, engagement, and support should be reinforced in an online environment, while in face-to-face classes, traditional lectures should be replaced with more engaging activities (Chiu, 2021).

The use of ICTs and Blended Learning in Algerian Higher Education

With the rapid advancement of the information age, the incorporation of ICTs into education has become a major priority in many developing countries. In Algerian higher education, the use of ICTs is still in its infancy. In 2006, an e-learning system was established in conjunction with foreign companies such as Microsoft in an attempt to incorporate ICTs into Algerian higher education (Bin Herzallah, 2021). In addition, updated classrooms and virtual laboratories were developed (Mebarkia et al., 2016), and additional efforts were made to leverage social media as an e-learning tool (Laifa, 2018).

Furthermore, several studies have been carried out to examine the benefits and drawbacks of deploying technology in Algerian higher education programs (Berbar & Ait Hamouda, 2019; Kouninef et al., 2015; Ladaci, 2017). These studies emphasize that Algerian students and instructors are willing to include ICTs in the classroom because they consider it vital to assist the learning process and acquire the necessary skills. Nonetheless, a lack of training and administrative support, the large number of students in universities, a paucity of resources, and an unstable Internet connection are obstacles that prevent instructors and students from fully exploiting the potential of these technological tools. Moreover, it is difficult to raise the necessary funds to make ICTs available in all Algerian higher education institutions, as higher education in Algeria is free to the public.

When COVID-19 reached Algeria in March 2020, universities seized all activities and remained closed for six months owing to two main reasons. First, online-only learning is not an option because ICTs were not well implemented. Second, faculty and students were unprepared for an abrupt shift to online learning strategies. As a result, Algerian institutions have opted for the less extreme option of blended learning to avoid jeopardizing the academic year.

This mixed learning model is still being used in the third year after the pandemic outbreak, as the country continues to face new waves of new virus strains. Furthermore, the Algerian Minister of Higher Education and Scientific Research stated that blended learning is being studied as a possible permanent plan for Algerian higher education.

Challenges of Blended Learning in Developing Countries

Although blended learning has been used in many developing countries for over a decade, there are still various challenges they are facing in this regard, which is calling for further examination. The most common challenges can be categorized in three main groups:

• Technological accessibility issues = A key challenge for low- and middle-income countries when it comes to blended learning is ensuring that the necessary internet access and computing equipment is available (Ashraf et al., 2021; Bordoloi et al. 2021). As many countries are still struggling to provide strong internet infrastructure and logistical support, teachers and students are often faced with the financial burden of internet and personal computers (Ashraf et al., 2021). One solution is to provide computer laboratories on higher education institutions campuses (Namyssova et al., 2019). However, Accessibility challenges extend beyond internet and devices, as some developing countries, such as Bangladesh, are still contending with providing a consistent power supply (Chowdhury, 2020).

• Quality and skills issues = because technology use in education in developing countries is relatively new compared to developed countries, teachers and staff lack ICTs competencies and require adequate training (Ashraf et al., 2021). Moreover, many studies reported that the current designed educational software and mobile apps are inadequate and do not meet the needs of educators and students who are new to blended learning (Al-Hunaiyyan et al., 2021). Inadequacy transcends the tools design and reach the quality of designed learning materials and courses content due to insufficient planning and lack of funding. In many cases, blended learning is introduced in a case of emergency without the proper policy to guide that shift nor enough time to prepare for it (Chowdhury, 2020).

• Equity and social issues = Incorporating traditional instructions and online learning is a difficult system to implement when quality education is not accessible to all in most economically disadvantaged countries. Additionally, teachers and instructors reported a lack of support while shifting to blended learning leaving them overburdened with courses adjustment and technological adaptation (Tupas & Linas-Laguda, 2020). On the other hand, students’ isolation and need for engagement seemed to decrease the acceptance of blended learning (Dwivedi et al., 2019). Finally, a major challenge is lack of awareness of stakeholders and their resistance to blended learning. While technological issues can be solved by providing cheaper and better equipment and infrastructure, changing rigid mindsets is more challenging in many developing countries (Islam et al., 2022).

Blended learning is relatively a recent concept in the teaching and learning arena in Algeria as compared to other developing nations. Despite the efforts provided by the government to ease the shift towards blended learning, Algerian higher education institutions are facing similar impediments that other countries are facing. For instance, it was reported before COVID-19 that both teachers and students were against a blended learning strategy for two main reasons. First, instructors and students acquired a fear of technology because they were aware of the accessibility problems and lack of technical assistance they will face (Guessabi, 2021). Second, students preferred in-class communication and engagement to be able to discuss the studied topic and to avoid feeling isolated (Rahmani & Zitouni, 2022).

Another reason that is keeping both parties resistant to blended learning is the current higher education system LMD (Licence/Bachelor, Master, Doctorate) that Algeria has adopted since 2004 where students are required to attend roughly 400 h of class over the course of a 16-week period (Sarnou et al., 2012). One proposed solution was dividing each group of students in a class into sub-groups and alternate between their attendance. Due to the large number of students in a single normal group (40 on average), teachers’ workload doubled with no compensation (Guessabi, 2021). Finally, a lack of confidence seemed to stifle instructors’ and students’ enthusiasm, who viewed themselves as having rudimentary computer skills but lacking the expertise necessary to adequately use blended learning (Amouri & Benyagoub, 2022).

