Motivators for University of Professional Studies Accra Students to Adopt a Learning Management System in Ghana

1 Introduction

Organizations are devising new business models in order to avoid the ravages of coronavirus. While some organizations allow employees to work from home, others have divided the workforce so that one part can work from home, while the other is physically present in the office. These arrangements are currently available in Ghanaian universities. At the University of Professional Studies, Accra (UPSA), for example, one group is engaged in face-to-face teaching, while the other group relies on distance learning for the first half of the semester. The groups then reverse their mode of study during the second half of the semester. As a result of these arrangements, these universities have invested in learning management systems (LMSs) to manage online learning.

LMS software is a collection of technologies that let educational instructors communicate learning information to students via the internet while also managing and monitoring their learning progress throughout the learning process (Ramesh & Ramanathan, 2013). It provides the “big picture” of learning objectives, content to be taught, learning activities, requisite resources, and class management – all while adhering to the instructional paradigm chosen for the unit (Metzler, 2017). Despite the existence of numerous popular LMSs around the world, such as Sakai, Moodle, Blackboard, ATutor, and others, most institutions in Ghana have installed Moodle’s open-source LMS. The UPSA, and the University of Education, Winneba, which are public universities in Ghana, and Methodist University Ghana, a private university, are some of the universities in Ghana using Moodle for online learning. The University of Ghana on the other hand uses Sakai to deliver course content to the learners. Hundreds of millions of students utilised Moodle LMS, a free and open-source LMS. It is extensively utilised in educational institutions because it supports a flexible learning environment, is simple to use on mobile devices, available to everyone, and is incredibly secure.

To avoid disrupting the academic calendar, most universities have resorted to using the LMS to transmit course content to students (Olujuwon, 2021). While the LMS may be effective in certain advanced countries, it may not be so in developing nations. Most of the university remedies have failed to address the problem from the students’ standpoint. According to Kaewsaiha and Chanchalor (2019), students find it difficult to use the LMS from outside the university due to issues with internet connectivity, the user interface, using specific LMS capabilities, unclear instructions, confusing evaluation criteria, lack of examples, time to complete activities did not match the actual time required, and receiving no relevant feedback or no feedback at all, among other things. According to Alebaikan and Troudi (2010), higher education institutions must offer instructional practice training for instructors, user-friendly LMS features, emphasize the importance of avoiding plagiarism, and provide orientation, resources, and time management strategies for instructors and students in order to ensure efficiency in online discussions using the LMS.

Many research on the use of LMS by Ghanaian students have employed technology acceptance model (TAM) or a mix of TAM and other factors as their underlying model, as well as a variety of analytical approaches, to investigate the phenomenon. TAM and multiple regression (Essel & Wilson, 2017); TAM and general linear model (Asampana et al., 2017); TAM plus computer self-efficacy, structural equation modelling (Boateng, Mbrokoh, Boateng, Senyo, & Ansong, 2016; Budu, Yinping, & Mireku, 2018); and TAM plus institutional factors and correlation and regression (Okantey & Addo, 2016). The findings of these research show that several variables influence the students’ behavioural intentions (BIs) to use and actual use of LMSs. The work of Budu et al. (2018) found perceived ease of use (PEOU), perceived usefulness (PU), and self-efficacy significantly influencing students’ BI to use LMS. In a related study by Boateng et al. (2016), they identified attitude and PU as the first and second most significant factors in determining students’ BI to use an LMS.

Boateng et al. (2016) proposed that it is important to take into account the antecedent of PU and its impact on BI, as PEOU did not directly correlate with the students’ BI, but rather, did through PU. Essel and Wilson (2017) argue that the benefits of employing interactive and collaboration technologies to promote learning should be increased because students appear to use these features infrequently. They also proposed that efforts be made to alleviate students’ issues with Moodle, such as internet connectivity and navigation, in order to boost their Moodle usage rate.

