Student adoption of e-learning at makerere university during COVID-19: A structural equation modelling based on the unified theory of acceptance and use of technology

ELearning in resource-constrained settings

Electronic learning in developing countries has received increased attention in recent years as a way to increase access to education and improve the quality of education in these countries (Arinto, 2017). Several studies have shown that eLearning can increase access to education for disadvantaged populations, such as those in remote or rural areas, those with disabilities, and those with limited financial resources (Zongozzi, 2022). In addition, eLearning can also provide flexible and personalized learning opportunities and support the development of 21st-century skills, such as critical thinking, problem-solving, and digital literacy (Arinto, 2017). However, the adoption and implementation of eLearning in developing countries is not without challenges. These challenges include limited ICT infrastructure, including access to technology and Internet connectivity, limited digital literacy among teachers and learners, and limited resources for the development and implementation (Chen et al., 2013; Martin and Ravanonoarivony, 2018). Therefore, there is a need for a well-coordinated and sustained effort from governments, education institutions, and other stakeholders to address these challenges and ensure that eLearning can contribute to the development and improvement of education systems in developing countries. This may involve the development of policies and programs to support the development and implementation of eLearning, investment in infrastructure and digital literacy, and ongoing evaluation and improvement of eLearning programs (Arinto, 2017; Martin and Ravanonoarivony, 2018).

This study holds significant importance for the institution, as it highlights the continuation of the status quo in the utilization of MUELE and Zoom. Subsequently, there has been a notable influence from students, instructors, and management advocating for blended learning. Nonetheless, comprehending the specific e-learning challenges within the context of Uganda’s institutions and Makerere University is crucial. According to a study conducted by Wanga and Ngumbuke (2012), Ugandan educational institutions encountered numerous challenges in embracing e-learning. These challenges included insufficient and inadequate infrastructural penetration, such as limited telephone connectivity, electricity, and internet access. Additionally, there was a lack of skills and e-competence to effectively utilize the technologies, coupled with a negative attitude towards technology. They also observed that the expense associated with computers and internet services remains elevated, and this is compounded by the prevalence of counterfeit products. These imitations not only come with a high price tag but also lack durability. Environmental factors, including prolonged dry seasons leading to overheating and dust accumulation, further contribute to a shortened lifespan of the equipment, worsening the challenges associated with e-learning. Moreover, findings from the research conducted by Mugenyi (2021), highlighted that Makerere University continued to grapple with various e-learning challenges affecting both students and instructors. These challenges encompassed obsolete and insufficient computers, unreliable internet access and speed, the absence of network cables as an alternative to wireless connections, and restrictions on accessing YouTube. Furthermore, students at Makerere University lacked familiarity with MUELE to the extent that they were unaware of how to access learning materials (Kharobo, 2022).

Amidst the challenges presented by COVID-19, e-learning obstacles became prominent, particularly as MUELE stood as the sole option and platform for engaging students, particularly in content uploads and assessments. For instance, some instructors and students, particularly from humanities colleges, faced digital illiteracy issues concerning MUELE. The university’s internet bandwidth was notably low, resulting in slow MUELE performance, especially during peak hours, and occasional outages. Access became a challenge as users were required to subscribe to daily internet, which proved to be expensive, especially during Zoom classes. As it was the initial experience with Zoom, conducting classes posed digital challenges since both students and instructors were unfamiliar with the technology. The procurement of Zoom licenses proved expensive for the university, especially considering the timing of the pandemic. However, with the assistance of NITA-U, Makerere successfully obtained some Zoom licenses. Each license was intended for installation on one machine and shared among two instructors.

Despite investments in technology, the adoption of e-learning remains notably low in developing countries, as noted by Farid et al. (2015). Therefore, this study aimed to evaluate the factors influencing students’ adoption of e-learning at Makerere University during the COVID-19 pandemic. The research utilized the Unified Theory of Acceptance and Use of Technology (UTAUT), which incorporates elements from eight other theoretical models related to technology acceptance. The subsequent sections of the paper are structured as follows: Theoretical perspective, existing studies, and hypotheses. The research design for the study, details of the data analysis, as well as the results of the measurement and structural model have been presented. Lastly, we offer a discussion of the primary findings of the study, concluding the article with implications and limitations.

Related work and conceptual framework

The study’s conceptual framework is based on The Unified Theory of Acceptance and Use of Technology (UTAUT). UTAUT is a widely used theoretical framework in research to understand the factors that influence the acceptance and use of technology (Venkatesh et al., 2003). In the context of eLearning research, the UTAUT model can be used to identify the key factors that influence the adoption and usage of technology-based learning systems. By applying the UTAUT model to eLearning research, researchers can gain insight into the attitudes and behaviours of students, teachers, and other stakeholders toward the adoption and use of technology-based learning systems. This information can inform the design and implementation of technology-based learning systems to ensure that they are more effective and user-friendly.

The conceptual framework of the study consists of six constructs and three moderating variables. These include Performance Expectancy (PE), Effort Expectancy (EE), Social Influence (SI), Facilitating Conditions (FC), Behavioural Intention (BI), and Usage Behaviour (UB). The moderating variables encompass age, gender, and experience. The inclusion of Performance Expectancy (PE) in the conceptual framework of the study is grounded in its pivotal role as a core element in diverse technology adoption models. Notably, it features prominently in models like the Technology Acceptance Model (TAM) and its extensions, such as the Unified Theory of Acceptance and Use of Technology (UTAUT). Venkatesh et al. (2012) emphasized the significance of performance expectancy as a predictor of users’ intentions to adopt new technologies. Furthermore, Bellaaj et al. (2015) discovered that PE significantly influenced individuals’ inclination to use eLearning at the University of Tabuk in Saudi Arabia. Likewise, PE emerged as a potent determinant of the intention to embrace eLearning in public universities in Kenya. Recognizing the significance of PE in determining students’ intention to use eLearning platforms in public universities under voluntary circumstances, it became imperative to investigate its role in the adoption of eLearning during mandatory conditions like those imposed by COVID-19. In such situations, where the continuity of learning was essential amidst lockdowns and social distancing policies, understanding the influence of PE became crucial.

The study incorporated effort expectancy (EE) from various information system models, including perceived ease of use from the Technology Acceptance Model (TAM/TAM2), complexity from the Model of PC Utilization (MPCU), and ease of use from the Innovation Diffusion Theory (IDT) (Kijsanayotin et al., 2009). Additionally, Bellaaj et al. (2015) discovered that Effort Expectancy (EE) played a substantial role in determining an individual’s intention to utilize eLearning at the University of Tabuk in Saudi Arabia. Additionally, EE was recognized as a significant predictor of behavioral intention in the adoption of eLearning within private universities in Indonesia (Indrati et al., 2014). Given its effectiveness in forecasting the behavioral intention to use eLearning technologies, it was essential to incorporate EE in the study investigating factors influencing the adoption of eLearning tools at Makerere University during the imperative period of COVID-19, where targeted interventions to bolster eLearning were vital.

Social influence is the sway that individuals exert on each other concerning their proficiency in utilizing a new system (Venkatesh et al., 2003). It’s crucial to recognize that this influence can take on either a positive or negative form, presenting a pivotal consideration in the daily lives of individuals as it holds the potential to impact future opportunities (Venkatesh et al., 2003). The impact of family, friends, and colleagues holds significant sway in shaping an individual’s choice to embrace electronic learning (Irani et al., 2009; Tan and Teo, 2000). Subjective norms are integrated into models such as the Theory of Reasoned Action (TRA), Theory of Planned Behavior (TPB), and TAM2, while social influence is included in models like MPCU and IDT, highlighting their role in shaping individual behaviour within society. Numerous studies have explored the relationship between SI and BI. For example, Kocaleva et al. (2015a) discovered a robust impact of SI on the intention to use eLearning among teaching staff in Macedonia. Similarly, in Higher Education Institutions in the United Arab Emirates and Nigeria, Alblooshi and Abdul Hamid (2021) and Daud Mahande and Makassar (2019) respectively found that SI strongly influenced behavioral intentions to adopt eLearning technologies. Considering the noteworthy findings in analogous studies, it was deemed appropriate to incorporate this factor in the current study. Given, the significance of its results in similar studies, it was deemed fit to include in this study.

