Introduction

Educational institutions face significant challenges resulting from the dynamic technical, economic, social, and political trends that have intensified during the last two decades, including globalization, technological advancement, digital transformation, and the widespread adoption of information and communication technologies (ICT). Adapting to these new trends includes enabling the smooth transition from traditional learning to distance learning (DL) that aligns with the new educational paradigm (Arsenijević & Mesaroš-Živkov, 2021) and the necessity for the establishment of a knowledge economy (Carayannis et al., 2012; Carayannis & Campbell, 2021; Carayannis & Rakhmatullin, 2014). These trends have particularly impacted higher education as the main generator of professionally qualified and trained staff for the upcoming jobs emerging from Industry 4.0 (Radun, 2020). Besides needing to be a specialist in a certain field and possess the job-specific skills (Bolade & Sindakis, 2020), high-educated graduates need to possess advanced technical skills, which can be acquired faster in distance than traditional learning.

Besides improved adaptability to new global trends, DL also addresses the issue of higher education accessibility in times of social, political, and health crises. It offers many potential benefits, not the least of which is the ability to get beyond traditional learning settings’ time and space limitations (Bates, 2005), evidenced during the COVID pandemic. Like in other countries worldwide, Serbian universities also introduced DL to overcome disruptions in education due to the adoption of anti-pandemic measures. “Lockdown, quarantine, stay home, stay safe” became the new paradigm to reduce the spread of the pandemic (Zhang et al., 2023). During the home confinement caused by COVID-19, universities closed and all teaching became virtual, prompting consideration of the question of the sustainability of the education system. (Faura-Martinez et al., 2021). A large number of faculties utilized the Moodle platform as a proven method of DL even before the pandemic (Eraković & Lazović, 2020) in combination with some web-based video conference tools, including Zoom, Microsoft Teams, and Google Meet.

DL during the pandemic was not a planned and designed procedure in advance, so from the standpoint of researchers in the field of online learning, it would not be accurate to categorize it as an example of distance education. The phrase “emergency remote teaching,” proposed by Hodges et al. (2020), is probably best for this circumstance since it relates to a temporary solution and alternative education delivery method in an online environment. However, the fact that lessons were delivered online and students were educated in this way is essential because it provided a valuable experience that can help higher education institutions learn from it and better prepare for future similar circumstances. Namely, DL should not be simply abandoned after the pandemic, and professors should not return back to their previous practices of teaching exclusively in the classroom without making sure that the lessons learned from 2020 were not lost for upcoming safety and public health crises (Barbour et al., 2020). It can be stated that the COVID-19 pandemic has prompted an investigation into the factors that influence the benefits of DL (Jo, 2023) to prepare efficient DL systems for eventual future crises.

In addition, it should be emphasized that today’s generation comprises digital natives (Al-htaybat et al., 2018), positioning them at the forefront of emerging technologies. Consequently, DL addresses the requirements of younger generations and presents a significant opportunity for aligning with the expectations of modern students (Rajeh et al., 2021). From an economic perspective, adopting advanced DL ensures equitable access to education for everyone. This is achieved by substantially lowering the expenses associated with studying, making it more accessible to individuals residing in smaller towns. In the traditional education model, the expenses related to commuting and lodging in some large cities can pose significant barriers to pursuing education (Stecuła & Wolniak, 2022).

Before the pandemic, the application of technology in accounting classes had been sporadic and primarily depended on the lecturers’ individual preferences. The pandemic era has reversed the burden of demonstrating that increased technological integration in education is required, given the state of the market. Earlier, professors advocated for more technology in the classroom did so by convincing management and colleagues of its benefits. Now, professors who prefer not to do so must explain why (Fogarty, 2020).

Due to the extensive use of DL before the pandemic, developed economies were better prepared than developing ones (Abbas, 2016; Mohammadi, 2015, Ebner et.al., 2020), and research on this issue is more prominent. The scientific literature in developing countries is very modest (Bhuasiri et al., 2012; Van Vu et al., 2022). Intensifying research in developing countries can add more insight to this crucial subject, providing a rewarding topic for additional study and advancements in practice, which was one of the reasons for selecting the Serbian case for this research. Serbia is considered as one of the developing countries with a perspective educational sector. Serbia has one of the best educational systems in the region, and its education sector contributes to the country’s overall economic progress. It is also worth noting that, despite being a developing country, the level of ICT usage in the country is relatively high. According to data from the Statistical Office of the Republic of Serbia (2021), 76.7% of households have a computer, and 81.5% have an internet connection, with 91.7% of them having broadband internet. In this sense, it can be said that this country has favourable conditions for the future development of distance learning.

Many universities adopted open-source learning management systems to supplement their teaching efforts. These systems served as additional platforms for instructing, gathering and presenting educational program data and disseminating pertinent information to students. During the COVID-19 pandemic, classes shifted online, allowing instructors to select their preferred platforms, while universities offered guidelines on suitable digital tools. The institution extended technical assistance, recognizing that most educators had to work from home and utilize their personal resources (as reported by Koruga et al., 2023).