Theoretical Model

Many models and theories of new technology acceptance have been proposed to predict and explain users’ behavior toward new technologies or information systems. Davis’ Technology Acceptance Model (TAM) (Davis, 1989) is one of the most informative and powerful models that has been widely used in many contexts. The TAM model assumes that users’ perceptions of the usefulness and ease of use of a technology or system are the key factors in its acceptance. In the TAM, users’ attitudes toward a particular technology, and thus their decision to adopt it, are impacted by their perceptions of how useful and easy it is to use (Davis, 1989). This model has been used extensively to identify the factors that affect students’ satisfaction and intentions to use various types of learning systems (Al-Maroof et al., 2021; Mustafa & Garcia 2021).

To assess the initial adoption of blended learning by students in this study, we used perceived usefulness and perceived ease of use. Perceived usefulness is the extent to which one believes that adopting a given system can increase one’s performance (Mailizar et al., 2021). Perceived ease of use, on the other hand, indicates the extent to which one trusts that using a particular system is easy and effortless (Bazelais et al., 2018).

Based on TAM, perceived ease of use has a positive impact on perceived usefulness. This has been confirmed in many studies on the use of new information technologies or new strategies in the learning environment in general (Joo et al., 2018; Laifa, 2018; Wu & Chen, 2017) and in a blended learning context in particular (Huang, 2021). Therefore, we aim to test the same hypothesis in our research model:

• H1. Perceived ease of use has a positive effect on perceived usefulness.

Students’ satisfaction alludes to the learner’s positive subjective and internal emotions evoked by the learning process (Huang, 2021). Learners feel satisfied when they reach or surpass expectations of their learning performance. According to the TAM, the perceived ease of use and perceived usefulness of a system also influence the user’s behavioral intention to use it in the future. In the context of blended learning, this suggests that the perceived usefulness and ease of use of a blended learning strategy will improve students’ opinions of their learning experience, making them more inclined to be satisfied with the blended learning strategy and consequently prefer it over a traditional face-to-face learning strategy (Lei & So, 2021). Hence, we hypothesize the following hypotheses:

• H2. Perceived ease of use has a positive effect on satisfaction.

• H3. Perceived usefulness has a positive effect on satisfaction.

A satisfying learning experience encourages investment in learning efforts and continued participation in learning activities (Huang, 2021). This leads to the continuation and recurrence of the learning process, creating a positive cycle. In this study, we consider learning satisfaction as a variable that mediates the relationship between students’ initial acceptance of blended learning and their future preferences regarding whether to continue to use the system or not. Following the work of Lei and So (Lei & So, 2021) and Joo et al. (Joo et al., 2018), this conjecture is based on the Exception-Confirmation Model (ECM) (Bhattacherjee, 2001). According to the ECM, individuals’ inclination to continue using a particular service or system is mainly determined by their satisfaction with previous experiences.

A number of studies from several disciplines have frequently established the association between satisfaction and future preferences. According to previous research, students’ satisfaction with their online learning experiences had a significant impact on their future preferences and behavioral intentions. For instance, Wang et al. (Wang et al., 2021) found that learners’ perceived satisfaction had a positive effect on their intention to continue using MOOCs during COVID-19. In Korea, students’ satisfaction with a Korean-MOOC course had a significant and positive effect on students’ intention to continue using the course in the future (Joo et al., 2018). Similarly, research has found that students’ and instructors’ satisfaction with online learning has a significant and positive influence on future preferences (Lei & So, 2021). Accordingly, we propose that:

• H4. Students’ satisfaction has a positive effect on their future preferences.

Methodology

Research Objective and Conceptual Model

Although there are few studies on the effectiveness of blended learning in developing and underprivileged African countries, blended learning remains a beneficial learning strategy while the world is still facing new waves of COVID-19. As the higher education community is keenly interested in learning how blended learning can lead to improved outcomes and academic performance, in this study we provide an exploration of students’ acceptance of and satisfaction with blended learning during COVID-19.

Given the relative newness of Internet-based education in Algeria, our study is particularly suited to investigate the elements that influence students’ satisfaction and their future preference using technology acceptance theory, and an unsupervised sentiment analysis approach.

For this end, we first proposed the following conceptual model (Fig. 1) based on the previous studies presented in the preceding section. We aimed to test students’ initial experience and satisfaction with blended learning (H1, H2, H3) and their acceptance of blended learning by examining their future preferences (H4). We hypothesized three factors: perceived ease of use, perceived usefulness, and satisfaction. We examined the relationship between these factors and how they affect students’ future preferences regarding blended learning.

Fig. 1.

Conceptual model

To achieve our goal, we relied on a mixture of qualitative and quantitative data collected in an Algerian context during COVID-19. We then used structural equation modeling to test the conceptual model. The qualitative data, on the other hand, were analyzed using an unsupervised sentiment analysis approach.

Context and Data Collection

For data collection, we developed a questionnaire following the scale development procedure recommended by (Churchill, 1979). The created questionnaire contained three sections. The first section was devoted to measuring different constructs of our conceptual model, using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Items were taken from three distinct survey instruments that examined the same variables in comparable contexts.

