Virtual reality and perceived learning effectiveness in accounting studies: the mediating role of task-technology fit

Abstract

The use of virtual reality (VR) for the improvement of learning outcomes for accounting studies is an area in which the research is still limited. Technology-based accounting learning is very important because, currently, the accounting profession is very much affected by rapid technological change which means that it must adapt to remain relevant in the business world. This study aims to examine the role of the mediating effect of task-technology fit (TTF) on the relationship between the use of VR and learning outcomes for accounting studies. This research model states that the use of VR will be able to increase TTF, technology quality and accessibility, and then increase reflective thinking and reduce cognitive overload thereby increasing perceived learning effectiveness (PLE). The hypotheses derived from this model were tested empirically using survey data from 199 users of VR engaged in accounting studies. The data analysis uses partial least squares-structural equation modeling (PLS-SEM), and the results support the hypothesis that TTF mediates the relationship between VR use, learning behavior, and PLE.

Keywords:

• Virtual reality
• task-technology fit
• technology quality
• technology accessibility
• reflective thinking
• cognitive overload
• perceived learning effectiveness

REVIEWING EDITOR:

• Collins Ntim
• University of Southampton
• United Kingdom of Great Britain and Northern Ireland

SUBJECTS:

• Accounting
• Accounting Education
• Financial Accounting
• I.T. Teaching

1. Introduction

In the current era known as the industry 4.0, there has been an acceleration of digitalization and information technology in business and government such as e-commerce and e-government. The COVID-19 pandemic also accelerated digitalization in various sectors. In these circumstances, various professions are currently facing the challenge of having to adapt to a very rapidly changing environment, especially in relation to technology. The accounting profession is also affected by technological change, so it must adapt to remain relevant in the future (Jackson et al., 2022; Tavares et al., 2023). One way to adapt is technology-based accounting learning (Tavares et al., 2023).

This current digital transformation in the field of education has seen the development of e-learning which is a more flexible learning method without limitations of time, distance, and space (Umair et al., 2022; Zhang et al., 2017). The challenge for educators is to use technology to support the achievement of high learning outcomes. The educational community has been increasingly concerned about the effectiveness of technologies such as virtual learning environments (VLEs). One of the educational technologies for VLE is virtual reality (VR) (Jena, 2016; Jeon, 2023; Sari et al., 2023; Umair et al., 2022; Zhang et al., 2017). VR is a technology-based learning media that has interactive characteristics and high immersiveness, and it creates a sense of presence which means it can improve learning outcomes (Sari et al., 2023; Umair et al., 2022; Zhang et al., 2017). VR has been widely used in the educational sector because it has three main characteristics which are its advantages, namely, immersion, interaction, and imagination (Umair et al., 2022).

Nowadays, VR has become an integral part of VLE. VR is a useful tool that brings people together from various parts of the world to engage and interact without geographical, social, and economic boundaries. VR technology creates an imaginary world that can transcend the boundaries found in traditional education. Instead of the traditional method of learning that involves listening to a lecture in a classroom, students can absorb words by using a headset and therefore have what feels like an authentic experience in a virtual space. The constructivism and game-based learning that VR provides can improve students’ learning capabilities when compared to the teaching and learning process offered by traditional pedagogy. VR can provide real-time visualization and interaction in a virtual world that can be a way to approach the real world (Chuan et al., 2008; Jeon, 2023). Users can wear three-dimensional (3D) glasses and feel as if they are floating in a projected world in which they can move freely (Laver et al., 2015). VR makes it possible, in the learning process, to visualize and stimulate events that are not perceivable in real life; therefore, users can obtain perceived learning effectiveness (PLE) in the fields of architecture, medicine, computer science, construction, and business (Antonieta, 2014; Green et al., 2014; Griol et al., 2014; Kim & Choo, 2023; Sari et al., 2023; Umair et al., 2022; Zhang et al., 2017).

Previous research has tested the effectiveness of VR in the learning process in various fields such as engineering (Alhalabi, 2016), the military, (Webster, 2016), robotic surgery (Bric et al., 2016; Francis et al., 2020), firefighting (Çakiroglu & Gökoglu, 2019), negotiation training (Ding et al., 2020), health-care training (Chow-White et al., 2017), ethics education (Sholihin et al., 2020), safety training (Pedram et al., 2020), construction design (Umair et al., 2022), business ethics (Sari et al., 2023) and marketing (Kim & Choo, 2023). However, empirical research on the use of VR in accounting studies is still very limited. This study expands previous research by analyzing the use of VR in improving learning outcomes in accounting. Studies on the use of technology in accounting studies are significant because, currently, the accounting profession must adapt to the process of technological evolution and transformation (Tavares et al., 2023). Education and training of individuals capable of adaptation in this era of digital transformation are necessary for the survival and sustainability of the accounting profession in the future (Tavares et al., 2023).