Teachers had the flexibility to choose assessments, such as quizzes, assignments, and exams, that aligned with the course objectives (along with deadlines to ensure students stayed on track). While most faculties required teachers to conduct online classes according to a predetermined schedule at the beginning of the semester, some faculties allowed teachers to decide whether the course would be delivered in real-time (synchronous) or if students could engage at their own pace (asynchronous). However, in both scenarios, they were required to post learning materials. These materials included presentations, recorded class videos, text-based resources, multimedia content, and case studies designed to address practical problems, all by the subject’s specifications. To enhance the quality of content delivery, most teachers implemented mechanisms for obtaining ongoing student feedback in addition to their regular communication efforts. This feedback allowed for continuous improvements to be made during the course.

The focus on accounting education is provoked by some specificities of teaching this economic discipline. First, teaching accounting subjects assumes intensive interaction between the professor and the students due to their specialized nature compared to other economics subjects. The efficient DL in accounting assumes enabling students to work in small groups, be encouraged to participate in discussions with the assistance of a professor, use the learning-by-doing, complete tasks independently, follow recommendations for group work, collaborate on results interpretation, and use simulation and optimization to foster critical thinking (Gainor et al., 2014; Januszewski & Buchalska-Sugajska, 2022). Second, accounting has a considerable practical dimension (Ali, 2020; Rameez et al., 2020). ICT is required to prepare students to adapt to business practice needs and adhere to their professional tasks (Busulwa & Ebans, 2021). So, enhanced ICT implementation in accounting curricula is an important competitive advantage of modern economic faculties. Third, several authors (Arbaugh, 2014; Arbaugh & Rau, 2007) suggested that the effects of DL may differ depending on the academic discipline and whether the course is quantitative or qualitative. So, a more thorough investigation of implications in the accounting field, which features numerous quantitatively oriented courses, seems warranted. Consequently, the specificity of this work compared to previous research is the distinct focus on the consideration of the usefulness of the satisfaction of distance learning in accounting subjects. Especially in the post-pandemic period, research results can be a framework for further improvements in distance learning or digital improvement of traditional learning. Considering that there have been no similar studies in Serbia, especially concerning accounting subjects, this research highlights a completely new dimension of studying accounting disciplines that was not present in Serbia before COVID-19. It emphasizes the advantages, issues, and benefits of DL.

Although there are many studies on DL systems, the contribution of this study is reflected in the advancement of knowledge in the field of DL analysis in the following contexts:

  1. 1.

    Evaluating the key factors impacting the successful adoption of DL in the Republic of Serbia, which could exemplify other developing countries. Like other developing economies, the Republic of Serbia has a low adoption and usage rate for DL, so fostering DL research can boost its adoption in practice.

  2. 2.

    Exploring a previously unexplored link related to the impact of the compatibility of DL with social and pandemic conditions on students’ perceived usefulness and satisfaction. This unique exploration highlights the importance of considering broader socio-environmental factors when evaluating DL outcomes.

  3. 3.

    Emphasizing the importance of DL in accounting education. It highlights the need for graduates to possess advanced technical skills beyond traditional accounting practices. This finding has implications for the evolving role of accountants in the dynamic business environment.

The paper aims to evaluate the impact of the technical qualities of DL platforms, teaching process quality, and compatibility with social and pandemic conditions on students’ perceived satisfaction and usefulness of emergency DL. The remainder of the paper is organized as follows. The next section presents the theoretical background with the development of the hypotheses. The “Data and Methodology” describes the research data and methodology. The fourth section presents the results of the research with appropriate discussion. The final section concludes the paper.

Theoretical Background and Hypothesis Development

Theoretical Background

Prior studies have mostly concentrated on the factors that affect the adoption and use of DL in the absence of extraordinary conditions that could change these practices (Parahoo et al., 2016; Al-Rahmi et al., 2019; Gómez-Ramirez et al., 2019; Yadegaridehkordi et al., 2019). Amidst the sudden transition to online learning prompted by the COVID-19 outbreak, the significance and complexity of various factors influencing perceived satisfaction and usefulness have become more pronounced. Numerous research studies have delved into the realm of distance learning in the context of the global health crisis (Ali, 2020; Alismaiel, 2021; Holzer et al., 2021; Ranadewa et al., 2023; Simamora, 2020; Yandra et. al., 2021).

After the pandemic, the knowledge gap has become even wider, given the lack of consensus among researchers about which factors are the most important in such extraordinary conditions. In developing countries, there is an especially wide spectrum of factors. The most commonly indicated are lack of resources, limited accessibility, infrastructure unavailability, communication issues, and the important role of numerous social factors (Adedoyin & Soykan, 2023; Hazzam & Wilkins, 2023; Islam et al., 2023; Venter et al., 2022).

The geographical distribution of the research population is quite uneven, with Asian countries being the most analyzed ones (Alismaiel et al., 2022; Nordin & Nordin, 2020; Chen et al., 2020; Bao, 2020; Howshigan & Nadesan, 2021; Jiang et al., 2021; Kornpitack & Sawmong, 2022; Selvanathan et al., 2023), followed by African counters (Kaisara & Bwalya, 2021; Ouajdouni et al., 2022; Bossman & Agyei, 2022; Ouajdouni et al., 2022: Ennam, 2024), and European countries (Cofini et al., 2022; Dragomir & Dumitru, 2023; Faura-Martinez et al., 2021; Nikou & Maslov, 2023; Stecuła & Wolniak, 2022). Also, the authors investigated factors that impacted satisfaction, usefulness, and learning outcomes in different areas of studies, like engineering (Pham, 2022), hospitality (Tavitiyaman et al., 2021), business (Alam et al., 2022; Fang et al., 2023), and economics (Dragomir & Dumitru, 2023). In order to take into account specificities of particular subjects, some papers are devoted to analysis of factors impacting learning on particular subjects and one of them is accounting (Alshurafat et al., 2021; Krasodomska et al., 2022; Lazim et al., 2021; Lux et al., 2023; Shabeeb et al., 2022; Terblanche et al., 2023; Tettamanzi et al., 2023; Viviers et al., 2023). However, most of these papers consider subjective factors of students’ US, like performance expectancy, effort expectancy, hedonic motivation, habit, attitude, stress, anxiety, student engagement, influence from family and other important people, and self-efficacy. The impact of these factors is usually evaluated using the technology acceptance model (TAM).