The original items were adapted to fit our context and then translated into Arabic by two independent researchers, followed by a back translation. The translated statements were then compared and appropriately corrected or further updated. Some items were eliminated because their meaning was redundant when translated into Arabic. Table 1 lists the final measurement items used. It is worth noting that we used the term “hybrid” and not “blended” in the questionnaire because the official Arabic terminology used in Algerian institutions to define blended learning is “اﻟﮭﺠﯿﻦ”, which translates to hybrid.

Table 1.

Measured variables

Constructs	Items	Source
Perceived usefulness	PU1- The hybrid learning approach helps me learn better. PU2- The hybrid learning approach is more useful than the traditional face-to-face teaching. PU3- The hybrid learning approach is more useful than the purely online approach. PU4- I find hybrid learning useful for my studies. PU5- Using hybrid learning would enhance my productivity.	(Huang, 2021), (Arbaugh, 2000)
Perceived ease of use	PEU1- Adopting hybrid learning is not difficult for me. PEU2- The learning activities in hybrid learning are easy to understand and follow. PEU3- I quickly learned to apply hybrid learning. PEU4- I think the hybrid learning approach is easy to use.	(Huang, 2021)
Satisfaction	SAT1- I think studying courses in a hybrid learning approach is a wise choice SAT2- I feel this hybrid way of learning satisfies my learning needs very well. SAT3- I feel the quality of the courses I took was largely unaffected by the hybrid learning approach. SAT4- Overall, I am satisfied with the hybrid learning arrangement during the coronavirus situation. SAT5- Overall, the attitudes and performances of my instructors who taught hybrid classes are acceptable.	(Huang, 2021), (Arbaugh, 2000), (Lei & So, 2021)
Future preferences	FP1- Overall, I prefer hybrid learning more than traditional face-to-face classes. FP2- Overall, I prefer traditional face-to-face classes more than hybrid learning.	(Lei & So, 2021)

The second section of the questionnaire contained seven demographic questions. The third and final section was an open-ended question about students’ thoughts and feedback on the blended learning experience.

To collect our data, we followed the following sampling process.

1. Defining our target population: Students from Algerian higher education institutions were the target demographic of our present study.

2. Determining the sample size: About 1.5 million students were enrolled in Algerian Higher Education institutions between 2020 and 2021. Following Bartlett et al. (2001) recommendations, a sample size of about 1036 is required to represent our studied population, with a margin of error of 4%, and a confidence level of 99%.

3. Choosing the sampling method: Because there were stay-at-home and quarantine regulations in effect at that time and few students were permitted into institutions, it was difficult to reach students for participation in the study. For this reason, we relied on the snowball sampling method. This technique is used in many fields when it is hard to locate the subjects, and it is known for being fast and cost effective (Taherdoost, 2016).

4. Collecting data: Once the targeted population and sample selection and size were determined, we started the data collection process. The questionnaire was distributed online between June and July 2021. To reach a wide range of students, we distributed the questionnaire via email to students we were teaching that semester and we encouraged them to share the link to the questionnaire with their fellow students via groups and social media channels of Algerian higher education students.

Structural Equation Modeling

Structural Equation Modeling (SEM) is a multivariate statistical technique that allows researchers to examine the relationships between latent and observed variables (Kline, 2016). It consists of a series of mathematical models and statistical methods that fit constructs to specific data to test, validate, or extend a theoretical model. SEM provides a thorough explanatory analysis with a required level of quality (Schumacker & Lomax, 2016).

SEM assesses both the relationship between latent constructs and their indicators, known as the measurement model, and the relationships between constructs, known as the structural model (Bagozzi & Yi, 1988). In addition, SEM enables testing models with different levels of hierarchically structured data. A well-known feature of SEM is its ability to model mediating variables and error terms (Schumacker & Lomax, 2016).

However, there are several assumptions and limitations of SEM that researchers should be aware of when using this technique. SEM assumes that the data is normally distributed, the relationships between variables are linear, there is no multicollinearity between predictor variables, there is no measurement error in the observed variables, and there is no missing data (Kline, 2016). Additionally, SEM requires a large sample size, can become very complex, assumes that the model set by the researcher is correct, and can only detect correlations between variables but not causality (Kline, 2016).

These assumptions and limitations may impact the accuracy and generalizability of the results and underscore the need for researchers to carefully consider and report on these factors when using SEM in their survey research. Despite its limitations, SEM remains a widely used statistical technique for survey research that provides a powerful tool for analyzing complex models and test hypotheses about the relationships between variables (Kline, 2016).

Sentiment Analysis

Sentiment analysis is the use of natural language processing and machine learning to systematically identify, extract, quantify, and study affective and subjective information from human-generated text data. Sentiment analysis is commonly applied to analyze user documents such as comments from surveys and online sources (Rani & Kumar, 2017).

Manual analysis of qualitative data in survey responses can be an endless effort and time-consuming labor. Instead, examining qualitative survey data using natural language processing techniques and tools can provide better and more accurate insights. To analyze our textual data collected from students’ feedback, we used a rule-based model for general sentiment analysis known as Valence Aware Dictionary for Sentiment Reasoning (VADER). VADER is a text sentiment analysis model that relies on a dictionary linking lexical features - combined with general rules embodying grammatical and syntactic conventions - to emotion intensities called sentiment scores (Hutto & Gilbert, 2014). The sentiment score of a text can be calculated by summing the intensity of each word in the lexicon, adjusting it according to the rules, and then normalizing it to a value between − 1 ( extremely negative) and + 1 ( extremely positive). The following formula is used for normalization:

$$x=\frac{x}{\sqrt{x^2+\alpha }}$$

where x is the total score of the constituent words and α is the normalization constant.