According to the findings of Adjin-Tettey (2014) from the UPSA, students’ perceptions of the Moodle course management system are overwhelmingly positive, and this has a direct impact on users’ intentions to use the system for the remainder of their campus studies, as evidenced by the correlation test results. Similar research from the same university, however, by Asampana et al. (2017), indicates that students’ acceptance of LMS is very poor, blaming it on a lack of information technology infrastructure, insufficient training, and the system’s relevance to great lecture delivery. Furthermore, despite the plethora of studies on students’ LMS adoption in the literature, evidence on the motivators of Ghanaian students’ BI and actual LMS usage behaviour in respect to the third version of TAM (TAM3) is scarce. Boateng et al. (2016) used a covariance-based structural equation model approach in Ghana, which has less statistical power to examine the students’ intentions when several factors are present. In addition, the research did not look at how PEOU and PU were modified by external circumstances. It is therefore important to address the gaps in the literature with regards to Ghanaian students’ adoption of LMS.

Therefore, the purpose of this research is to explore the motivators for LMS adoption by UPSA students in Ghana using the partial least squares structural equation model (PLS-SEM) and a simplified version of the TAM3 model. The study specifically attempts to analyse the external factors that influence PEOU and PU, as well as the value of the constructs to students’ use of the LMS.

2 Literature Review

Administrators at higher education institutions deal with risks and uncertainty in a variety of ways. Some are useful, while others have a negative impact on the students’ academic achievement. In a period of major curricular changes and vastly variable student performance in a chaotic environment, it is critical to examine and understand the students’ behaviour on the use of LMS in relation to their academic achievement.

An LMS is a collection of software tools for managing user learning interventions, courses, instruction, online communication, and assessment and testing (Nof, 2009). It also enriched instruction by encouraging communication among students and between students and teachers, sharing documents by providing access to course materials and assignments, tests and test results, an e-discussion and chat room, and other educational features (Taylor, Handler, FitzPatrick, & Whittingham, 2020). Furthermore, it allows teachers to assess the depth of student learning and provide helpful comments by facilitating self-directed learning practice (Dent, Harden, & Hunt, 2021). The most extensively used LMS is Moodle in the Ghanaian context.

Despite the LMS’s many features, there are several drawbacks to its success when it comes to student usage. According to Ssekakubo, Suleman, and Marsden (2011), the most common causes of LMS failure are poor levels of student familiarity with technology, low levels of information and communication technology literacy, and a lack of assistance. The hexagonal e-learning assessment model identifies six types of success factors that influence e-learner satisfaction (supportive issues; social issues: learner perspective; social issues: instructor attitudes; technical issues: system quality, information (content) quality, service). Notwithstanding the significant expenditure to build and maintain the LMS and other technologies, there is evidence that users do not fully utilise them (Mtebe, 2015).

The interaction between students and LMS during the COVID-19 period in Ghana has been the subject of several research, which have revealed significant evidence. Dampson et al. (2020) contend that most users of the University of Education, Winneba will stop using the LMS platform for teaching and learning activities after COVID-19, and that the use of the LMS will cease whenever users have the option of choosing between the LMS and the traditional classroom approach to instruction. The majority of University of Cape Coast (UCC) students, according to Agormedah et al. (2020), were aware of e-learning and use tools like Google Classroom, Alison, and the UCC Moodle platform. However, they were unprepared for the transition to remote learning because they had not been introduced to it, and they perceived constant internet connectivity as a challenge due to financial insecurity, which leads to resistance from some students, which could affect their remote instruction engagement activities and academic success. Furthermore, Agormedah et al. (2020) argue that students’ unpreparedness, as well as a lack of formal training and experience in e-learning platforms, can have a negative impact on their behaviour outcomes such as engagement in learning, satisfaction, participation, motivation for learning, online work skills proficiency, self-directed learning, and efficacy in the use of e-learning devices, as well as their academic performance. In a related study in Ghana, Opoku (2021) found that learning distractions from family, environment, and technology negatively impact students’ Moodle usage, which in turn negatively impacts the academic performance. The study further reveals that students’ academic performance had declined after taking tests and quizzes during the Covid-19 lockout.

TAM was created to forecast information technology acceptance and utilisation on the job. It is suitable for information systems environments (Lee, Kozar, & Larsen, 2003). TAM has been applied to a wide range of technologies in a variety of settings. According to the authors, PU and PEOU have a direct impact on BI. In addition, the authors claimed that PEOU has a direct impact on PU.