The factors considered in this paper are defined based on the user satisfaction model and the findings of Händel et al. (2020), Adedoyin and Soykan (2020), Terblanche et al. (2023), and Lux et al. (2023), and they are related to TQ, TPQ, and CSPC as external factors. Namely, the unprecedented circumstances of immediate distance teaching and learning witnessed during this crisis underscore the need to investigate how TQ, TPQ, and CSPC impact students’ perceived US in this unique setting. Furthermore, there is a lack of substantial empirical evidence in existing literature to substantiate these links in accounting education in developing countries, particularly within the context of a global pandemic. Establishing this connection empirically can aid educational decision-makers in preparing for potential future disruptions and shed light on strategies to enhance student satisfaction and usefulness, thereby maximizing the benefits of DL during such challenging times.

The mentioned papers dealing with students’ perception of DL employed a variety of methodologies, like paired t-tests and independent t-tests, correlation measures, principal component analysis (PCA), qualitative analysis of interview data, confirmatory factor analysis (CFA), back propagation (BP) neural network model, covariance-based structural equation model (CB-SEM), and partial least squares structural equation model (PLS-SEM).

The PLS-SEM method is the most commonly used methodology in this area, and the following justifications support the use of PLS-SEM in this research. First, the research model exhibits a relatively complex structure featuring several observable latent variables. Second, it is thought that the structural relationships within the model are still in the initial phases of theoretical development or expansion in these new pandemic conditions. Third, this research took advantage of the methodology’s benefits, which included fewer stringent requirements and less restrictive assumptions, that facilitated the creation and estimation of the model without imposing additional constraints (Hair et al., 2019; Hair et al., 2023; Richter et al., 2012; Sarstedt et al., 2021).

Hypothesis Development

Technical Qualities of the DL Platform

The technical quality DL platform encompasses its ability to facilitate diverse learning methods, ensure reliable access to online resources, enable accurate self-assessment via assessment tools, and suit the educational context, notably in accounting education. Several studies have examined these attributes concerning perceived US. Shehzadi et al. (2021) found that high-quality ICT, e-services, and e-information enhance students’ satisfaction with e-learning. So, the satisfaction of students with e-learning and, consequently, the improvement of learning outcomes can be elevated by the quality of interaction and the system’s reliability, including platforms for video meetings (Favale et al., 2020).

Al-Fraihat et al. (2020) and Fogarty (2020) echo findings that the quality of technical systems, information, and support drives perceived usefulness. Fogarty notes that today’s students, accustomed to constant internet use, question the delayed implementation of online learning. DeLone and McLean’s (2003) model identifies two technical elements: information quality, regarding content accuracy and accessibility, and system quality, focusing on performance and user-friendliness. Positive correlations between information quality, perceived satisfaction, and usefulness are observed by Seddon and Kiew (1994), Seddon (1997), Eom et al. (2012), Lwoga (2012), and Alotaibi and Alshahrani (2022), reinforcing the importance of robust technical infrastructure in enhancing online learning experiences.

Several authors, including Seddon and Kiew (1994), Seddon (1997), Po-An Hsieh and Wang (2007), Liaw (2008), and Hassanzadeh et al. (2012), have observed a significant positive correlation between system quality and perceived US in their empirical studies. Alotaibi and Alshahrani (2022) conducted a comprehensive analysis of the impact of both information and system quality on perceived usefulness and satisfaction. They found that a well-structured learning platform with updated content positively influences student satisfaction by enabling efficient task completion. However, they noted that information quality does not affect students’ perceived usefulness, possibly because DL platform usage was mandatory during the pandemic, providing students with alternative avenues, such as direct teacher contact, for precise information. Moreover, they concluded that system quality did not affect perceived usefulness due to the compulsory use of DL platforms, implying that despite quality variations, students benefitted from their usage. According to previous research results, the following is hypothesized:

  • H1. There is a statistically significant positive relationship between the technical qualities of the learning platform and students’ usefulness and satisfaction.

Teaching Process Quality

Teaching process quality in distance learning (DL) encompasses various factors like fostering student communication, effective utilization of ICT by teachers, thorough preparation, easy communication, and enhancing student learning experiences. Interactions between students and teachers, attitudes towards online learning, and course design positively influence perceived satisfaction and usefulness (Barbour et al., 2020; Cidral et al., 2018; Hossain et al., 2019; Liaw & Huang, 2013; Shin & Cheon, 2019; Wei & Chou, 2020). Human resources are pivotal for educational organizations, as teacher productivity relies heavily on their educational and pedagogical capacities. It is crucial during emergencies when creating online lessons is more time-consuming than traditional ones (Almazova et al., 2020; Omebe, 2014). Educators tended to be unprepared for online delivery, and students had reduced access to digital technology and stable and reliable internet. This likely affected low socioeconomic and vulnerable student populations the most (Crawford & Cifuentes-Faura, 2022). This underscores the importance of well-prepared educators and effective communication channels in DL settings, emphasizing the human aspect of facilitating effective online education.