Our methodology for qualitative data analysis involves 5 steps.

1. Translation: the comments received were mostly in English. However, some of them were written in Arabic or French. Therefore, we translated non-English comments into English to ensure consistency in the sentiment analysis.

2. Lowercasing: to facilitate the process of matching words in students’ comments, characters were converted to lowercase.

3. Removal of irrelevant content: to increase the efficiency of the model, stop words, unnecessary punctuation, and emoticons were removed.

4. Tokenization: students’ comments were split into tokens (i.e., individual separate words).

5. Sentiment analysis: after cleaning and preprocessing the data, VADER was used to generate sentiment scores for each comment. The compound scores were then next converted into labels: “Positive”, “Negative” or “Neutral”.

Results

Demographics and Descriptive Statistics

We received a total of 830 anonymous responses from students, with a response rate of 80.11%. After reviewing the collected data, 48 invalid responses were removed, and the remaining 782 valid responses (47.6% males) were used for further analysis. Note that 60% of enrolled students in Algerian tertiary education were females (Statista, 2022), making our sample more representative of our targeted population.

The characteristics of the students who participated in the present study are presented in Table 2. It is worth mentioning that the total number of enrolled students in each department is about 850 students, thus, our sample is sufficient. The majority of respondents were from BBA University (64.7%) and M’sila University (32.6%). Sétif University and Batna University were among the “others”.

Table 2.

Descriptive statistics of students’ profiles

Students (n = 782)
Variable	N	%
Gender
Male	372	47.6
Female	410	52.4
Age
17–19	20	2.6
20–23	451	57.7
23–25	219	28.0
26–29	38	4.9
>=30	54	6.9
Institution
BBA	506	64.7
M’sila	255	32.6
Others	21	2.7
Department
Computer sciences	637	81.5
English	131	16.8
Other	14	1.8
Level of education
BS	507	64.8
MSc	264	33.8
PhD	11	1.4
Learning location
Home	669	85.5
University library	29	3.7
Friend/Family member	35	4.5
Others	49	6.3
Preferred language
Arabic	298	38.1
English	347	44.4
French	137	17.5

If each measurement variable in the SEM does not have a normal distribution, a biased estimate is generated, and adequate statistical inspection is impossible. To test the assumption of multivariate normality, the mean, standard deviation, skewness, and kurtosis were examined.

The descriptive statistics for the latent variables in Table 3 reveal that the data is narrowly scattered around the mean. Additionally, the skewness index of our data ranges from − 0.3 to 0.15 (not more than |2.3|), and the kurtosis index ranges from − 0.70 to 2.53 (not more than |7|), reflecting a reasonable degree of multivariate normality in the data (Kallner, 2018).

Table 3.

Evaluation of the validity and reliability of the scale

Items	Factor loading	Explained variance (%)	Cronbach’s Alpha	Mean	SD	Skewness	Kurtosis
Perceived usefulness		60.97	0.83	2.77	0.97	0.15	-0.51
PU1	0.860			2.62	1.30
PU2	0.754			2.19	1.27
PU3	0.483			3.34	1.28
PU4	0.873			2.88	1.25
PU5	0.864			2.81	1.23
Perceived ease of use		71.81	0.87	3.01	1.01	-0.06	-0.42
PEU1	0.815			3.13	1.17
PEU2	0.845			2.87	1.20
PEU3	0.869			2.97	1.20
PEU4	0.859			3.08	1.20
Satisfaction		62.05	0.85	2.84	0.99	0.05	-0.70
SAT1	0.847			2.85	1.28
SAT2	0.829			2.65	1.25
SAT3	0.717			2.72	1.24
SAT4	0.823			2.91	1.36
SAT5	0.711			3.08	1.18
Future preferences
FP	-		-	2.35	1.34	-0.3	2.53

Construct Validity and Reliability

The results of the exploratory factor analysis (EFA) to assess the construct validity of the measurement scales are presented in Table 3. The EFA was conducted using the principal components method. The factor loadings for each item and the explained variances are also presented in Table 3. All items had acceptable factor loadings (λ > 0.40), and the variance of the items explained by the scales was also large enough (> 50%), indicating the validity of the constructs. Cronbach’s alpha, as a measure of inter-consistency reliability, was calculated for the scales and showed acceptable reliability of the instruments (alpha > 0.70).

The results of the Pearson (and Spearman) correlation test between the measurement scales are presented in Table 4. All scales showed pairwise strong and significant positive associations (P < 0.001), indicating a strong relationship between the variables.

Table 4.