Facilitating conditions are frequently integrated into various technology adoption models, playing a crucial role as a determinant in the acceptance and utilization of technology. According to Venkatesh et al. (2003), facilitating conditions refer to the extent to which individuals perceive that organizational and technical infrastructure are in place to support the use of the technology. In the UTAUT model, facilitating conditions directly influences usage behavior. Previous studies and literature consistently demonstrate that facilitating conditions have a positive impact on the intention to use technology (Chang et al., 2007; Moore and Benbasat, 1991). The results of the study indicate that facilitating conditions act as a favourable predictor of technology usage. Considering the consistent performance of FC in prior studies, it was considered essential to include it in the current research. Consequently, the study formulated and examined the conceptual framework depicted in Figure 2, incorporating six variables.OPEN IN VIEWER

Furthermore, the study examined how age, gender, and experience moderate the relationship between FC and UB, EE, PE, and SI on BI. The examination of the moderating variables was deemed necessary, considering their impact observed in previous studies. For example, Venkatesh et al. (2003) discovered the substantial moderating effects of age, gender, and experience in their research. Similarly, Acheampomg and Boateng (2018), identified a statistically significant moderating effect of age on the relationship between FC, SI, EE, and PE on BI.

Utilizing the conceptual framework illustrated in Figure 2, the research examined 16 null hypotheses derived from the nine constructs. Notably, the study did not assess the moderating variables for voluntariness, as the utilization of MUELE and Zoom for learning during the COVID-19 pandemic was obligatory, with no alternative mode of learning available.

Study hypotheses

The study was conducted at Makerere University, Kampala, Uganda. The following alternate hypotheses (H_a1-H_a16) were formulated and tested.

Research design

This study employed a descriptive research design, a common choice for investigations into factors influencing technology adoption due to its effectiveness in providing a comprehensive understanding of the phenomenon. The primary aim of this study was to identify the factors influencing the adoption and subsequent use of eLearning technologies among undergraduate students at Makerere University. According to Davis (1989) and Rogers (1995) technology adoption is a multifaceted process influenced by diverse factors, and a descriptive research design allows for an in-depth examination of these factors within their natural settings. Building on Davis and Rogers’ arguments, we utilized a descriptive research design to conduct a thorough analysis of the factors affecting the use of MUELE and ZOOM in the context of Makerere University. Additionally, descriptive research is valuable for establishing a baseline understanding of the current state of technology adoption, facilitating subsequent research and interventions (Aquino et al., 2018; Johnson et al., 2017), and offering a holistic examination of the adoption process. This approach contributes to a nuanced understanding that informs both academia and practical applications, including policy decisions. Therefore, the selection of a descriptive research design in this study aligns with the need for a detailed, contextualized, and holistic exploration of the factors influencing the adoption of MUELE and ZOOM at Makerere University during the COVID-19 pandemic. The survey technique was employed in this study, allowing for data collection from a large population while maintaining the anonymity of the respondents (Neill et al., 2021).

Makerere University is one of the oldest institutions of higher learning in Africa and the largest in Uganda. Established in 1922, as a small technical school, today, the institution accounts for over 80% and 95% of the annual graduate and university related research output respectively Kasozi (2016). The institution was eventually called Uganda Technical College commencing in January with 14 day students offering carpentry, building and mechanics. Presently, the institution has a total of nine colleges and one school with a total student population of 27 thousand five hundred as of January 2024. It is located in Kampala, the capital of Uganda sitting on 300 acres. The choice of conducting this study at Makerere University was based on the fact that it is the largest and one of the first institutions in Uganda to introduce e-Learning under the platform called blackboard. Conversely, it was convenient for the researchers to collect data since they work at the institution. Also, few studies have examined the eLearning adoption factors at Makerere University hence contributing to this thinly researched area would be highly impactful.

Study participants and settings

Undergraduate students from the College of Computing and Information Sciences (COCIS) and the College of Natural Sciences (CONAS) were enlisted in the research. COCIS consists of two schools: the School of Computing and Informatics Technology and the East African School of Library. CONAS is composed of two schools: Physical Sciences and Biosciences. Students from these colleges were chosen based on their previous experience and training in utilizing MUELE for both learning and assessment purposes.

Students who lived on or near campus had access to the university’s unlimited Wi-Fi Internet connectivity, while others used private internet connections. These students accessed learning materials such as recorded lessons and lecture notes through MUELE and YouTube, with live classroom sessions conducted via Zoom. Consequently, all course modules were exclusively available through online platforms. Moreover, students, particularly first-years, underwent a series of usability training sessions aimed at enhancing their proficiency with MUELE. These sessions encompassed tasks such as opening accounts, enrolling in courses, uploading course assessments, downloading grade books and study materials, and setting quizzes.

Sampling procedure

A multistage sampling technique was adopted to select study participants. Multistage sampling was deemed appropriate, as the students in the seven colleges in Makerere University were enrolled in various courses with different study backgrounds and subjected to different study conditions, which makes it possible to divide the student population into distinct groups. The selection of the colleges to recruit students for the study was guided by an evaluation report that suggested that College of Computing and Information sciences (COCIS) and College of Natural Sciences (CONAS) had fully uploaded teaching materials on the digital platform (Mugenyi, 2021). First, we purposively recruited students from COCIS because it had the highest number of courses with uploaded learning content, suggesting a greater use of MUELE. According to Mugenyi (2021), CONAS is one of the colleges that least used MUELE therefore, the students from CONAS were recruited in the study to enable a comparative analysis of factors that influenced the adoption of MUELE in a college that least used MUELE and the counterpart with higher utilization levels. Finally, students were randomly recruited based on the course enrolment list, henceforth contacted through an instant messaging feature of MUELE while others were telephoned. Questionnaires were administered either at the end of the lectures or after writing exams. Student leaders assisted in administering questionnaires.

Sample size

The sample size was determined using Kish Leslie’s formula (Kish, 1965) for estimating sample sizes in cross-sectional studies (Formula (1)). Kish Leslie’s formula was chosen because Makerere University boasts an extensive and diverse population, making it impractical to collect data from the entire population. Makerere University has a population of 31,000 undergraduates and 4,000 postgraduates, offering 136 undergraduate programs and 179 graduate programs (CIVIS, 2024). The adoption of Kish Leslie’s formula is particularly pertinent in this context as it enables the selection of a representative sample that can be extrapolated to the broader university community, ensuring statistical reliability (Formula (1)).n = z α 2 2 * ρ 1 − ρ ÷ α 2(1)Where z is the z-value corresponding to marginal error (α = 0.05), and ρ is the estimated response rate of the specific characteristic under the study. In this study, the response rates to specific covariates were not determined, but the assumed response rate that yields the largest sample size (ρ = 0.5) was used. Assuming a power of 90%, and a two-tailed alpha of 0.05, the required sample size was determined as 384. Also, in accordance with the guidelines outlined by Hair et al. (2011), a recommended sample size of 15–20 observations per variable is suggested for ensuring generalizability. Considering that this study encompassed a total of nine variables, inclusive of the actual use of the e-learning system, age, gender and experience, the sample size was deemed adequate.