Koruga et al. (2023) and Kyerewaa et al. (2021) emphasized that educators in higher education faced time constraints during the initial stages of the COVID-19 crisis, necessitating creative approaches to online teaching challenges. Prior research stressed the importance of educators focusing on pedagogical aspects of technology, prioritizing effective interaction and communication skills in the online realm (Kyerewaa et al., 2021). The findings underscored the need for educators to enhance pedagogical skills related to technology rather than solely focusing on technical proficiency. Teachers relied on past online teaching experiences to adapt to new working conditions, addressing disparities in digital resource access (Laudari et al., 2021). Success in online education requires educators to actively navigate and understand digital spaces, utilizing various learning tools and platforms effectively.

Careful planning of DL classes involves determining the content and fostering interactions pivotal for student satisfaction and usefulness. Educators in distance education must digitize traditional materials (Tere et al., 2020) and engage in digital communication, notably in providing feedback on student interactions (Yengin et al., 2010). Effective teachers enhance student satisfaction and encourage using electronic learning platforms (Cidral et al., 2018; Lwoga, 2012). Their efforts in planning and facilitating DL classes play a crucial role in student engagement and the overall success of the learning process.

Recognizing learning as a social and cognitive activity, not just a question of passing across information (Barbour et al., 2020), Liaw and Huang (2013) revealed a positive and significant link between interactive learning environments and perceived US. Sun et al. (2008) found a positive and statistically significant relationship between educator quality (timely responses and positive attitudes toward e-learning) and satisfaction. According to the presented theoretical background, the following hypothesis is defined:

  • H2. There is a statistically significant positive relationship between the teaching process quality and students’ usefulness and satisfaction.

Compatibility with Social and Pandemic Conditions

The pandemic brought significant stress to university students, affecting their academic, personal, and relational lives, as well as raising concerns about disease transmission (Zurlo et al., 2020). Lockdown measures shifted interactions to online platforms, leading to feelings of boredom, anxiety, and frustration (Aristovnik et al., 2020; Elmer et al., 2020). Contagion anxiety heightened stress levels among students (Brooks et al., 2020), and personal risk perception influenced responses and mental health (Capone et al., 2020). Students with higher risk perception levels may view DL as a useful tool to mitigate contagion risk, impacting learning outcomes and satisfaction (Baber, 2020). Studies suggest DL environments can be adaptable, sociable, and personalized (González-Gómez et al., 2016; Westermann, 2014), offering flexibility and work-life balance (Herrador-Alcaide et al., 2019). These factors significantly shape students’ perceptions of the US amidst the pandemic’s challenges.

DL, crucial for student safety during COVID-19 (Saxena et al., 2021), addresses the potential cabin fever from extended home stays due to government regulations. Often emotionally sensitive and aware of virus risks, college students tend to adhere to social distancing guidelines (Jo, 2022). In a university context where face-to-face interactions are challenging, students increasingly recognize the significance of DL for both safety and continued education. Namely, students who prioritize social distancing benefit more from distance learning (DL), influenced by their perceived risk and behavioral intentions (Bae & Chang, 2021). Those who view COVID-19 as a serious threat are more likely to adhere to distancing measures. Cabin fever, common among students in quarantine, may exacerbate discomfort and perceptions of COVID-19 risks, impacting attitudes toward social distancing and DL engagement (Jo, 2023). Heightened risk perception influences students’ readiness to embrace DL amidst lockdown measures and social isolation.

Recent studies (Behzadnia & FatahModares, 2020; Cantarero et al., 2021; Martela & Sheldon, 2019) highlight a positive connection between student satisfaction and well-being. Learners’ mental well-being during COVID-19 is crucial for coping and focusing on learning (Ranadewa et al., 2023). Factors influencing DL include a lack of social support and isolation and impact satisfaction (Ghaderizefreh & Hoover, 2018). Student emotions, including anger, anxiety, and enthusiasm, further affect satisfaction and perceived DL usefulness. The asynchronous nature of DL, allowing flexibility in a pandemic with family responsibilities, technological constraints, and limited access, contributes to its utility. Interviews conducted by Zhao et al. (2021) suggest increased satisfaction may stem from a comfortable environment aligning with students’ habits, emphasizing the role of motivation and habits in the relationship between the learning environment, physical conditions, and satisfaction. Alismaiel et al. (2022) point up that the utilization of social media for collaborative learning and student engagement has a direct and positive impact on US.

Considering all factors, it can be asserted that DL ensures equal access for both students and instructors. Its flexibility allows participation from any location and at convenient times, reducing the workload on educators during the COVID-19 era. Educators need not rearrange schedules or provide extra hours, as resources are easily accessible through cloud-based systems like Google Drive (Basir et al., 2023). This adaptability contributes to a more accessible and efficient learning experience for both students and instructors. Thus, the research of previous literature proposed the following hypothesis:

  • H3. There is a statistically significant positive relationship between the compatibility with social and pandemic conditions and students’ usefulness and satisfaction.

The established hypotheses are tested by applying the PLS-SEM methodology to the appropriate data collected by surveying.