Correlation matrix

	A	B	C	D
A. Perceived ease of use	1
B. Perceived usefulness	0.72	1
C. Satisfaction	0.72	0.78	1
D. Future preferences	0.55	0.65	0.60	1

Structural Model

To ensure that the proposed model described the observed data, various model fit indices were tested (see Table 5). The original model did not have enough acceptable fit indices according to (Bagozzi & Yi, 1988; Hooper et al., 2008; Schumacker & Lomax, 2016). As a result, we modified the model by entering additional parameters to reduce the chi-square value and improve the fit indices of the model, so a modified model was proposed and tested. These modifications are theoretically justified and are addressed in the discussion section. The fit indices of the modified model, listed in the second row of Table 6, meet the guidelines, thus, the model is a good fit for the current sample.

Table 5.

Data model fit indices

Recommended guidelines	Chi-square (x2)	df	SRMR	RMSEA	GFI	AGFI	CFI	NFI	IFI
			< 0.05	≤ 0.08	> 0.90	≥ 0.90	≥ 0.90	> 0.90	> 0.90
Original Model	974.9	87	0.048	0.090	0.89	0.85	0.92	0.91	0.92
Modified Model	493.5	83	0.040	0.080	0.92	0.90	0.94	0.93	0.94

Table 6.

SEM results for the modified model

Independent variables	Dependent variables	Direct effect β, P-value	Indirect effect β, P-value
Ease of use →	Usefulness	0.86, < 0.001
Ease of use →	Satisfaction	0.16, 0.015	0.72, 0.002
Usefulness →	Satisfaction	0.83, < 0.001
Ease of use →	Future preferences		0.85, 0.003
Usefulness →	Future preferences		0.81, 0.003
Satisfaction →	Future preferences	0.97, < 0.001

Figure 2 illustrates the results of SEM for the modified models. Table 6 presents the results in detail. Based on the modified model, the direct positive association between perceived ease and perceived usefulness (β = 0.86, P < 0.001) and satisfaction (β = 0.16, P = 0.015) was significant. Perceived usefulness directly increased satisfaction (β = 0.83, P < 0.001) and satisfaction had a significant positive association with future preferences (β = 0.97, P < 0.001). For indirect effects calculated and tested using the bootstrapping method, perceived ease of use indirectly increased satisfaction over perceived usefulness (β = 0.72, P = 0.002). In addition, perceived ease of use (β = 0.85, P = 0.003) and perceived usefulness (β = 0.81, P = 0.003) significantly increased future preferences indirectly via increasing satisfaction.

Fig. 2.

SEM results

Sentiment Analysis

In total, we received 122 comments from respondents. Qualitative data can provide considerably more insights into our sample. Sentiment analysis used lexical signifiers of emotion to classify the comments received into three main categories: positive, negative, or neutral.

As can be seen in Fig. 3, most of the comments received were positive (55.73%), while 27.04% of the comments were negative. The remaining 17.23% were neutral. Once the comments were classified, we took a closer look at each category.

Fig. 3.

Labels distribution

Positive comments included mostly gratitude notes to teachers for their efforts during the transition to blended learning. For example:

Thank you professor for all your efforts with us. You gave us the best.

Likewise, “Thank you so much”, and “Thank you” were frequently used to end positive feedback.

Some students spoke in favor of blended learning and increased incorporation of online learning, because it provides more flexibility, less time wasted on campus, and more productivity. Here are some examples:

Hybrid education and distance learning is better for those who have occupations other than studying, as this technology enabled me to learn […] and also work at the same time, which means 0 wasted time.

Distance or hybrid learning contributes more to the productivity of the student who tries at home and learns everything new to advance in any field towards building and developing [skills] and competing with everything that is modern now, especially via the Internet. Thank you so much.

I feel like studying this way [...] is great, first of all because it gives me free time to go do my research about what I was studying. [...] teachers are giving their 100% since they are controlled [by limited] time so we are more efficient …

Although students find BL exciting and useful, they are aware of the problems they face, such as the way BL has been implemented in Algeria and Internet access and quality.

Hybrid learning is good, but with the obligatory application of business-oriented classes and practical work and their attendance, and with the study of the lessons online in full, for example, in the Wilaya of M’sila, we studied half of the classes online and half of the classes in [person] and they were far apart, so it is better to study them all online, thank you.

Overall, I think the hybrid learning is way better than the traditional one but there are some issues that can affect its efficiency like the quality of the internet connection and the commitment of students and the [coordination] between the online and face to face method.

Neutral comments, on the other hand, were either: suggestions, requests, or statements of fact, as the following examples show:

• Suggestions: e.g., “We suggest returning to traditional learning.”

• Requests: e.g., “The time [dedicated] for traditional face-to-face learning should be much greater than for distance learning.”, “I prefer to study in class with teachers because the interaction [is] better”.

• Stating facts: e.g., “Hybrid learning is nice if implemented wisely.”, “in Algeria, the quality of the internet introduces big difficulties.”

Even though these comments have a neutral tone, they tend to be less positive towards blended learning as it is currently implemented in Algerian universities. This is consistent with the quantitative data, as we find that students tend to not be very satisfied with BL (M = 2.84, SD = 0.99).

Finally, negative comments reflect students’ dissatisfaction with blended learning overall. For instance, one student stated: “Hybrid learning has shown its abject failure in our universities.”

Another student said, “Just this year is GONE in a bad way for the study situation without course classes… I think that online teaching is not beneficial for students”.