Data collection instrument

Likert-scale questionnaire was developed based on the Unified Theory of Acceptance and Use of Technology (UTAUT) and tailored to our study context as shown in Table 1. It was self-administered to collect data on the relationships between the six latent variables in the UTAUT model. The questionnaire comprised three parts. The first part consisted of eight demographic questions to gather information on the participants’ demographic characteristics. The second part comprised 24 questions measuring five of the six variables in the research model. Each measurement element was rated on a five-point Likert scale, ranging from 1 (‘strongly disagree’) to 5 (‘strongly agree’). Additionally, the option ‘not sure’ was included to provide flexibility in completing the questionnaire. The third part included four open-ended questions designed to collect qualitative data on the challenges encountered during eLearning and proposed solutions. This structure facilitated a comprehensive data collection approach, combining quantitative measures of UTAUT constructs with qualitative insights into challenges and solutions related to eLearning.

OPEN IN VIEWER

Table 1.

Constructs, their items and sources.

Construct	Descriptions	Source
Performance expectancy (PE)	Using MUELE during COVID-19 enhanced my motivation to learn (PE1)	(Venkatesh et al., 2003)
	Using ZOOM during enhanced my motivation to learn during COVID-19 (PE2)	(Fleury et al., 2017)
	Using MUELE increased my performance in my learning activities (PE3)	(Mongkolnam et al., 2013)
	Using ZOOM increased my performance in my learning activities (PE4)	(Venkatesh et al., 2003)
	Using MUELE improved my learning results (PE5)
	Using ZOOM improved my learning results (PE6)
	I found MUELE useful in my university study (PE7)
	I found ZOOM useful in my university study (PE8)
	Using MUELE saved time with regard to submission of coursework (PE9)
	Using ZOOM saved time with regard to submission of coursework (PE10)
Effort Expectancy (EE)	I found it easy to become skillful at using MUELE (EE1)	(Venkatesh et al., 2003)
	I found it easy to become skillful at using ZOOM (EE2)	(Fleury et al., 2017)
	I found MUELE easy to use (EE3)	(Venkatesh et al., 2003)
	I found ZOOM easy to use (EE4)
	My learning activities with ZOOM were clear and understandable (EE5)
	My learning activities with MUELE were clear and understandable (EE6)
	I became proficient at using ZOOM (EE7)
	I became proficient at using MUELE (EE8)
	I did not take long to learn how to use MUELE (EE9)
	I did not take long to learn how to use ZOOM (EE10)
Social influence (SI)	People who affected my learning behavior thought I should use MUELE (SI1).	(Venkatesh et al., 2003)
	People who affected my learning behavior thought I should use ZOOM (SI2).	(Fleury et al., 2017)
	People who were important to me thought I should use MUELE (SI3).	(Venkatesh et al., 2012)
	People who were important to me thought I should use ZOOM (SI4).	(Venkatesh et al., 2003)
	My peers thought I should use MUELE (SI5).	(Fleury et al., 2017)
	My peers thought I should use ZOOM (SI6).	(Mongkolnam et al., 2013)
	My teachers thought I should use MUELE (SI7).	(Venkatesh et al., 2012)
	My teachers thought I should use ZOOM (SI8).
Facilitating conditions (FC)	I had the knowledge necessary to use MUELE (FC1)	(Venkatesh et al., 2003)
	I had the knowledge necessary to use ZOOM (FC2)	(Fleury et al., 2017)
	I had the resources necessary to use MUELE (FC3)	(Mongkolnam et al., 2013)
	I had the resources necessary to use ZOOM (FC4)
	I thought using MUELE fitted well with the way I want to learn (FC5)
	I thought using ZOOM fitted well with the way I want to learn (FC6)
	I quickly solved problems related to using MUELE (FC7)
	I quickly solved problems related to using ZOOM (FC8)
Behavioral intention (BI)	I would use MUELE to improve my performance (BI1).	(Venkatesh et al., 2003)
	I would use ZOOM to improve my performance (BI2).	(Fleury et al., 2017)
	I intend to use MUELE in my future learning activities (BI3).	(Mongkolnam et al., 2013)
	I intend to use ZOOM in my future learning activities (BI4).
	I plan to use MUELE in my future learning activities (BI5).
	I plan to use ZOOM in my future learning activities (BI6).
Usage behavior (UB)	I was satisfied with using MUELE for learning purposes (UB1)	(Venkatesh et al., 2012)
	I was satisfied with using ZOOM for learning purposes (UB2)	(Bhattacherjee, 2001)
	I had a good experience using MUELE for learning purposes (UB3)	(Venkatesh et al., 2003)
	I had a good experience using ZOOM for learning purposes (UB4)
	I was satisfied with the effects of using MUELE for learning purposes (UB5)
	I was satisfied with the effects of using ZOOM for learning purposes (UB6)

Demographic characteristics of the respondents

The demographic characteristics of the participants in Table 2 reflects a fairly balanced gender distribution, with 194/374 (52%) male and 180/371 (48%) female respondents. Across academic years, Year 3 has the highest representation at 187/374 (50%), followed by Year 1 at 99/374 (27%), Year 2 at 86/374 (23%), and a minimal presence of Year 4 students at 2/374 (1%). Regarding age, the majority fall within the 18-22 years range, constituting 227/374 (61%) of the respondents, followed by 23-27 years at 123/374 (33%), while older age groups have lesser representation. In terms of programs, Bachelor of Information Systems Technology is the most popular at 229/374 (61%) followed by Bachelor of Science in Computer Science at 50/374 (13%), Bachelor of Science in Software Engineering at 54/374 (14%), and Bachelor of Science, Flat at 41/374 (12%). Most participants are affiliated with the College of Computing and Information Sciences (COCIS) at 333/374 (89%), with the remaining 41/374 (11%) from the College of Natural Sciences (CONAS). The majority of participants are undergraduates at 339/374 (91%), with 35/374 (9%) being postgraduates. During COVID-19, a significant portion of students utilized MUELE 328/374 (88%) and Zoom 339/374 (91%) for their educational activities.

OPEN IN VIEWER

Table 2.

Demographic information of the participants.

Number that participated and returned questionnaires	Number/Percent of students N = 374
	Gender
	Male	Female	Total
Number of respondents	194 (52%)	180 (48%)	374 (100%)
Year of study
Year 1	58 (30%)	41 (13%)	99 (27%)
Year 2	53 (27%)	33 (18%)	86 (23%)
Year 3	81 (42%)	106 (59%)	187 (50%)
Year 4	2 (1%)	0 (0%)	2 (1%)
Age
18–22 years	99 (51%)	128 (71%)	227 (61%)
23–27 years	78 (40%)	45 (25%)	123 (33%)
28–32 years	8 (4%)	6 (3%)	14 (4%)
33- 37 years	8 (4%)	0 (0%)	8 (2%)
38-42 years	1 (1%)	1 (1%)	2 (1%)
Programs
Bachelor of information systems technology	140 (72%)	89 (49%)	229 (61%)
Bachelor of science in computer science	15 (8%)	35 (19%)	50 (13%)
Bachelor of science in software engineering	12 (6%)	42 (23%)	54 (14%)
Bachelor of science, flat	27 (14%)	14 (9%)	41 (12%)
Colleges
COCIS	172 (89%)	161 (89%)	333 (89%)
CONAS	22 (11%)	19 (11%)	41 (11%)
Type of program
Day	130 (67%)	139 (77%)	269 (72%)
Evening	64 (33%)	41 (23%)	105 (28%)
Type of student
Undergraduate	179 (92%)	160 (89%)	339 (91%)
Postgraduate	15 (8%)	20 (11%)	35 (9%)
Student used MUELE during COVID-19
No	25 (13%)	21 (12%)	46 (12%)
Yes	169 (87%)	159 (88%)	328 (88%)
Student used zoom during COVID-19
No	26 (13%)	9 (5%)	35 (9%)
Yes	168 (87%)	171 (95%)	339 (91%)

Across gender lines, there appears to be a notable difference in the distribution across years of study. While Year 3 dominates among female respondents with 106/374 (59%), Year 1 and Year 2 students comprise a larger proportion of male respondents at 58/374 (30%) and 53/374 (27%) respectively. This suggests potential variations in the progression rates or enrolment patterns between male and female students within the institution. Additionally, the distribution of age groups also displays a discrepancy between genders. A majority of female respondents fall within the 18-22 years’ bracket, constituting128/374 (71%) of their cohort, whereas male respondents are more evenly distributed across the age groups, with 99/374 (51%) falling within the 18-22 years range. Furthermore, the analysis of program enrolment reveals intriguing patterns when dissected by gender. Male students exhibit higher participation rates in programs such as Bachelor of Information Systems Technology and Bachelor of Science, flat, with 140 (72%) and 27 (14%) respectively, compared to their female counterparts. Conversely, female students demonstrate a stronger presence in Bachelor of Science in Software Engineering and Bachelor of Science computer science, comprising 35 (19%) and 42 (23%), respectively. These findings underscore the importance of considering gender dynamics in educational settings and highlight potential areas for targeted support or intervention to promote gender equity and diversity across academic disciplines.