Data and Methodology

Research Design

The research utilized a quantitative study method to test and analyze the factors influencing students’ perceived (US) toward (DL) application in accounting during the pandemic. Four variables were considered in this study, comprising the independent variables TQ, TPQ, and CSPC, along with the dependent variable US. The data collected from the questionnaires were processed using the PLS-SEM methodology. The theoretical model served as the foundation for its application.

The PLS-SEM testing yields two models: the outer and inner models. The outer model, also known as the measurement model, assesses the reliability and validity of the indicators associated with variable constructs. Once the reliability and validity tests are successfully completed, the subsequent phase involves the structural analysis, commonly known as the inner model analysis. The inner model represents the structural relationships between constructs, illustrating how they influence each other and serving as the platform for testing hypotheses related to these relationships.

Sampling and Data Collection

The data for the analysis was collected from primary sources via online research (Google Forms) over 6 months, from November 2020 to May 2021. The target group was students from state economics faculties who enrolled or intended to enrol in the accounting, auditing and business finance modules. In line with the research objective, a quantitative research technique was used. Students were selected through stratified random sampling to ensure equal representation of students in terms of gender, age, and year of study. Only students actively engaged in DL during the COVID-19 pandemic were invited to participate in the research.

Additionally, the authors guided their choice of sampling technique based on similar studies conducted worldwide (Alshurafat et al., 2021; Binyamin et al., 2020; De, 2020; Herrador-Alcaide et al., 2019), as well as studies mentioned in the literature review of previous research. It is important to emphasize that this research has an advantage because it was carried out during a period of certain adaptation to the pandemic circumstances (at the beginning of the second and during the third semesters of the pandemic), which is important in such studies (Milheim, 2012). The advantage lies in the fact that students’ perceptions and impressions of adjusting to the new scenario have subsided, and the number of students who have taken part in DL has increased, so the findings are more constant and, thus, more reliable.

The initial questionnaire was designed according to the review of questionnaires from similar studies (Al-Fraihat et al., 2020; Cidral et al., 2018; Herrador-Alcaide et al., 2019), and it contained 45 questions. Since the COVID-19 situation was novel, authors modified already-existing items or created new ones to address the actual circumstances appropriately. The first seven questions focused on the general characteristics of the respondents. The remaining questions were divided into four groups: technological characteristics of the learning platform, teaching process quality, compatibility with social and pandemic conditions, and usefulness and satisfaction. A five-point Likert scale was employed to evaluate students’ opinions about presented statements (ranging from 1, meaning completely disagree, to 5, meaning completely agree). The initial questionnaire was piloted with a small group of students to ensure the reliability and validity of the considered items. After the pilot survey, the final questionnaire was defined by excluding questions that negatively affected the constructs’ reliability and validity. The 22 questions are retained for the empirical research (presented in Appendix in Table 7).

The questionnaire was sent to students through learning platforms. The total number of students who filled out the questionnaire was 373, which gives a response rate of 38.02%, bearing in mind that the number of students registered on learning platforms when conducting the research was 981. The stated research rate is fully in line with previously conducted research in this area, listed in the review of previous research. Table 1 presents the demographic information of the study sample (373 respondents).

Table 1 Descriptive statistics

The descriptive statistic from Table 1 shows that most of the surveyed respondents are University of Kragujevac students (33.8%), and 71.6% are female. As expected, most are aged 19–25 (85.8%), and 87.9% are at the bachelor level of studies. As the most commonly used device for DL, the laptop was used by 61.4% of them, 39.1% of respondents used the Moodle platform, and only 19.6% had experience with DL before the pandemic. At this point, it is important to emphasize that most students who had prior experience with DL before the pandemic primarily utilized the Moodle platform, which had been introduced at certain faculties before the pandemic. Using Moodle alongside traditional classroom learning before the pandemic likely enhanced students’ digital literacy by fostering familiarity with online platforms and developing specific digital skills. However, the differences in digital literacy among students were relatively low, considering that Moodle is used in the vast majority of universities in Serbia.

Theoretical Model Development

The DeLone and McLean information systems success model, the technology acceptance model (TAM), the user satisfaction model, and the e-learning quality models are the four most commonly used types of models for evaluating the success of DL used in previous literature (Sun et al., 2008; Ozkan & Koseler, 2009; Alismaiel et al., 2022). The theoretical model used in this study (Figure 1) is mainly based on the user satisfaction model. This approach suggests that satisfaction is a major indicator of information systems’ success, effectiveness, utilization, and acceptance (Harter & Hert, 1997; Seddon, 1997; Thong & Yap, 1996). The current model also incorporates the DeLone and McLean information systems success model’s constructs of system quality (divided into technical characteristics of the learning platform and the teaching process quality), perceived usefulness, and satisfaction. Technical characteristics include elements like system availability, the usability of system features, and system reliability, while teaching process quality incorporates elements like interactivity and communication components and diversity of learning styles. Besides satisfaction, this model also incorporates the concept of usefulness, which is taken from the TAM model. In Seddon’s (1997) model, usefulness was operationalized as a factor influencing user satisfaction and viewed as a generic perceptual indicator of user benefits. As the positive correlations between satisfaction and usefulness are supported empirically by the existing literature (Arbaugh, 2000; Limayem & Cheung, 2008; Seddon, 1997), they are used as a joint construct in this model. In this study, students’ perceived usefulness and satisfaction encompass the extent to which DL is perceived by students as facilitating more effective learning compared to traditional methods, enabling quick problem-solving and knowledge acquisition, supporting adequate exam preparation, being an attractive educational format in contrast to traditional methods, advocating for its permanent integration into the accounting education process, and generating satisfaction with the achieved results in accounting exams when utilizing DL. Similar to Sun et al.’s (2008) model, the environmental component is introduced to represent social, health, and psychological issues during the pandemic, and it is named compatibility with social and pandemic conditions.