Other students, however, are not bothered by blended learning, but by the problems they faced. We summarize these problems below, along with examples from the data:

• Intense lessons: depending on the course, some lessons were stuffed to reduce class time, and students were unable to keep up with the workload assigned at home (e.g., “Hybrid learning is good, but stuffing the lessons is not a good thing.”).

• Inadequate teaching methods: unfortunately, some teachers did not adapt their teaching to a blended learning environment which caused students to struggle with keeping up. (e.g., “Hybrid learning is good but in the last pandemic, teachers didn’t really put their efforts into explaining instead they took advantage of this new hybrid learning to charge students from most levels and programs with projects and homeworks that they don’t even know what they are really about. So no, I highly disagree with this new method and I highly think that the problem is coming from our dear teachers.”)

• Adaptation and confusion: some students found it more difficult to adapt to the requirements of blended learning and were confused. One of the main reasons students reported was the lack of social presence and interaction. To illustrate, students stated the following:

We thank the teachers for their efforts in teaching us remotely, but we faced many difficulties in adapting, especially due to the poor internet and inappropriate teaching time

In short, we do not want hybrid learning. We want to study like previous years (face to face traditional learning) or distance education. As for hybrid learning, it has become a confusing and frustrating matter for us as students.

Distance learning did not help me with anything because it needed to study at home and this did not help me because I am one of the people who study in the department and pay attention and focus with the professors, …

• Lack of engaging activities and practice: other students also reported being dissatisfied with the way blended learning was implemented, as tutorials where students could directly engage with their instructors and classmates were virtually nonexistent.

very sad of how this was implemented.

It was the approach in which this hybrid learning idea was implemented that pretty much killed its potential. I’m fine with online learning in general also I find face-to-face learning a bit overrated anyways. The actual problem is cutting down tutorials and practice hours, and the execution could have been way tighter and overall just more organized.

The application is the basis for learning, and so I was not able to absorb and understand the lessons as before, …

Discussion

The main objective of this study is to investigate students’ acceptance of and satisfaction with blended learning. We posited four conjectures. Theoretically, we hypothesized that students’ perceived ease of use of blended learning would influence their perceived usefulness of blended learning. We also hypothesized that both perceived ease of use and perceived usefulness would have a direct positive effect on students’ satisfaction with blended learning. Moreover, we assumed that students’ satisfaction with their blended learning experience would influence their future preferences regarding blended learning. In this section, we review the findings and their implications for the conceptual model, in addition to the results of the sentiment analysis of students’ feedback.

First, and in line with existing research in learning contexts (e.g., (Huang, 2021; Joo et al., 2018; Wu & Chen, 2017)), our results confirm that perceived ease of use positively influences perceived usefulness (H1). This implies that students find blended learning usable with valuable applicability in times of crisis. Consequently, perceived ease of use can be used to determine whether blended learning is designed to be accessible and engaging for students. A user-friendly and convenient implementation of blended learning would increase its practical worth for students’ learning. In addition, it is vital to consider the varied technical capabilities and age groups of students when designing learner-friendly features so that students do not encounter technical issues while using blended learning for the first time and recognize its usefulness for their learning goals.

Second, students’ satisfaction with blended learning was positively influenced by perceived ease of use and perceived usefulness (H2 and H3) which is coherent with existing literature such as (Arbaugh, 2000; Joo et al., 2018; Lin & Chen, 2012). Moreover, perceived ease of use also affects usefulness, which in turn affects students’ satisfaction. This is consistent with the results of Lin and Chen (2012). However, the impact of perceived usefulness was far greater than that of perceived ease of use, inferring that students’ satisfaction with blended learning is more influenced by the usefulness of blended learning features and its functionality.

The usability of blended learning, on the other hand, still has a significant influence on students’ satisfaction. The ease of use of blended learning is reflected in its flexibility and use of technology. Students are allowed to either attend classes in person or online from any location at any time. To earn academic credit, they can also submit their homework and take-home exams via an online platform or email. Additionally, learning presentations can be reviewed and completed. Some researchers claim that the flexibility promised in blended learning can increase learning motivation and improve students’ access to various contents and resources (Jun & Ling, 2011; Medina, 2018). Therefore, instructional content for blended learning should be designed with usefulness and ease of use in mind.

Our hypothesis that students’ satisfaction with blended learning has a positive impact on their future preferences (H4) was also supported by the data, which is in harmony with existing research (Huang, 2021; Joo et al., 2018; Lei & So, 2021; Wang et al., 2021). This means that students who are satisfied with blended learning will continue to choose this learning design. Furthermore, our findings echo that perceived ease of use and perceived usefulness have a positive mediation effect on future preferences. This is also consistent with previous research in an online learning environment (Huang, 2021; Joo et al., 2018).

Although students are keen to keep up with global learning tendencies such as blended learning, the obstacles they face limit the potential benefits of blended learning for Algerian higher education. The most frequently cited challenge was Internet access and its quality. Like in many developing countries, the Internet is not universally accessible in Algeria. Moreover, providing good quality Internet can be a costly burden for students. Without the proper technology, there can be no blended learning. This can diminish students’ perceived ease of use and consequently affect their overall acceptance of blended learning as a long-term strategy.