Venkatesh and Davis (2000) examined the consistent and large influence of PU on BI across all research investigations, as well as the inconsistent effect of PEOU on BI. Venkatesh and Davis (2000) devised the extended TAM (TAM2) as a way of studying the elements that influence PEOU after discovering these trends in the literature. TAM2 incorporates additional social influence processes (subjective norm (SN), voluntariness, and image) as well as cognitive instrumental processes (job relevance (REL), output quality (OUT), result demonstrability, and PEOU). Also, Venkatesh (2000) looked into the factors that influence PEOU and came to the conclusion that individuals form early impressions of a system’s PEOU based on a number of factors related to people’s general attitudes toward computers and how they are utilised. Venkatesh and Bala (2008) established TAM3 by integrating the factors from Venkatesh and Davis (2000) and Venkatesh (2000).

Figure 1, which was adapted from Venkatesh and Bala (2008), shows a graphic representation of TAM3. TAM3 expands on the technology acceptance model by arguing that image, REL, OUT, result demonstrability, and SN all predict PU. Furthermore, they propose that computer self-efficacy, perception of external control, computer anxiety, and computer playfulness are anchor variables that have a direct influence on PEOU. The model also suggests that experienced enjoyment and objective usability are adjustment factors that predict PEOU. According to the authors, experience moderates the relationships between (i) PEOU and PU, (ii) computer anxiety and PEOU, and (iii) PEOU and behavioural intention. The relationship between SN and BI is also moderated by voluntariness of use.

Figure 1

The TAM3 model.

Although TAM3 could not account for crossover effects, Abdullah and Ward (2016) developed the General Extended Technology Acceptance Model for E-Learning to be used in the context of e-learning adoption.

Recent empirical studies on technology adoption in higher education that used TAM or a modified version of TAM as a conceptual framework to investigate students’ adoption of technology suggest that PU and PEOU have varying effects on students’ intention to use the technology. A slew of research has confirmed that PU and PEOU have a major impact on students’ intentions to use LMS (Al-Azawei, Parslow, & Lundqvist, 2017; Kaewsaiha & Chanchalor, 2021; Tarhini, Hone, & Liu, 2013). Other studies (Agudo-Peregrina, Hernández-García, & Pascual-Miguel, 2014; Boateng et al., 2016; Binyamin, Rutter, & Smith, 2018) claim that PEOU has no influence on students’ BI. Furthermore, while Binyamin et al. (2018) and Kaewsaiha and Chanchalor (2020) discovered a substantial association between BI and actual usage of LMS in higher education institutions, Agudo-Peregrina et al. (2014) found no support for this assumption.

The literature also shows that other external variables influence students’ perceptions of the usefulness of LMSs. For example, Agudo-Peregrina et al. (2014), Al-Gahtani (2016), and Kaewsaiha and Chanchalor (2020), discovered REL as an important variable in determining students’ LMS PU. Furthermore, Hanif, Jamal, and Imran (2018), as well as Abdullah, Ward, and Ahmed (2016), demonstrated that students’ perceived enjoyment (ENJ) determines PEOU of LMS. Furthermore, Agudo-Peregrina et al. (2014) confirmed that personal innovation in IT has an impact on PEOU in higher education. In light of the disparate findings from the preceding studies, which may be attributed to differences in culture and student familiarity or experience with LMS, a study employing a modified TAM to explore the factors affecting UPSA students’ LMS use behaviour (UB) is required.

3 Conceptual Framework and Hypotheses

The technology acceptance model and its extensions have been applied in a variety of situations and have been found to be useful in explaining students’ intention and actual use of LMS (Al-Azawei et al., 2017; Binyamin et al., 2018; Kaewsaiha & Chanchalor, 2020). Thus, the study adapted TAM3 as the base model for the study. There were four core variables: REL and OUT, which directly influence PU, and SN and ENJ, which directly influence PEOU. There are three mediating variables (PEOU, PU, and BI). Finally, actual UB was an endogenous variable. The conceptual model that the study is based on is depicted in Figure 2.

Figure 2

Conceptual framework.