Fig. 1
figure 1

Theoretical model

Methodology

PLS-SEM is a highly utilized approach in multivariate data analysis in social sciences. It is the preferred methodology, especially in cases where the study subject lacks a well-established theoretical foundation, particularly when there is a limited prior understanding of causal relationships (Tilahun et al., 2023). PLS-SEM operates by maximizing the explained variance of the endogenous latent variables, making it especially suitable for exploratory and predictive studies (Manley et al., 2021). Among the most commonly cited advantages of PLS-SEM methodology are also facts that this methodology enables overcoming small population constraints with models that include a large number of items and constructs permitting the use of non-normal data (Anand et al., 2023; Pramudita et al., 2023; Ramli et al., 2019; Yesuf et al., 2023). In this research, this methodology is implemented through the SmartPLS 4 software.

Results and Discussion

The procedure established by Hair Jr et al. (2016) and Sarstedt and Cheah (2019) is applied to access measurement and structural models.

Measurement Model Assessment

The convergent validity and construct reliability of the measurement model are accessed by item loadings (with appropriate p values), Cronbach’s alpha (CA), composite reliability (CR), and average variance extracted (AVE), which are presented in Table 2 (Hair Jr et al., 2016).

Table 2 Convergent validity and construct reliability measures

Item loadings must be greater than 0.6, indicating that structures have incorporated more than 50% of the variance (Benitez et al., 2020; Teng et al., 2021). The results in Table 2 indicate that the loadings for the items were significantly higher than the recommended 0.6 value and statistically significant (p ≤ 0.01). CA and CR values also indicate pleasant internal consistency since they exceed the recommended value of 0.7 for all constructs (Nannally, 1978). Additionally, convergent validity is confirmed by AVE values higher than the suggested threshold of 0.5 (Hair Jr et al., 2016).

Besides convergent validity, the discriminant validity should also be checked before assessing the structural model. The discriminant validity is assessed by the heterotrait-monotrait ratio of correlations (HTMT) ratio (Table 3).

Table 3 HTMT ratio

Table 3 indicates that all HTMT ratio values are lower than the suggested threshold of 0.9 (Franke & Sarstedt, 2019; Henseler et al., 2015), confirming the existence of discriminant validity of the model.

Structural Model Assessment

Since the validity and reliability of the constructs are empirically verified, the structural model can be evaluated. Variance inflation factor (VIF), coefficient of determination (R2), effect size (f2), and predictive relevance (Q2) are used to assess the structural model, as they are the most commonly used indicators in the previous literature (Alami & Idrissi, 2022; Mustofa et al., 2022). The VIF values for considered items are presented in Table 4.

Table 4 VIF values

A VIF value greater than 5 suggests a potential collinearity issue. The derived VIF values from Table 4 are all within the acceptable threshold values (VIF < 5). Thus, collinearity was not an issue with these data.

The next step in structural model assessment is the evaluation of path coefficients, obtained via the bootstrapping procedure in SmartPLS 4 software (5000 bootstrap samples were generated) and presented in Table 5.

Table 5 Results of the PLS-SEM model

The results from Table 5 suggest that all path coefficients are positive, as established hypotheses suggested. However, the path coefficient indicating the relationship between the TQ of the learning platform and the US is low (0.022) and statistically insignificant (p-value = 0.626), suggesting that hypothesis H1 should be rejected. This finding contradicts extensive literature emphasizing technical aspects in enhancing online learning experiences (Al-Fraihat et al., 2020; Alotaibi & Alshahrani, 2022; Delone & McLean, 2003; Eom et al., 2012; Favale et al., 2020; Fogarty, 2020; Lwoga, 2012; Seddon, 1997; Seddon & Kiew, 1994; Shehzadi et al., 2021). While the TQ encompasses features like diverse learning methods and reliable access, the study’s findings suggest these may not strongly influence the US. Despite prior research emphasizing the positive correlation between system quality and perceived usefulness, the current study highlights discrepancies (Hassanzadeh et al., 2012; Liaw, 2008; Po-An Hsieh & Wang, 2007).

Such a result can be explained by the fact that the students were obliged to use DL platforms since there was no other option. So, they used and benefited from it regardless of its qualities. Also, the DL platform may not be necessary for students to obtain reliable information. They can contact the teacher if they require more specific information (Alotaibi & Alshahrani, 2022). Additionally, it should be said that the reason for this result could be that a larger number of students were already familiar with the Moodle platform as an asynchronous platform, which improved their digital literacy to some extent.

Furthermore, these generations spend a lot of time using computers and smartphones, so using new technologies does not affect their satisfaction. Hence, in the digital transformation era, education must evolve to match the digital landscape, with high-quality digital platforms supporting DL and preparing students for the demands of the digital workforce, which is crucial for Serbia’s competitiveness. Serbia should bolster its competitiveness by investing in education and embracing digital advancements, signalling a commitment to innovation and a skilled labour pool. Such a strategy supports lifelong learning, recognizing the necessity for continuous skill development amid a swiftly evolving job market.