Similarly, students complain about the inadequacies of implementing blended learning, which have led to confusion and difficulties to adapt to the new learning environment. These inadequacies manifest themselves in a lack of engaging activities, the use of traditional teaching methods, the presentation of a large amount of content in a very short time, and the elimination of face-to-face courses for tutorials and exercises, especially in areas that rely heavily on practice such as computer science, and the lack of social presence and interaction. Students in different contexts and countries have identified comparable barriers (Abu Talib et al., 2021; Berbar & Ait Hamouda, 2019; Radia, 2019) and have been found to negatively affect students’ perceived ease of use and perceived usefulness, which in turn reduces students’ satisfaction with blended learning.

Theoretical Implications

This study has important implications for future research in the context of blended learning in developing countries such as Algeria. Firstly, the findings provide support for the theoretical model proposed by (Davis, 1989) and advances our understanding of the relationship between of perceived ease of use and perceived usefulness in shaping students’ satisfaction with blended learning and choice to continue using it in the future. Results show that students who are satisfied with blended learning are more likely to choose this learning strategy in the future. Consequently, the significance of establishing a positive blended learning experience cannot be overemphasized.

Moreover, the study shed light on difficulties faced by students in adapting to blended learning and stresses the need of resolving these issues. These difficulties can have a detrimental influence on students’ perceived ease of use and usefulness, and as a result, decreasing their engagement. Furthermore, the results highlight the need of emphasizing the utility of blended learning features and functionality in order to boost students’ satisfaction.

Ultimately, it can be inferred that theoretical implications of this analysis reveal the current state of blended learning in Algeria, and cast light on the necessity to encourage consistent communication and interaction between and among all involved parties (students, teachers, and institutional leadership) to grasp the challenges and requirements that can enhance the implementation of blended learning in a developing country.

Practical Implications and Recommendations

Overall, the results of this study have practical implications for designing blended learning that is both user-friendly and useful for students, considering their needs and technical capabilities. Furthermore, the findings are relevant to educators, policy makers, and educational institutions, as they provide insights into important factors to consider while designing and implementing blended learning effectively, thereby enhancing students’ acceptance and satisfaction with blended learning.

Based on the findings and theoretical implications of the study, we suggest the following recommendations for the effective implementation of blended learning in Algeria, which can also be considered for similar developing countries.

First and foremost, policymakers should address the Internet accessibility and quality. Given that this issue is a significant barrier to the effective implementation of blended learning, it is recommended that the educational institutions take the necessary steps to deliver quality and affordable Internet access to students and teachers as well. This can be achieved by increasing the government support and investment in internet infrastructure including building more cell towers, expanding fiber optic networks, and upgrading existing infrastructure to support higher bandwidth. However, a major limitation in Algerian telecommunication industry is lack of competition. There is only one state-owned telecommunication company (Algérie Télécom) that provides fixed-line telephone, mobile and internet services, and only two privately owned mobile phone operators and 4G internet, namely Djezzy and Ooredoo. Encouraging competition in this industry can help drive innovation while improving services and lowering costs. One way to increase competition is to encourage public-private partnerships nationally and internationally to leverage resources and expertise.

Secondly, we recommend that higher education institution offer the necessary technology and technical support and skills for all involved parties to ensure that teachers are capable of adapting their courses to the used technology while students are able to navigate the technology used in blended learning. With this aim in mind, institutions must afford access to needed resources in classrooms so students can participate in blended learning activities if they can’t afford high-speed internet, computers, or software. Training sessions and workshops should also be organized for this goal. Those sessions should focus primarily on explaining the blended learning strategy and used software. However, we should keep in mind that students and instructors in a scientific or engineering field may find it easier to use technology than those in more literary domains. Thus, institutions can provide resources such as tutorial, videos, and documentations, to help instructors and students to better understand how to use blended learning tools. Additionally, support system should be created, which may include technical staff as well as pedagogical staff who can assist students with any technical and learning-related issues that may arise.

Thirdly, students’ willingness to continue using blended learning is largely dependent on their satisfaction with their previous experience. This satisfaction is significantly influenced by the ease of use and usefulness; thus, it is important to refine usability features of the implemented blended learning strategy by implementing learner-friendly features to students of different age groups and technical capabilities. A refined and user-friendly blended learning strategy can boost student engagement and motivation, ultimately leading to better learning outcomes. This is because students are more likely to be engaged and motivated when they can easily navigate and interact with the learning process. Hence, improving the usability features of the blended learning strategy can lead to a more effective and efficient learning environment for Algerian higher education students and teachers.

One of the most essential learner-friendly features is an intuitive interface. A user-friendly interface makes it easier to navigate through different courses and access various resources. It should also be visually appealing. Furthermore, learner-friendly features must include simplified instructions and interactive learning tools. Simplified instructions are essential for enhancing the overall usability of a blended learning strategy as they can make it easier for students to understand the content and complete necessary tasks, whereas complex instructions can be a significant barrier to learning, especially for new online learners. Interactive learning tools, such as games and simulations, can make learning engaging and reinforce key concepts, encouraging critical thinking and practical application. This, however, would require a significant effort from instructors to adapt their lessons and activities to different formats. For instance, the presentation of a large amount of content in a very short time should be avoided as it may lead to confusion and difficulties in adapting to the new learning environment. Instead, lessons content should be divided into smaller parts, over a period of time, to ensure students will have enough time to take in the content. Furthermore, teachers should abandon traditional lectures, and not simply mix online and offline courses. Instead, they should encourage teamwork and provide engaging activities.