4 Methodology

The study used cross-sectional survey design and a convenient sampling technique to elicit data from the population of first-year students in the faculty of information technology and communication studies at UPSA to address the objective of the study. The study chose that population because they were new to using the LMS for academic work. The findings may drive administrators to develop rules that could encourage new students to use the LMS. According to Hair, Hult, Ringle, and Sarstedt (2017), PLS-SEM performs well with fewer data and a more complicated model and makes few assumptions about the underlying data. Thus, due to the study’s structural model complexity, which includes various variables, indicators, and model linkages, the study used PLS-SEM to better understand UPSA students’ behaviour by exploring TAM3 performance in the Ghanaian educational context. The study relied on first-year student class representatives from the faculty to identify volunteers who agreed to participate in the survey. Following that, they received an email at the end of the academic year where all students have used the LMS with a link to a Google form. Only 188 of the 200 volunteers completed the survey.

The online survey was created with Google Forms by adapting Venkatesh’s and Bala’s (2008) questionnaire items and scales. The items were assessed using a seven-point Likert scale. As part of the survey’s introduction, participants were told that the responses they supplied would be kept private. The second and third sections were utilised to collect demographic information and data relevant to the study’s goal. Prior to collecting data, the Institutional Research Ethics Committee granted its approval for this study. All individuals provided informed consent prior to participation in the study, which was carried out in compliance with the ethical guidelines specified in the Belmont Report and the Declaration of Helsinki. The rights, welfare, and dignity of the study participants were protected by the researchers according to all ethical rules and regulations established by the institutional review board or ethical review committee

To evaluate the reflective measurement model, the study used composite reliability and individual indicators to examine the internal consistency reliability, the average variance extracted (AVE) to test convergent validity, and the Heterotrait-Monotrait ratio (HTMT) to assess discriminant validity. Composite reliability values range from 0 to 1, with higher values indicating higher levels of reliability. In exploratory research, measured values of 0.600–0.700 as composite reliability are acceptable. In complex research, values between 0.700 and 0.900 are considered satisfactory (Hair et al., 2017). Furthermore, factor loadings greater than 0.708 and AVE greater than 0.500 are sufficient to assess convergent validity. According to Hair et al. (2017), to confirm discriminant validity, the HTMT values must be less than 0.850 for the stricter condition and less than 0.900 for the lenient condition; additionally, the HTMT values must be significantly different from 1.

To evaluate the structural model, the study looks at the model’s predictive capabilities as well as the relationships between the latent variables. The primary evaluation criteria for PLS–SEM results are the check for multicollinearity (VIP), coefficients of determination (R ²) values, and the t-values of the path coefficients. The study used the bootstrapping routine to assess the significance of the path coefficient. As a rule of thumb, the study used 5,000 subsamples, which exceeded the number of valid observations. Other measures examined in the study include effect size (f ²), predictive relevance (Q ²), and effect size (q ²). Path coefficients in structural models represent the hypothesized relationships between latent variables. Critical values for a two-tailed test at 10, 5, and 1% significance levels are 1.65, 1.96, and 2.57, respectively. In this study, path coefficients with a 5% chance of error were considered statistically significant.

5 Results

5.3 Structural Modelling

The results of each predictor construct’s tolerance, variance inflation factor (VIF), given in Table 4 fall within the acceptable range of 0.2–0.5 using the rules of thumb for structural model evaluation outlined by Hair et al. (2017). As a result, there are no issues with collinearity between the constructs.

Table 4

Variance inflation factor

	BI	ENJ	OUT	PEOU	PU	REL	SN	UB
BI								1.000
ENJ				1.420
OUT					2.035
PEOU	1.921				1.978
PU	1.921
REL					1.290
SN				1.420
UB

To determine the significance of path coefficients, the researcher utilised a bootstrapping approach with 5000 samples. Apart from the association between PEOU and BI, which is negligible, the bootstrapping findings in Figure 3 and Table 5 reveal that all the relationships in the model are significant at the 5% significant level. As a result, the data support hypothesis H1, H2, H3, H4, H5, H6, and H8, but not H7.

Figure 3

Bootstrapping t values in the structural model.