The path coefficients for the remaining links are statistically significant, so H2 and H3 are supported. Both of them range from 0.2 to 0.5, which is considered moderate, according to Cohen (1988). The path coefficient representing the link between TPQ and US is higher, amounting to 0.474, indicating the highest positive impact of this factor on students’ US among considered factors. Such a result aligns with insights from the literature review. TPQ, encompassing factors such as communication, effective use of ICT, preparation, and interaction, has been consistently associated with perceived US (Barbour et al., 2020; Cidral et al., 2018; Hossain et al., 2019; Liaw & Huang, 2013; Shin & Cheon, 2019; Wei & Chou, 2020). The significance of well-prepared educators and effective communication channels in DL, particularly during emergencies like the COVID-19 crisis, is underscored (Almazova et al., 2020; Omebe, 2014). The research results are partially consistent with the findings of Alismaiel et al. (2022), which indicate that both collaborative learning and engagement positively impacted peer and instructor interaction; both factors influenced online learning during the COVID-19 pandemic, leading to improved student satisfaction and academic performance. The findings emphasize the human aspect in facilitating effective online education, aligning with the notion that educators, facing time constraints and challenges during crises, must prioritize pedagogical aspects of technology, focusing on interaction and communication skills (Koruga et al., 2023; Kyerewaa et al., 2021).

A slightly lower path coefficient (0.409) is obtained for the link between CSPC and US. However, the positive and statistically significant path coefficient obtained for this link is corroborated by the literature review, highlighting the pandemic’s multifaceted impacts on university students. The COVID-19 pandemic significantly impacted university students, evidenced by heightened stress, anxiety, and risk perception (Aristovnik et al., 2020; Brooks et al., 2020; Capone et al., 2020; Elmer et al., 2020; Zurlo et al., 2020). This prompted a shift towards online platforms like distance learning (DL), known for their adaptability and personalized nature (González-Gómez et al., 2016; Herrador-Alcaide et al., 2019; Westermann, 2014). DL became crucial for safety and continued education amidst challenges with face-to-face interactions (Bae & Chang, 2021; Jo, 2022; Jo, 2023). Martela and Sheldon (2019), Behzadnia and FatahModares (2020), Cantarero et al. (2021), and Ranadewa et al. (2023) also emphasized the positive link between student satisfaction, well-being, and DL engagement, supporting mental health during the pandemic. Also, it points up that online learning during the COVID-19 pandemic mediates the relationship between interactivity and student satisfaction (Alismaiel et al. 2022).

Finally, the obtained R2, f2, and Q2 evaluation should supplement the previous analysis. R2 has been used to determine the explained variance of the latent dependent variables in relation to the overall variance. The cutoff R2 values suggested by Chin (2010) are as follows: 0.190 weak, 0.333 moderate, and 0.670 substantial. The R2 value of 0.681 indicates that the three evaluated constructs explained a substantial percentage (68.1%) of e-learning usefulness and satisfaction. The factors’ effect size (f2) should be evaluated considering that values from 0.02 to 0.149 are considered small, from 0.15 to 0.35 is considered medium, and higher than 0.35 is considered large, according to Cohen (1988). Considering these thresholds, it can be concluded that the relationship between the technical qualities of the learning platform and usefulness and satisfaction has a small effect size (0.01), while a medium effect size is recorded for the remaining two relationships. A larger effect size is obtained for the link between TPQ and US (0.277) than for the link between CSPC and US (0.260). The reason for the medium impact of the TPQ may be that teachers used different platforms for different durations and in different ways to deliver lessons. On the other hand, the strength of the CSPC influence can be clarified by the fact that the research itself was conducted in a period when students were already mostly used to the teaching being conducted online. Digital technologies have been present in the Serbian education system for some years, but the COVID-19 pandemic has intensified the pace of education digitalization and highlighted its potential, possibilities, and risks.

The predictive relevance is another aspect of the structural model that should be evaluated (Q2). According to Fornell and Cha (1994), there is predictive relevance if the cross-validated redundancy value of the construct is greater than 0. The Q2 value for usefulness and satisfaction is 0.672, indicating that the model is sufficiently predictive. To assess the model’s robustness, we employed the non-linearity criteria for robustness validation, as suggested by Sarstedt et al. (2020). This approach is commonly adopted by other researchers, including Ghasemy et al. (2021), Yusfiarto et al. (2022), Werimon et al. (2023), and Li et al. (2023). It is important to recognize that although theory often assumes a linear relationship between constructs, empirical observations reveal that linear relationships may not always hold true. From a statistical perspective, when the relationship between two constructs is nonlinear, the effect’s magnitude does not depend only on the exogenous construct’s change but also on its specific value (Hair et al., 2019). Consequently, this study incorporates a polynomial model by introducing a quadratic effect.

The outcomes of the quadratic effect, as presented in Table 6, indicate that none of the path coefficients representing quadratic effects reach statistical significance. Therefore, we can infer that the lack of significance in these interactions supports the robustness of the linear effect, per the findings of Sarstedt et al. (2020).