To prevent students from feeling isolated and separated from their peers and instructors, blended learning should include activities that encourage social presence and connection. Examples of activities that promote social presence and interaction in blended learning include; incorporating online discussion forums or chat rooms, group projects, peer-to-peer review activities, and virtual office hours or live Q&A sessions with instructors.

Finally, we believe that regular evaluation of the effectiveness of blended learning is crucial in ensuring that it meets its intended objectives and facilitates the identification of areas that require improvement. These assessments should not focus only on students, but also include teachers’ feedback and suggestions to improve their working environment.

Conclusion

Aside from the negative impact of COVID-19 on higher education, the overwhelming effect of technological advances remains an important issue that offers a good chance to redefine higher education for underprivileged communities. This study delivers empirical evidence regarding what factors are likely to influence students’ satisfaction and future preferences regarding blended learning. Our study targeted students from a developing country that is still in the early stages of incorporating ICTs in higher education institutions.

To analyze our quantitative data and validate our conceptual model, we applied the structural equation modeling approach. The main findings are consistent with the existing literature, which found that the adoption of blended learning is influenced by students’ perceived ease and usefulness, and their satisfaction with the experience. Their satisfaction is also impacted by these factors, particularly, by their perception of how useful blended learning is in achieving their learning objectives.

Qualitative data were analyzed using unsupervised sentiment analysis. The results reflect students’ initial enthusiasm for blended learning, albeit low satisfaction with the way it is currently delivered. There are several common problems students face with blended learning, such as difficulty accessing the Internet, lack of social presence, and inadequacies of online material due to ' lack of teachers’ experience or adoption efforts. These concerns are compounded by a lack of research on blended learning in developing countries, which may lead policymakers and institutional leadership to endorse such an approach before their needs have been fully considered.

We caution that if blended learning continues to be perceived as a difficult learning environment, it is unlikely that its implementation can sustain the intent of providing free and equal access to learning for all eligible students, which is the primary perspective of Algerian higher education. The diversity of challenges currently faced by students highlights the need to design a context-sensitive blended learning approach. Based on our analysis of qualitative and quantitative data, we believe that future implementation of blended learning should support the fundamental objective of expanding educational opportunities for all by offering equality and enabling easy access to essential technologies.

Limitations and Future work

This paper reflects the current state of blended learning adoption and helps plan and develop future curricula. However, our study has three major limitations.

First, we focused in this study only on students’ perceptions regarding the blended learning experience using subjective measures. Future research should balance subjectivity and objectivity in the data. Also, other factors such as software used, the effects of technostress, and the topics of the courses offered should also be explored.

Second, this work was conducted in a short time frame. Students’ subjective perceptions may change over time. Therefore, a longitudinal research design may reveal further insight into best practices that can accentuate blended learning benefits for developing countries and communities with limited resources.

Third, the scope of this study was limited to investigating students’ satisfaction and acceptance of blended learning. Future studies should also address Teachers’ perceptions to grasp the differences between teachers’ and students’ needs and challenges.

Biographies

Meriem Laifa, Ph.D.

received her Bachelor’s degree in Information Technology and Communication from Sétif University in 2010. In 2012, she received her Master’s degree in Business Intelligence from Bordj Bou Arreridj University. Now, she holds a Ph.D. in Computer Science and is working as a professor and researcher at the Computer Science Department of Bordj Bou Arreridj University. She is interested in Social Network Analysis, Natural Language Processing, online education, and social movements.

Roya Imani Giglou, Ph.D.

is a Non-Resident Research Fellow and Associate Editor at the Brussels- based think tank, the European Center for Populism Studies. She has previously worked as a postdoctoral researcher on a project focused on social media use and social integration of refugees in Belgium. Her research areas include quantitative and computational methods, political communication, social use of technology, and migration studies. Roya has published numerous articles in prestigious journals, including New Media & Society, Telematics and Informatics, and The Information Society.

Pr. Samir Akhrouf

is currently a full Professor at the Computer Science Department, Mohamed Boudiaf University of M’sila, Algeria. He earned his Engineer Degree from Constantine University, Algeria in 1984; an M.S in Computer Science, from University of Minnesota, USA in 1988; and a Ph.D from University of Setif Algeria. He was the Dean of the Faculty of Mathematics and Computer Science for 6 years. He also has been the vice rector in charge of Development, Prospective and Orientation for 3 years at Bordj Bou Arreridj University, Algeria. He has been teaching for over 36 years and his main research interests are focused on Biometric Identification, Computer Vision, Computer Networks, Distance Learning and Social Network Analysis.

Author Contributions

Authors contributed to this work as follows: Laifa Meriem and Imani G. Roya conducted the conceptualization, methodology, and analysis. Data collection was conducted by Laifa Meriem and Akhrouf Samir. The first draft of the manuscript was prepared and written by Laifa Meriem, then reviewed by Imani G. Roya, then edited again by Laifa Meriem. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The authors did not receive support from any organization for the submitted work.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available due to privacy concerns as some revealing data was explicitly shared in students’ feedback.

Declarations

Competing Interest

The authors have no relevant financial or non-financial conflicting interests to disclose.