Table 5

Results of structural models

	Original sample (O)	Sample mean value (M)	Standard deviation (STDEV)	T Statistics (|O/STDEV|)	P Values	Effect size (f 2)	Decision
H1: REL→ PU	0.274	0.268	0.098	2.780	0.005	0.152	Supported
H2: OUT → PU	0.314	0.311	0.124	2.524	0.012	0.127	Supported
H3: ENJ → PEOU	0.584	0.577	0.097	6.024	0.000	0.571	Supported
H4: SN → PEOU	0.265	0.277	0.088	2.996	0.003	0.118	Supported
H5: PEOU → PU	0.360	0.368	0.132	2.731	0.006	0.172	Supported
H6: PU → BI	0.658	0.655	0.110	5.993	0.000	0.479	Supported
H7: PEOU → BI	0.095	0.097	0.118	0.806	0.420	0.010	Not Supported
H8: BI → UB	0.676	0.678	0.060	11.267	0.000	0.840	Supported

The coefficient of determination (R ²) shows the amount of variance in the endogenous constructs explained by all the exogenous constructs linked to it (Hair et al., 2017). That is, the coefficient measures the amount of variance in the endogenous constructs explained by all of the exogenous constructs associated with it. It is thus a measure of the model’s predictive power. Generally, R ² values of 0.250, 0.500, and 0.750 are considered to be weak, medium, and substantial, respectively (Hair et al., 2017). Thus, the R ² values shown in Table 6 suggest that ENJ and SN explained 60.1% of the variance in PEOU; OUT, REL, and PEOU explained 61.9% of the variance in PU; PU and PEOU explained 52.9% of the variance in BI, and BI explained 45.6% of the variance in UB. The predictive accuracy of the exogenous constructs on PEOU, PU, and BI is medium whereas the predictive accuracy of BI on UB is weak.

Table 6

Effect size analysis and predictive relevance

	Coefficient of determination			Predictive relevance
Constructs	R 2	f 2	Decision	Q²	q 2	Decision
UB	0.456			0.226
BI		0.840	Substantial
BI	0.529			0.276
PU		0.479	Substantial		0.176	Medium
PEOU		0.010	Small		0.009	Small
PU	0.619			0.417
OUT		0.127	Small		0.070	Small
REL		0.152	Medium		0.075	Small
PEOU		0.172	Medium
PEOU	0.601			0.360
ENJ		0.571	Substantial		0.263	Medium
SN		0.118	Medium		0.056	Small

The effect size is used to determine how much a predictor construct contributes to the R ² value of the target constructs (f ² ) (Hair et al., 2017). Effect sizes of 0.02, 0.15, and 0.35 are evaluated as small, medium, and substantial, respectively, by Hair et al. (2017). The f ² values are shown in Table 6. While SNs have a moderate influence on PEOU, enjoyment has a substantial effect. PEOU and REL have medium impact sizes on PU, but OUT has a modest effect size. Furthermore, in terms of BI, PU has a large effect size, but PEOU has a relatively tiny effect size. Finally, the impact of BI on actual UB is significant. Table 6 further shows the predictive relevance of the model with respect to endogenous variables. Since Q ² is greater than 0, the findings clearly corroborate the model’s predictive relevance with regard to the latent variables. The impact sizes of ENJ and SN are medium and small. OUT and REL also have a small impact size on PU. Furthermore, PU and PEOU have medium and small impact sizes on BI.

5.4 Importance-Performance Map Analysis (IPMA)

The IPMA seeks to identify which elements have poor performance but high importance for the target construct (Hair et al., 2017). The IMPA is a reliable and practical analysis in PLS-SEM that extends the usual path coefficients in a more practical method. In this study, BI to use the LMS predicts the students’ actual use of the LMS. The precursors of BI are PU and PEOU. REL and OUT, enjoyment and SN, respectively, predict the PU and PEOU. The study employed UB as the target construct and ran the IPMA. The performance scores were created by rescaling the latent variables score range, with 0 being the lowest score and 100 being the highest. The importance scores were obtained from the total direct influence of outcome variables using the structural equation model (Hair et al., 2017).

Table 7 presents the total effects (importance) and index values (performance) used for the IPMA. It can be seen that BI is the most important factor in order to determine the UB of LMS by students due to higher importance values compared to other latent variables. PU is at an intermediate level, while PEOU has the lowest importance level after PU. The level of importance and performance are shown in Table 7.