Table 6 Results of quadratic effect analysis

Conclusion

The COVID pandemic forced teachers to leave their classrooms and give lectures from home via DL platforms, which was especially challenging for developing countries because it was about emergency distance education applications. In that sense, this research assessed the factors affecting Serbian students’ perceived usefulness and satisfaction with accounting DL during the pandemic. Based on the findings, the TPQ and CSPC had a medium and statistically significant impact on the US, while the impact of technical qualities was rather low and statistically insignificant. This research contributes to the existing literature in several ways. It addresses a critical gap by examining the perceptions of DL among accounting students in Serbian public universities during the COVID-19 pandemic, particularly focusing on the contexts of developing countries. By investigating the US of DL, the study sheds light on the challenges and opportunities students and educators face in transitioning to remote learning environments. This comprehensive approach helps understand the multifaceted nature of DL experiences during the pandemic.

Nevertheless, it can be concluded that faculties in Serbia have mostly readily organized classes in extraordinary conditions. However, the pandemic revealed the importance of skilled and trained teachers for organizing classes in this mode, especially during public health and safety crises when additional support and effort are required. It should be highlighted that just a few higher education institutions in Serbia have applied for accreditation of DL programs and courses, indicating that there is more than enough space for their implementation in the future.

Hopefully, the COVID-19 pandemic will soon be a memory of the past. When it is, universities should not just return to our pre-pandemic teaching and learning practices, abandoning distance instruction. There will most likely be future public health and safety crises. Facilities have been closed in recent years in response to natural disasters such as wildfires, storms, and earthquakes. As a result, teaching in a DL system — in both emergencies and more planned scenarios — must become part of a teacher’s set of qualifications and experience. Learning in such an environment is particularly important for accounting graduates whose spectra of skills are continuously increasing due to dynamic changes in the business environment. They must possess the advanced technical skills necessary to meet the requirements of future employers.

The shifting role of accountants has long been considered to be one of the primary causes of the changes in tertiary accounting education. They should evolve from their historical role as scorekeepers of past business changes to that of designers of the vital management data for directing the organization’s future in a dynamic environment. The ability to work in accounting software alone, which is used to track economic events and produce required reports, is insufficient. The ability to use the software for simulations, optimization, and further advanced analysis is required. Hence, university professors must select the best teaching approaches to generate accounting graduates with appropriate skills.

Theoretical Implications

The theoretical implications of this research are numerous. This first theoretical contribution is the development of a multi-dimensional, comprehensive model for measuring the performance of DL. The model was created after a thorough assessment of the literature, and it is mainly based on the user satisfaction model, which was not sufficiently exploited in previous literature. Second, a previously unexplored link (between compatibility with social and pandemic conditions and perceived usefulness and satisfaction) was examined. Third, this paper is one of the rare papers on DL effectiveness that indicates the greater importance of teaching process quality over the technical qualities of the learning platform. Last but not least, this paper contributes to the modest body of literature on DL effectiveness in developing countries (especially in this part of the world) and accounting education.

Managerial or Policy Implication

The research findings shed light on crucial challenges and recommendations that should be considered to improve perceived US and better utilize the benefits of DL. The study offers practitioners in this area the following practical contributions. First, the study’ findings emphasize the importance of surveying students regularly. As a result, these systems must be continuously improved to solve any issues or limitations. Second, according to the study’s findings, perceived usefulness and satisfaction are substantially affected by the quality of the teaching process. As a result, effective, extensive, and continuous teacher training for using the DL system is a critical success factor. It will help teachers obtain a better understanding and confidence in utilizing the DL system and raise their awareness of all the features of this learning system. Third, universities should motivate teachers to increase the number of online accounting courses and raise students’ awareness about the benefits of attending online courses regularly rather than only in emergencies.

The importance of improving TPQ and TQ, and generally DL development, can be viewed in the context of their broader impact on Serbia’s real economy, telecommunication industry, and digital information sector. There are several ways to articulate this significance. A well-educated workforce is a cornerstone of economic growth, with improved education quality yielding a more skilled and knowledgeable labour force vital for various industries, reducing unemployment rates, and fostering economic stability. The telecommunication industry is pivotal in delivering DL services, prompting increased demand for reliable internet connectivity and telecommunications services as online education surges. A robust digital information sector hinges on a skilled workforce and solid digital infrastructure, with enhancements in TPQ and TQ fuelling sector growth by producing graduates proficient in digital tools. Commitment to high-quality education, including DL, attracts investment and innovation, positioning Serbia favourably in the global marketplace and enhancing its global competitiveness.

Limitations and Ideas for Future Research

Despite the highlighted contributions, this research has certain limitations. First, the proposed model explained 68.1% of perceived usefulness and satisfaction, suggesting that other factors not considered in the model account for approximately 31.9% of the variance in DL perceived satisfaction and usefulness. As a result, there is still potential to research and incorporate more factors influencing perceived satisfaction and usefulness. Second, this study was based on the perceptions of students. Different groups of DL stakeholders, most notably teachers, could enrich the research with their diverse perspectives and provide a deeper knowledge of the challenges in DL delivery. Also, differences in students’ age and diploma types should be considered. These limitations provide researchers with the basis for future research. The new factors impacting perceived usefulness and satisfaction should be incorporated into future models. With ICT and DL practices continuously evolving, comprehensive studies analyzing how the DL quality factors change over time may yield further interesting results. Future research should also be focused on teachers’ perceptions as well. Such analysis should be accompanied by comparing the differences between teachers’ and students’ attitudes. Finally, the current database can be employed to analyze the differences in the impact of considered factors on US of students from different age groups and diploma types.