Reliability and validity of the questionnaire

We tested the validity and reliability of the questionnaire before deployment (Taherdoost, 2016). In this study, reliability was tested using the Cronbach Alpha, while validity was tested using average variance extracted (AVE). The reliability results in Table 3, as indicated by Cronbach’s alpha and composite reliability values, demonstrate high internal consistency and reliability across all constructs in the study. Cronbach’s alpha values range from 0.872 to 0.929, indicating strong internal consistency (Tavakol and Dennick, 2011). Additionally, both composite reliability estimates (rho_a and rho_c) surpass the commonly accepted threshold of 0.70, with values ranging from 0.873 to 0.940 (Hair et al., 2019). These findings suggest that the measures used to assess behavioral intention, effort expectancy, facilitating conditions, performance expectancy, social influence, and usage behavioral are reliable and consistent, providing a solid foundation for further analyses and interpretations.

OPEN IN VIEWER

Table 3.

reliability results.

Constructs	Cronbach’s alpha	Composite reliability (rho_a)	Composite reliability (rho_c)
BI	0.908	0.909	0.929
EE	0.929	0.931	0.940
FC	0.891	0.898	0.913
PE	0.924	0.926	0.936
SI	0.907	0.909	0.925
UB	0.872	0.873	0.904

Convergent validity

The average variance extracted (AVE) values in Table 4 provide insights into the convergent validity of the constructs, indicating the extent to which the items within each construct measure the underlying concept. In this study, all constructs exhibit AVE values above the recommended threshold of 0.50, thus indicating satisfactory convergent validity (Hair et al., 2019). Specifically, Behavioural Intention has an AVE of 0.686, followed by Usage Behaviour with 0.610, Social Influence with 0.605, and Effort Expectancy with 0.609, indicating that these constructs explain more than 60% of the variance in their respective items. Performance Expectancy and Facilitating Conditions also demonstrate acceptable convergent validity, with AVE values of 0.595 and 0.571 respectively. These findings suggest that the constructs adequately represent their underlying concepts and provide reliable measures for assessing the targeted phenomena (Fornell and Larcker, 1981).

OPEN IN VIEWER

Table 4.

Convergent validity results.

Constructs	Average variance extracted (AVE)
Behavioral intention	0.686
Effort Expectancy	0.609
Facilitating conditions	0.571
Performance Expectancy	0.595
Social influence	0.605
Usage behavioral	0.610

Discriminant validity

Table 5 presents a correlation matrix that depicts the discriminant validity among the constructs in this study, illustrating the extent to which each construct is distinct from others. Discriminant validity is supported when the correlation between constructs is lower than the square root of the average variance extracted (AVE) of each construct (Fornell and Larcker, 1981). In this matrix, the diagonal values represent the square root of the AVE for each construct. Comparing these diagonal values with the correlations between constructs, it can be observed that the correlation values are generally lower than the square roots of the AVEs.

OPEN IN VIEWER

Table 5.

Discriminant validity results.

Constructs	AGE	Behavioral intention	Effort expectancy	Experience	Facilitating conditions	GENDER	Performance expectancy	Social influence	Usage behavioral
AGE	1.000
Behavioral intention	0.118	0.828
Effort Expectancy	0.089	0.687	0.780
EXPERIENCE	−0.006	0.258	0.323	1.000
Facilitating conditions	0.088	0.734	0.747	0.315	0.755
GENDER	−0.194	−0.062	−0.045	−0.003	−0.069	1.000
Performance Expectancy	0.167	0.692	0.747	0.414	0.677	−0.027	0.771
Social influence	0.147	0.706	0.658	0.333	0.662	−0.066	0.721	0.778
Usage behavioral	0.058	0.667	0.584	0.294	0.695	−0.092	0.649	0.609	0.781

Path coefficient

The path coefficients presented in Table 6 suggest various relationships among the variables. Age exhibits a minor positive influence on Behavioral Intention and a slightly stronger negative association with Usage Behavior. Both Behavioral Intention and Usage Behavior are positively affected by Effort Expectancy, with the former also positively linked to Performance Expectancy and Social Influence. However, Behavioral Intention is negatively influenced by gender, while its relationship with experience is negative. Gender interacts positively with Social Influence and Performance Expectancy but negatively with Effort Expectancy. Experience interacts positively with Facilitating Conditions, Social Influence, and Effort Expectancy but negatively with Performance Expectancy, with the interaction terms implying nuanced effects on Usage Behavioral and Behavioral Intention. Overall, these results depict a complex interplay among demographic factors, work experience, and organizational variables, indicating that certain factors like EE and gender exert notable influences on various aspects of the model.

OPEN IN VIEWER

Table 6.

Path coefficient.

Constructs	Path coefficients
AGE -> BI	0.023
AGE -> UB	−0.019
BI-> UB	0.334
EE -> BI	0.455
EXPERIENCE -> BI	−0.045
EXPERIENCE -> UB	0.068
FC -> UB	0.440
GENDER -> BI	−0.059
PE -> BI	0.187
SI -> BI	0.239
EXPERIENCE x FC -> UB	0.045
EXPERIENCE x SI -> BI	−0.085
EXPERIENCE x PE -> BI	0.035
EXPERIENCE x EE -> BI	0.053
AGE x FC -> UB	−0.001
AGE x SI -> BI	−0.019
AGE x PE -> BI	−0.040
AGE x EE -> BI	0.034
GENDER x SI -> BI	0.224
GENDER x PE -> BI	0.074
GENDER x EE -> BI	−0.289

Coefficient of determination (R2)

The R-square (coefficient of determination) and R-square adjusted in Table 7 are measures of the goodness-of-fit of a regression model. They indicate the proportion of variance in the dependent variable that is explained by the independent variables. In this context, for Behavioral Intention, the R-square value of 0.624 means that approximately 62.4% of the variance in Behavioral Intention can be explained by social Influence, Effort expectancy and performance expectancy. Similarly, the R-square adjusted value of 0.608 adjusts the R-square value to account for the number of predictors in the model, providing a more conservative estimate of the variance explained. For Usage Behavior, the R-square value of 0.544 indicates that approximately 54.4% of the variance in Usage Behavior can be explained by Behavioral Intention and Facilitating Conditions while the adjusted R-square value of 0.537 considers the number of predictors and provides a more accurate estimate of the variance explained. These results suggest that the independent variables considered in this study have a substantial explanatory power for both Behavioral Intention and Usage Behavior. However, there may still be unaccounted factors influencing these dependent variables, as evidenced by the portion of variance not explained by the model.

OPEN IN VIEWER

Table 7.

Coefficient of determination.