Table 7

Index values and total effects

Latent variables	Total effect of the latent variable LMS UB (importance)	Index values (performance)
BI	0.639	76.988
ENJ	0.093	66.750
OUT	0.084	64.274
PEOU	0.182	68.742
PU	0.279	68.986
REL	0.092	80.342
SN	0.046	72.090
Average	0.202	71.167

According to the IPMA in Figure 4, SN and REL are in the overkill quadrant, ENJ, OUT, and PEOU are in the lower priority quadrant, PU is in the concentrate here quadrant, and BI is in the keep the good work quadrant (Martilla & James, 1977).

Figure 4

Importance-performance analysis map.

The IPMA in Figure 4 clearly shows that higher education institutions must take suitable efforts to increase the students’ PU of LMS. This can be accomplished by utilizing resources that can be used to enhance students interaction among themselves and the relevance of LMS’s on enhancing students’ PU of LMS. Higher education administrators must also continue the positive effort that inspires students’ BI to use the LMS.

6 Discussion

The LMS is widely used by higher education institutions that use either an entirely online or a blended learning approach. However, only a few studies have used an extended TAM to investigate the determinants of Ghanaian higher education students’ BI and actual use of LMS. Based on the TAM3, this study chose two external variables to predict PEOU (SN and ENJ) and PU (REL and OUT). The results showed that the predictors of PEOU and PU explained 60.1 and 61.9% of the variance, respectively. Furthermore, the model explained 52.9% of the variation in BI and 45.6% of the variation in actual UB, respectively. Although the effect size of SN on PEOU is moderate, the effect size of ENJ is significant, and the effects of PEOU and REL on PU are moderate whereas the effect size of OUT is small. Given ENJ’s significant predictive power on PEOU and SNs’ moderate predictive power on PEOU, as well as REL’s and PEOU’s moderate predictive power on PU, it is crucial to pay attention to these factors while using LMS as a study tool.

First, the results show that ENJ is positively related to PEOU of LMS. This demonstrates that as students have fun with the use of the LMS, students’ perception of LMS’ ease of use increases. This finding is consistent with the findings of Venkatesh and Bala (2008) and other studies in higher education institutions (Al-Gahtani, 2016; Abdullah et al., 2016; Hanif et al., 2018). The students’ attention is drawn away from any difficulties in using the LMS due to their happiness as a result of their use of the LMS. It is therefore critical for designers to ensure that the LMS interface is user-friendly. Furthermore, essential options must be easily accessible to students in order for them to complete their assignments.

Second, the research found a connection between SN and PEOU. This illustrates that as key persons in students’ lives encourage them to use the LMS, students perceive the LMS to be simple, which increases the BI to use the LMS, which effects actual usage. This result is in line with the previous work (Farahat, 2012; Revythi & Tselios, 2019) in a higher education context. The finding, however, contradicts previous research in a similar context (Park, Nam, & Cha, 2012; Salloum, Alhamad, Al-Emran, Monem, & Shaalan, 2019; Binyamin et al., 2018). Higher education administrators must devise ways for peers and tutors to make other students aware of the LMS’s value.

Third, the study showed a strong positive relationship between REL and LMS PU. According to the findings, the more students perceived the LMS to be relevant to them, the more they perceived it to be useful. The outcome is consistent with the work of Kaewsaiha and Chanchalor (2020) and Venkatesh and Bala (2008) in the higher education context. As a result, it is critical for LMS administrators to ensure that students realise the significance of the LMS to their studies by responding to all academic concerns about its use.

Fourth, the study established that there is a significant positive relationship between OUT and PU of LMS, resulting in the higher the quality of the LMS output, the more students regard the LMS as useful. This finding supports the findings of Venkatesh and Davis (2000) and Venkatesh and Bala (2008). The finding is also consistent with the findings of Teo et al. (2019) in higher education settings. When the course materials on the LMS are simple, adequate, and easy to understand, students perceive OUT to be high. They also believe that lecture delivery is comparable to face-to-face learning and that internet connectivity is stable. Higher education administrators must ensure that course materials are available online.

Fifth, the study discovered that PU is influenced by PEOU. This finding suggests that as long as LMS are simple to use, students see them as useful. This conclusion is similar to that reached by Venkatesh and Bala (2008), as well as other empirical studies in higher education (Al-Gahtani, 2016; Kaewsaiha & Chanchalor, 2021). In a comparable setting, however, this result contradicts Saroia and Gao (2019). Though the relationship between PEOU and BI is small, it is consistent with the findings of Agudo-Peregrina et al. (2014), Boateng et al. (2016), and Binyamin et al. (2018) in the educational setting. It should be noted that there is an indirect relationship between PEOU and students’ BIs. Therefore, it is critical for administrators and designers to ensure that students perceive LMS use to be simple for learning.

Sixth, the study revealed that PU has a significant effect on BI. This finding shows that as students perceive the LMS to be useful, their proclivity to form an intention to use the LMS increases. This relationship supports the claims made by Davis (1989) and Venkatesh and Bala (2008). The discovery is also consistent with previous empirical work in higher education (Agudo-Peregrina et al., 2014; Al-Gahtani, 2016; Kaewsaiha & Chanchalor, 2021). The IMPA results also clearly show that PU is important to students’ actual use of the LMS. As a result, resources that may be utilised to enhance SNs and REL but are overkill can be used to improve students’ PU of LMS. Higher education administrators must develop rules that enhance students’ PU of LMSs.

Finally, the findings show that BI and actual usage have a substantial positive association. This finding shows that when students’ BI to use a LMS grows, so does their actual use of the system. Venkatesh and Bala’s (2008) hypothesis is supported by this finding. The result is also consistent with other empirical investigations in the field of higher education (Binyamin et al., 2018; Kaewsaiha & Chanchalor, 2021). However, the findings contrast those of Agudo-Peregrina et al. (2014). As a result, regardless of how simple or complicated the LMS is to use, administrators must ensure that learners find it useful. The IMPA finding confirms the importance of BI in students’ LMS use. As a result, higher education administrators must ensure that policies set in place to awaken students’ BIs are either maintained or improved so that students can use the LMS for heightened academic work.

7 Conclusion, Limitation, and Future Direction

The current study sought to explore the motivators for LMS adoption by UPSA students in Ghana using the PLS-SEM and a simplified version of the TAM3. The study evaluated the key factors that influence LMS usage and offers some implications for stakeholders in the higher education sector. The study drew a sample of 188 students from the faculty of Information Technology of the UPSA in Ghana. A modified version of TAM3 is used as the study’s framework. According to the study’s findings, BI has a significant influence and importance on students’ actual use of LMS. Though there is no direct relationship between PEOU and BI, there is an indirect relationship between PEOU and BI. Finally, REL and OUT had a significant impact on PU, whereas SN and ENJ had a significant impact on PEOU.

The findings of the study have a wide range of implications for practitioners, policymakers, and researchers. Given that ENJ has a significant influence on PEOU, LMS designers must ensure that students enjoy using it. This can be accomplished by ensuring that the LMS interface is consistent with other systems, easy to use, and standards-compliant. Administrators must ensure that students have easy access to course materials and that they can be used without difficulty. Policymakers can also collaborate with internet service providers to keep data costs low and internet access available in every corner of the country.

After determining that a SN predicts PEOU, higher education managers must persuade tutors to advertise the LMSs’ worth to students in order to encourage them to use it. Administrators must also guarantee that course materials available on the LMS are up to date and relevant to students’ academic work, given that REL predicts PU. Again, administrators must ensure that the LMS contains all the learning resources required by students for the course, having determined that OUT is connected to PU. Finally, lecturers must respond to student complaints on time and comment or mark all assignments distributed to students via the LMSs.

Though the study identified some important relationships that could aid in the resolution of issues related to the use of LMS, it is also important to acknowledge some limitations associated with this study. Despite the fact that external variables explained 61.9% of PU and 60.1% of PEOU in TAM3, the study did not include all of the external variables used in TAM3. Future studies could make use of all TAM3 variables as well as other additions. As in the original TAM3, the study failed to address the moderating effect of experience and voluntariness. Future research could look into the moderating effect of these variables. This study’s participants were drawn solely from a public university. When generalizing the findings to include all university students, caution must be exercised. In the future, efforts may be made to reach all public, private, and technical university students in the country.