Constructs	R-square	R-square adjusted
Behavioral intention	0.624	0.608
Usage behavior	0.544	0.537

Indirect and total effect of the independent variables on dependent variables

In the context of adopting eLearning in the post-COVID-19 era, the total effects of Performance Expectancy (PE), Effort Expectancy (EE), and Social Influence (SI) on Behavioral Intention (BI) shed light on crucial determinants influencing individuals’ decisions to engage with eLearning platforms. As shown in Table 8, among these factors, Effort Expectancy (EE) emerges as the most influential determinant of BI, with a substantial total effect of 0.455. This indicates that individuals’ perceptions of the ease of use and simplicity associated with eLearning systems significantly influence their intention to adopt such platforms. Following EE, Social Influence (SI) demonstrates a noteworthy total effect of 0.239, suggesting that the influence of social factors, such as recommendations from peers or social networks, plays a crucial role in shaping individuals’ behavioral intentions towards eLearning adoption. Additionally, Performance Expectancy (PE) exhibits a positive total effect on BI, albeit with a smaller magnitude of 0.187, highlighting the importance of individuals’ beliefs about the usefulness and benefits of eLearning in driving their intention to adopt. Thus, in this model, EE emerges as the factor with the greatest effect on BI, followed by SI and PE, respectively, underscoring the significance of user experience and social influences in shaping intentions to engage with eLearning platforms in the post-pandemic era.

OPEN IN VIEWER

Table 8.

Total effects of constructs.

Constructs	Indirect effects	Construct	Total effects
AGE -> BI -> UB	0.008	AGE -> BI	0.023
EE -> BI -> UB	0.152	AGE -> UB	−0.011
EXPERIENCE -> B -> UB	−0.015	BI-> UB	0.334
GENDER -> B -> UB	−0.020	EE -> BI	0.455
PE -> BI -> UB	0.062	EE -> UB	0.152
SI -> BI -> UB	0.080	EXPERIENCE -> BI	−0.045
EXPERIENCE x SI -> B -> UB	−0.028	EXPERIENCE -> UB	0.053
EXPERIENCE x PE -> B -> UB	0.012	FC -> UB	0.440
EXPERIENCE x EE -> B -> UB	0.018	GENDER -> BI	−0.059
AGE x SI -> BI -> UB	−0.006	GENDER -> UB	−0.020
AGE x PE -> BI -> UB	−0.013	PE -> BI	0.187
AGE x EE -> BI -> UB	0.011	PE -> UB	0.062
GENDER x SI -> BI -> UB	0.075	SI -> BI	0.239
GENDER x PE -> BI -> UB	0.025	SI -> UB	0.080
GENDER x EE -> BI -> UB	−0.096	EXPERIENCE x FC -> UB	0.045
		EXPERIENCE x SI -> BI	−0.085
		EXPERIENCE x SI -> UB	−0.028
		EXPERIENCE x PE -> B	0.035
		EXPERIENCE x PE -> UB	0.012
		EXPERIENCE x EE -> B	0.053
		EXPERIENCE x EE -> UB	0.018
		AGE x FC -> UB	−0.001
		AGE x SI -> B	−0.019
		AGE x SI -> UB	−0.006
		AGE x PE -> B	−0.040
		AGE x PE -> UB	−0.013
		AGE x EE -> B	0.034
		AGE x EE -> UB	0.011
		GENDER x SI -> B	0.224
		GENDER x SI -> UB	0.075
		GENDER x PE -> B	0.074
		GENDER x PE -> UB	0.025
		GENDER x EE -> B	−0.289
		GENDER x EE -> UB	−0.096

Also, the total effects of Performance Expectancy (PE), Effort Expectancy (EE), Facilitating Conditions (FC), Social Influence (SI), and Behavioral Intention (BI) on Usage Behavior (UB) reveal significant insights. Among these factors, Facilitating Conditions (FC) emerges as the most influential determinant of UB, with a substantial total effect of 0.440. This indicates that the presence of conducive conditions such as technological infrastructure, support mechanisms, and accessibility plays a pivotal role in encouraging learners to engage in eLearning activities. Following FC, Effort Expectancy (EE) exhibits a noteworthy total effect of 0.152, suggesting that the perceived ease of use and convenience associated with eLearning platforms significantly impacts users’ behavior. Additionally, Performance Expectancy (PE) and Social Influence (SI) demonstrate positive total effects on UB, albeit with smaller magnitudes of 0.062 and 0.080, respectively. This underscores the importance of users’ perceptions of the usefulness of eLearning and the influence of social factors in shaping their adoption behavior. Furthermore, Behavioral Intention (BI) exhibits a substantial total effect of 0.334 on UB, highlighting its role as a key precursor to actual usage behavior. Thus, in this model, FC emerges as the factor with the greatest effect on UB, followed by BI, EE, SI, and PE, respectively, emphasizing the multifaceted nature of factors influencing eLearning adoption post-pandemic.

Hypothesis testing

The hypothesis testing results in Table 9 provide insightful implications for understanding the factors influencing the usage of eLearning systems. Firstly, the hypothesis regarding the influence of Behavioral Intention (BI) on Use Behavior (UB) yielded a significant T statistic (T = 5.820, p = .000), indicating strong support for the hypothesis. This suggests that users’ intentions play a pivotal role in shaping their actual usage behavior towards eLearning systems. Such a finding underscores the importance of fostering positive intentions among users through various interventions and strategies to ultimately drive usage.

OPEN IN VIEWER

Table 9.

T statistics and p values.

Hypothesis	T Statistics (|O/STDEV|)	p Values	Comment
Ha5: Behavioural intention positively influences usage behaviour of eLearning technologies.	5.820	0.000	Supported
Ha2: Effort Expectancy positively influences behavioural intention to use eLearning technologies	5.490	0.000	Supported
Ha4: Facilitating conditions positively influences usage behaviour of eLearning technologies.	7.205	0.000	Supported
Ha11: Gender moderates the relationship between effort expectancy and behavioral intention	2.887	0.004	Supported
Ha3: Social influence positively influences the intention to use eLearning technologies.	2.588	0.010	Supported
Ha1: Performance expectancy positively influences behavioural intention to use eLearning technologies.	2.068	0.039	Supported
Ha12: Gender moderates the relationship between social influence and behavioral intention	1.946	0.052	Not supported
Ha15: Experience moderates the relationship between social influence and behavioral intention.	1.447	0.148	Not supported
Ha16: Experience moderates the relationship between facilitating conditions and use behavior.	1.195	0.232	Not supported
Ha14: Experience moderates the relationship between effort expectancy and behavioral intention	0.936	0.349	Not supported
Ha7: Age moderates the relationship between effort expectancy and behavioral intention.	0.648	0.517	Not supported
Ha10: Gender moderates the relationship between performance expectancy and behavioral intention.	0.625	0.532	Not supported
Ha13: Experience moderates the relationship between performance expectancy and behavioral intention.	0.623	0.533	Not supported
Ha6: Age moderates the relationship between performance expectancy and behavioral intention.	0.588	0.557	Not supported
Ha8: Age moderates the relationship between social influence and behavioral intention.	0.296	0.767	Not supported
Ha9: Age moderates the relationship between facilitating conditions and use behaviour.	0.037	0.971	Not supported

Secondly, the significant T statistic for Effort Expectancy (EE) on Behavior Intention (BI) (T = 5.490, p = .000) highlights the crucial role of perceived ease of use in influencing users’ actual behavior. This implies that efforts aimed at simplifying the use of eLearning systems and enhancing user-friendliness can potentially lead to increased adoption and usage rates.

Additionally, the significant T statistic for Facilitating Conditions (FC) on UB (T = 7.205, p = .000) underscores the importance of providing necessary resources, support, and infrastructure to facilitate users’ access and utilization of eLearning systems. This suggests that addressing barriers and enhancing facilitating conditions can significantly enhance users’ actual usage behavior.

Conversely, several hypotheses did not yield significant results, such as the influence of demographic factors like Age and Gender on users’ behaviors, as indicated by non-significant T statistics and p values. This suggests that demographic variables might not significantly impact users’ behaviors in the context of eLearning systems, emphasizing the need for a more nuanced understanding of user characteristics and their interactions with other factors. Overall, these findings highlight the multifaceted nature of factors influencing eLearning system usage and underscore the importance of addressing key determinants such as behavioural intention, effort expectancy, and facilitating conditions to promote successful adoption and usage.

Influence of performance expectations on the intention to use eLearning technologies

In this study, a notably substantial impact of PE on the behavioural intention to utilize eLearning systems was identified, aligning with findings in various studies (Kayali and Alaaraj, 2020; Kocaleva et al., 2015b; Mahande, 2016; Qiao et al., 2021). As per the findings of this study, learners expressed the belief that the utilization of e-learning systems heightened their motivation to learn, enhanced learning outcomes, and positively impacted their performance in learning activities, although this was attributed to Zoom rather than MUELE. Consequently, e-learning proved beneficial in their university studies, with MUELE playing a pivotal role in time-saving for coursework submission. This aligns with a study conducted at Hashemite University in Jordan (Abbad, 2021), where learning management systems were identified as useful, facilitating swift completion of learning activities and improving students’ overall learning productivity.

Several studies have observed that the more learners view e-learning systems as platforms to enhance their performance in e-learning, the greater their inclination to adopt them (Alblooshi and Abdul Hamid, 2021; Bellaaj et al., 2015), consistent with the findings of this study. Hence, it is not surprising that students at Makerere University universally perceived e-learning systems as beneficial in enhancing their learning experience during the COVID-19 period. The alignment of these results with prior studies may be linked to the predominantly public nature of many of these institutions. Specifically, Makerere University invested significantly during COVID-19, undertaking initiatives such as training both instructors and learners to enhance the overall perception of e-learning. Additionally, efforts were made to improve internet bandwidth, and instructors were provided with data internet bundles (for the initial month), contributing to increased instructor morale in utilizing the system, thereby positively impacting the learners.

Active participation of instructors within the eLearning ecosystem is crucial, as their engagement with digital learning systems naturally extends to the learners. Research emphasizing the impact of PE on BI among instructors suggests that the acceptance of eLearning is feasible when both key users (students and instructors) strongly believe in the utility of these systems (Riaz and Adnan, no date; Lin et al., 2013; Thomas et al., 2013; Thanh et al., 2014; Kocaleva et al., 2015a; Wang, 2016; Daud Mahande and Makassar, 2019). Abbad’s study revealed that students were primarily focused on improving their classroom performance, leading them to perceive Moodle as a technological aid for achieving that goal (Abbad, 2021). In summary, the findings support the connection between PE and the behavioral intention to use e-learning systems, thereby accepting hypothesis Ha1.

Influence of effort expectancy on behavioural intention to use eLearning

The research identified a statistically significant and positive correlation between PE and BI. These results align with and are supported by earlier comparable studies (Kline, 2015; Persada et al., 2019). Similar to other studies (Bellaaj et al., 2015), this study revealed that learners became skilful at using Zoom and MUELE. Although they found it easy to use zoom than MUELE, which aligns with Venkatesh et al.’s. (2003) findings, in certain studies (El-Masri and Tarhini, 2017), it was discovered that the ease of use had no notable impact in developed countries but was significant in developing countries. Their research indicated that students in the USA perceived e-learning as easy to use, attributing it to widespread internet usage, which enhances familiarity with the technology. In the context of this research, the challenges encountered by Makerere students in using MUELE may be partially linked to the fact that some were first-year students, having spent less than 6 months at the university when COVID-19 emerged.

These students had minimal or no familiarity with MUELE, and they had not received any training before the onset of COVID-19. This circumstance clarifies why they experienced a prolonged learning curve in adapting to the system. Additionally, the research highlighted that the students eventually became adept at utilizing both systems (Zoom and MUELE), and their learning activities were clear and comprehensible. This aligns with the results of a comparable study conducted among university students in Indonesia (Indrati et al., 2014). In their findings, the virtual class application was easy to use and useful in achieving the success of teaching and learning process. Although a study conducted by Jayaseelan (2020) and Namatovu et al. (2021) examined the influence of EE on BI in eHealth, a context different from eLearning, their findings are similar to this study’s findings.

Contrastingly, certain studies, such as the one examining the utilization of Massive Open Online Courses by graduate students in India (Mohan et al., 2020), have reported no correlation between EE and BI. However, these scholars noted that the findings differed from students’ location, especially among those with limited experience and access to the internet. Other studies alike found no effect between these two variables (Thomas et al., 2013). The findings suggest that students at Makerere University are inclined to embrace eLearning technologies when they perceive them as user-friendly and easy to grasp.

In line with the recommendations of numerous researchers (El-Masri and Tarhini, 2017; Namatovu and Oyana, 2021), policymakers should consistently emphasize the importance of training novice learners in diverse e-learning services. This training may include aspects like material upload and download, participation in online assessments such as quizzes and discussion forums, and course enrollment, among other elements. Such initiatives aim to influence learners’ perceptions not only regarding the system’s usefulness but also its ease of use. Hence, training plays a pivotal role in the adoption of technology, as increased digital experience correlates with heightened confidence. In this study, effort expectancy emerged as the most influential factor in determining BI, thus affirming the support for the alternative hypothesis Ha2.

Influence of social Influence on behavioural intention to use eLearning technologies

The findings of this study demonstrated a favorable impact of social influence (SI) on learners’ intention to utilize eLearning systems. This outcome aligns with findings from various other studies (Abdou and Jasimuddin, 2020; Daud Mahande and Makassar, 2019; Jaya, 2017; Kayali and Alaaraj, 2020; Sung Youl Park, 2009; Venkatesh et al., 2003; Wang, 2016), thus supporting the acceptance of the alternative hypothesis Ha3. These research endeavors observed that elements like peer support, instructor encouragement, and organizational support had notable effects on an individual’s inclination to adopt educational technologies. In the present study, participants expressed that their peers, teachers, and influential figures in their learning environment, possibly including parents, exerted a strong influence on their decision to engage in e-learning. Similar findings were identified in a study conducted by Wang (2016). In his research conducted in Taiwan, individuals of significance, those exerting influence on learners’ behaviours, and senior management were found to have a substantial impact in motivating learners to adopt e-learning systems.

The favorable influence of social influence (SI) on behavioral intention (BI) in this study is linked to the training provided by lecturers and eLearning coordinators to students. Additionally, administrators played an advocacy role, encouraging students to embrace and utilize eLearning technologies. While SI has been identified as a factor influencing the intention to use digital technologies in this context, it’s worth noting that several studies have reported contrasting findings (Abbad, 2021; Thomas et al., 2013). As exemplified in a research undertaking by Riaz and Adnan (no date), senior students exhibited a higher tendency to remain uninfluenced. This inclination is attributed to the South Asian cultural context, where individuals in this category tend to be less opinionated and more driven by their personal volition, irrespective of external social influences.

Influence of facilitating conditions on usage behaviour of eLearning technologies

The research confirmed a notable connection between FC and UB, thereby supporting the acceptance of the alternative hypothesis Ha4. These results imply the crucial role of FC in shaping the utilization of eLearning systems. As per Venkatesh et al. (2003), FC encompass the organizational and technical infrastructure that facilitates the system’s use. In this investigation, the facilitating conditions that significantly influenced learners to engage with the e-learning system were the availability of resources and the alignment of these systems with students’ preferred learning methods. These findings are consistent with those of previous studies (Riaz and Adnan, no date; Venkatesh et al., 2003; Šumak et al., 2010; Lin et al., 2013; Thanh et al., 2014; Kocaleva et al., 2015b; Wang, 2016).

Makerere University addressed the challenges posed by COVID-19 by providing learners with resources such as digital skills training, internet accessibility, and enhanced bandwidth. Additionally, each college received dedicated e-learning coordinators tasked with resolving digital learning issues for both students and instructors. Training plays a crucial role in reshaping individuals’ perceptions toward technology, reducing resistance and increasing willingness to use a system as proficiency grows. Similarly, the availability of stable and high-speed internet serves as a catalyst for users to engage with a system. It is highly premised that these facilitating conditions immensely contributed to UB, as revealed in a similar study conducted by Riaz and Adnan (no date). In their study, it was revealed that training elderly people to use tablet PC’s for learning purposes improved usage behaviour. Other studies have revealed that possessing digital literacy promotes the adoption of digital systems (Namatovu, 2018; Namatovu and Magumba, 2023).

While the university administration aimed to establish a conducive environment for e-learning adoption, various technical challenges persisted. Instances of the MUELE system being slow or inaccessible occurred frequently, and many colleges faced a shortage of functional computers. Additionally, students lacked Zoom licenses, resulting in disconnections during longer lectures as the non-licensed Zoom application has a time limit of 40 minutes. These infrastructural challenges, to some extent, impeded usage. Contrary to our findings, Alblooshi and Abdul Hamid (2021) discovered that facilitating conditions did not impact the actual usage of eLearning systems among students. Their findings suggest that the presence of technical and operational resources did not necessarily translate into the effective utilization of the systems.

Influence of behavioural intention on usage behaviour of eLearning technologies

The research uncovered a highly significant correlation between the learner’s intention to use eLearning systems and their actual use behavior (UB), aligning with findings from various other studies (Abbad, 2021; Daud Mahande and Makassar, 2019; Jaya, 2017; Kayali and Alaaraj, 2020; Kocaleva et al., 2015a; Wang, 2016; Šumak et al., 2010). The results of this study suggest that students are motivated to use e-learning systems based on their belief that these platforms can improve future performance and enhance learning activities. Additionally, the study showed a positive correlation between PE, EE, and SI on BI. This implies that students generally view e-learning systems as valuable and user-friendly, and when influenced by peers, it contributes to their actual usage.

The latent variables of PE, EE, and SI collectively accounted for 60% of the variance in BI, signifying their substantial influence on students’ intention to use e-learning systems. In turn, BI strongly affected learners’ UB, explaining 69% of the variance in UB. The robust influence of BI on UB, as evidenced in our study, indicates that students with a strong intention to utilize digital learning technologies also exhibited high usage behavior.

Moderating role of age on PE and BI, EE and BI, SI and BI, FC and UB

The research discovered that age does not have a significant impact on the relationships between PE and BI, EE and BI, SI and BI, and FC and UB. This finding does not align with the alternate hypotheses Ha6, Ha7, Ha8, and Ha9. In essence, our study indicates that regardless of whether the eLearning system is perceived as useful or easy to use and learn, age does not play a decisive role in influencing the behavioral intention to use these systems. The result can be elucidated by the fact that the majority of participating students fall within a similar age range. These young adults exhibit a heightened inclination toward utilizing technology, primarily due to the value they ascribe to its usefulness. Importantly, younger scholars are more adept and responsive to technology use, making it easier for them to navigate and employ. A similar study conducted by Bellaaj et al. (2015) seems to also suggest that age is not a moderator of PE and EE. They emphasized that the most important thing is how much the user is familiarized with the system.

Equally noteworthy, the study found that age does not serve as a moderator for the impact of social influence on the behavioral intention to use. This implies that the intention to use remains consistent across students of different ages, irrespective of the presence of opinions from peers, family, or teachers. The influence of social factors tends to diminish from childhood through mid-adolescence and beyond, potentially elucidating why, in this instance, the impact among these students was negligible (Knoll et al., 2017).

Lastly, of equal significance, the study determined that age does not act as a moderator between facilitating conditions (FC) and use behavior (UB). This suggests that the age of the students did not impact their system usage, and this pattern is likely to persist in the future regardless of the facilitating conditions. Our findings, however, are in contrast to Venkatesh et al. (2003), who found age to be a moderating factor in social influence, performance expectancy, and effort expectancy. Nevertheless, we firmly believe that the results might vary if learners of an older age group are taken into consideration.

Influence of gender on the relationship between performance expectancy and Behavioural Intention

The study revealed that the moderating effect of gender on the relationship between PE and BI was not significant. This contrasts with the outcomes of previous studies (Akbulut and Kaya, 2015; Huang et al., 2017; Liaw and Huang, 2013). These studies propose that gender may play a moderating role in the connection between perceived ease of use (PE) and behavioral intention (BI) within different technological contexts, such as eLearning and the internet. The nature of the moderation effect differs among studies, with some indicating a more pronounced effect of PE on BI for women and others observing a stronger effect for men. Our study results can be clarified by the fact that all students, irrespective of gender, shared the same learning environment, received training from identical instructors, and had equitable access to the same ICT infrastructure. This implies that the intention to use the e-learning system, given the constancy of all aforementioned conditions, would be uniform regardless of the perceived usefulness of the system by the students. In line with our study, Bellaaj et al. (2015), also discovered that gender does not serve as a moderating variable for the relationship between PE and BI. They propose that the pivotal determinant is not the user’s gender but, instead, their familiarity with the technology.

Influence of gender on the relationship between effort expectancy and behavioural Intention

The research identified a notable impact of gender on the connection between EE and BI. This implies that a student’s gender played a significant role in influencing their intention to use e-learning systems, particularly when they perceived these systems as user-friendly. While we did not explicitly determine the gender more inclined towards adoption, which is a limitation of the study, the research consistently aligns with other comparable studies. For example, Alaiad and Zhou (2017) study found that the relationship between EE and BI was moderated by gender. Several prior studies have presented conflicting results. For instance, Menachemi (2011) discovered that EE was a significant predictor of BI for both male and female healthcare providers, with no notable interaction effect of gender on this relationship. Similarly, in a comparable study, it was revealed that EE significantly predicted BI for both male and female eLearners, without any notable interaction effect of gender on the relationship between EE and BI (Esteban-Millat et al., 2018).

Influence of gender on the relationship between social influence and behavioural intention

Regarding the moderating impact of gender on the association between SI and BI, the research did not identify any effect. The results suggest that gender did not play a role in shaping students’ intentions to use e-learning, even in the presence of motivation from their peers to utilize the system. Whereas the study indicated a positive effect between SI and BI, the moderating effect of gender did not affect this relationship. In summary, it can be concluded that at Makerere University, the intention to use e-learning was not influenced by peers or individuals within one’s social circle, irrespective of whether the student was male or female. This phenomenon can be elucidated by the notion that university students, being deemed adults, autonomously make decisions and are therefore less susceptible to peer influence in adopting e-learning, regardless of their gender. This aligns with a study in which social influence demonstrated a notable positive impact on students’ behavioral intention to use eLearning environments, without any significant interaction effect of gender on this relationship (Tarhini et al., 2013). Likewise, in the public sector of Taiwan, Wang (2016) discovered that gender had no discernible impact on the relationship between social influence (SI) and behavioural intention (BI), signifying that gender did not serve as a significant moderator between SI and BI in the context of e-learning.

Influence of experience on the relationship between performance expectancy and behavioural Intention

This study unveiled the absence of a noteworthy moderating effect of experience on the relationship between PE and BI, aligning with similar findings in studies such as Wang (2016). In Wang’s research, it becomes evident that the moderating influence of experience on the connection between PE and BI is not prevailing among e-learning users. In essence, our study’s findings imply that the e-learning experience of learners did not impact their behavioral intention to use eLearning systems, even when the system was perceived as useful. The outcomes in this study might be elucidated by the fact that over 70% of the participating students were enrolled in computing programs and had spent no more than 1 year at the college, thereby sharing relatively similar MUELE experience. Contrary to our findings, Bellaaj et al. (2015), observed that the impact of performance expectancy on the intention to use eLearning systems became more pronounced as experience increased.

Influence of experience on the relationship between effort expectancy and social influence upon the student’s behavioural intention

This study did not identify a moderating effect of experience on the relationship between EE and BI, as well as social influence (SI) and BI. These findings align with those of other studies (Riaz and Adnan, no date; Wang, 2016), leading to the rejection of the alternate hypotheses Ha14 and Ha15. The results suggest that the students’ experience had no impact on their intention to use e-learning, even when they perceived it as easy to use or were influenced by their peers to do so. As the research was conducted among undergraduate students who had limited or no prior experience with MUELE, their ability to influence each other’s intention to use is considerably low. This aspect may contribute to the interpretation of these findings. Likewise, these learners underwent identical training and utilized the same infrastructure, suggesting that the level of experience in using e-learning systems could be assumed to be relatively uniform among them. While some students may have had experience with various learning management systems, MUELE, which was predominantly utilized, could differ from those they had previously encountered. In their results, Riaz and Adnan (no date) categorized the respondents’ experience into three groups: those with less than a year of experience, 1-3 years, and more than 3 years. They did not observe any effect of individuals with different years of experience on the relationship between EE and BI.

Influence of experience on the relationship between facilitating conditions and use behaviour

Equally noteworthy, our study unveiled a robust impact of the experience moderator variable on the relationship between FC and UB, in contrast to the findings of Wang (2016). These results suggest that as students gain more experience, their utilization of e-learning systems increases, and consequently prompting enhanced support from management in the form of improving existing infrastructure. As a student accumulates more experience with a technology, their self-efficacy also increases, directly contributing to a heightened motivation to persist in using the technology. The ongoing utilization results in added strain on current IT resources, equating to a heightened demand for enhanced infrastructure, including fast and reliable internet, functional computers, and more. In the context of Makerere University, the College of Computing and Information Sciences achieved the top ranking in the utilization of MUELE, as reported by Mugenyi in 2021. Therefore, it is unsurprising that students from this college persist in using the system, contributing to the ongoing updates, with the most recent one implemented in the last quarter of 2023. A study conducted in the United States among high school and college graduates uncovered that individuals with more experience in the workforce were more inclined to adopt and persist in using computers compared to their counterparts (Weinberg, 2004).

Theoretical implications

Utilizing the UTAUT approach, the theoretical implications of this study arise from our endeavor to broaden the literature concerning the utilization of video conferencing software (Zoom) in conjunction with a learning management system (MUELE) during the challenging period of COVID-19, an area that has not been extensively explored. The results provide valuable insights into the adoption and acceptance of these two systems within the context of a developing country.

Moreover, the research underscores that performance expectancy, effort expectancy, and social influence are influential factors shaping the behavioral intention to use eLearning at Makerere University. Simultaneously, it identifies behavioral intention and facilitating conditions as predictors of usage behavior, with the latter exerting the most substantial influence on use behavior, as evidenced by a total effect of 0.44. A noteworthy aspect of this study is the prominence of effort expectancy as the most influential predictor of behavioral intention, yielding a total effect of 0.455 and accounting for 61% of the variance in the behavioral intention variable. Social influence emerged as the second most influential predictor, registering a total effect of 0.239 and explaining 61% of the variance. In contrast, performance expectancy exhibited the least influence, with a total effect of 0.187. We consider the explained variance of behavioral intention to be high because it is approaching 70% of Venkatesh et al. (2003). In contrast to age and experience, which were determined to lack a moderating effect on BI and UB, gender was identified as a moderator for both SI and EE on BI. These findings can be generalized to other public universities in Uganda, as the funding source, governance structure, and student admission criteria are slightly similar across these institutions.

Practical implications

Firstly, e-learning developers should create useful products for learners for instance, building intuitive systems that facilitate easy navigation and give learners a better user experience. The intuitiveness could involve incorporating intelligent features such as chatbots and personalized resource features such as intelligent tutoring systems, utilizing artificial intelligence to adapt to the individual needs of students.

Secondly, universities should provide regular training sessions for students to improve their proficiency in using digital systems and enhance their confidence. The study results indicated low scores on specific questions regarding whether students found MUELE easy to use and the time it took them to learn to use it. Policymakers and advocates for e-learning can utilize these findings to develop customized training programs specifically concentrating on system usability. This may include instructions on how to (i) upload or download a file, (ii) create a comment, (iii) attempt quizzes, and (iv) enrol in a course. Additional features to augment the learner’s understanding, such as user manuals for MUELE, an online FAQ chatbot, and peer discussion forums, can be integrated into the system, complemented by offering personalized support through e-learning champions.

In conclusion, this study offers valuable insights into the factors impacting the adoption of eLearning among students at Makerere University. This information can be utilized by policymakers, management, and eLearning system developers to shape policies. As an example, a policy mandating all instructors to make their teaching materials available on MUELE could significantly facilitate learners’ easy access to materials. This, in turn, may encourage continuous use of the system, ultimately enhancing their proficiency in navigating it.

Limitations and future directions

Firstly, because of the emergence of the COVID-19 pandemic, this study was confined to a questionnaire approach, and we acknowledge that interviews might have provided the study with more comprehensive information. The primary challenge we encountered revolved around the effective administration of questionnaires due to mutual apprehension between respondents and researchers. This fear led to some respondents not returning the questionnaires, initially impacting the sample and subsequently influencing the research outcome.

Secondly, the study did not explore the impact of certain parameters, including attitude, trust, and users’ location, on the adoption of eLearning systems. These variables were omitted from the investigation as it was the researchers’ intention to limit the number of predictors and instead focus on the constructs of the original UTAUT model (Venkatesh et al., 2003). Nevertheless, we firmly assert that the influence of these variables should be explored in future research. Introducing trust into the model assesses learners’ decisions, predominantly influenced by security concerns and their trust in the system (El-Masri and Tarhini, 2017). Measuring attitude would help measure one’s way of thinking since it is highly hypothesized that a positive attitude has an uphill effect on technology acceptance and adoption. Lastly, adding user’s location to the model would be beneficial especially in the developing world were internet access is limited to the urban areas. On the flip side, determining the moderating effect of gender on EE and BI, specifically identifying which gender is more inclined to adopt, is crucial for university management in determining where to focus their efforts.

Thirdly, our pool of respondents was restricted. Subsequent research endeavors should encompass additional stakeholders, including university top management, lecturers, and system administrators. Incorporating these stakeholders is essential as the e-learning ecosystem involves various participants who can directly or indirectly impact its adoption. Examining their role in adoption is significant, as it can serve as a foundation for establishing or revising certain e-learning policies, such as making it mandatory for all instructors to have MUELE accounts and Zoom licenses. Lastly, incorporating these stakeholders aids in comprehending the barriers and/or facilitators to adoption from both the supply and demand sides, facilitating the creation of a comprehensive solution.

Finally, over 70% of the participants were affiliated with the College of Computing and Information Sciences, where students possess a specific level of digital literacy and self-efficacy. Consequently, the findings may not accurately represent the overall adoption scenario within the university. The selection bias limits the broader generalizability of the study findings. Subsequent research should focus on humanities colleges, where the adoption of MUELE is still notably low, as indicated by Mugenyi (2021). Gaining insights into e-learning adoption from the viewpoint of e-learners lacking or having minimal digital skills offers a comprehensive understanding of e-learning challenges. This, in turn, equips management with focal points to concentrate human, financial, and infrastructural resources. Presently, the findings are derived from science-based colleges, and their inclusion would mitigate bias by ensuring the representation of arts-related colleges in the context of e-learning adoption.