One of the latest technological innovations in the accounting field is the use of VR to help accountants understand company financial reports better and faster (Al-Gnbri, 2022; Chukuwani, 2022; Egiyi, 2022; Zou, 2019). Egiyi (2022) shows that there are several benefits of VR in the accounting field, namely: (1) creation of financial statements, (2) accounting data visualization, (3) accounting education, recruitment and training, (4) virtual communication, networking and customer service, (5) corporate reporting, and (6) auditing. VR will make all accounting activities and preparation of financial reports more automatic in the future. VR will make tasks such as budgeting, invoicing, customer management, inventory, auditing of stock and others more efficient and effective. VR can enable accountants wherever they are to work remotely and can provide users with more efficient experiences such as team meetings, making presentations, training programs, and others. Furthermore, VR will enable accountants to check inventory in real time and with VR, it can reduce costs such as transportation costs because there is no longer a need to go directly to the field when you want to check inventory availability. In accounting learning, VR is used to understand accounting standards and systems in a real-life situation. One example of VR application in accounting learning is Top Education Institute (TOP) in collaboration with PwC Australia, developed to help students understand accounting concepts. The TOP application helps students in the process of learning accounting concepts in an interactive and fun way because they move from a standard classroom situation to a virtual store and apply these accounting principles in a real life.

Although several previous studies have provided empirical evidence that users have more learning experiences with VR, there is still very little research that focuses on the mechanism of how VR can increase the PLE of users (Zhang et al., 2017). Most previous research focuses on user’s intention to use VR based on the technology acceptance model (TAM) theory. There is still very limited existing research analyzing how VR affects learning outcomes, especially PLE. Although several studies have described VR as improving learning outcomes when a task is designed well, few have based it on the task-technology fit (TTF) theory when developing research models.

Zhang et al. (2017) have tested how the use of VR, according to TTF theory, can increase PLE in business analytics courses. They tested the role of the TTF variable as a moderator in the relationship between VR, technology quality and accessibility, reflective thinking, and PLE. However, empirical evidence does not support the hypothesis about the moderating effect of the TTF variable. This study extends the research of Zhang et al. (2017) by basing it on a mediation model, rather than a moderation one. In order to explain the mechanism or process of how independent/exogenous variables (use of VR) can influence dependent/endogenous variables (perceived learning effectiveness), it is more appropriate to use the mediation model than the moderation one (Hair et al., 2022; Hayes, 2022; Hayes & Rockwood, 2020). Apart from that, expansion was also carried out by adding the cognitive overload variable as one of the important learning behavior variables in research on learning (Buchner et al., 2022).

2. Literature review and hypothesis development

2.1. Task-technology fit theory

The task-technology fit (TTF) theory has been used in empirical research to test the usefulness of technology in the work environment and education (Jung et al., 2023). TTF theory explains the level of conformity between the requirements of a person’s task and the ability of technology to support that person’s skills when performing the task (Dishaw et al., 2002). TTF theory is part of contingency theory which explains the usefulness of technologies (Jung et al., 2023). The right technology for a specific task can help individuals or organizations achieve their goals and improve performance. In addition, TTF theory is useful for understanding the technology-to-performance chain, focusing on the match between user task needs and the functionality provided by the technology (Howard & Rose, 2019). TTF theory complements the existing focus on utility in research on acceptance based on models such as the technology acceptance model (TAM) theory, in which the user’s attitude toward technology influences his or her choice (Klopping & McKinney, 2004). However, the user’s skill selection behavior does not always guarantee that task performance will improve (Jung et al., 2023). On the other hand, the use of the TTF model has been shown to improve personal and organizational task performance (Goodhue & Thompson, 1995; Jung et al., 2023). TTF theory shows the importance of decision-making and choosing appropriate information technology for different tasks. TTF theory has been used in previous research to explain the successful selection of information technology in various digital transformation contexts such as online educational content that ensures sustainable use (Wu et al., 2018), Internet of Things (IoT) network systems for disaster management (Sinha et al., 2019), learning about VR technology to increase its effects (Zhang et al., 2017), and decision support systems to improve the quality of decisions (Erskine et al., 2019).

2.2. Virtual reality (VR)

VR is defined as "the sum of the hardware and software systems that seek to perfect an all-inclusive, sensory illusion of being present in another environment" (Biocca & Delaney, 1995; Radianti et al., 2020). VR has three core characteristics, namely interactivity, presence, and immersion (Radianti et al., 2020; Ryan, 2015; Walsh & Pawlowski, 2002). Interactivity is the degree to which a user can modify the VR environment in real time (Steuer, 1995). Presence is considered to be "the subjective experience of being in one place or environment, even when one is physically situated in another" (Witmer & Singer, 1998). According to a technological perspective, the term immersion means “the extent to which the computer displays are capable of delivering an inclusive, extensive, surrounding, and vivid illusion of reality’’ (Slater & Wilbur, 1997). More precisely, this includes the degree to which the physical reality is excluded, the range of sensory modalities, the width of the surrounding environment, and the resolution and accuracy of the display (Slater & Wilbur, 1997). The technological attributes of VR technology—such as the frame rate or the display resolution—consequently determine the degree of immersion that a user experiences (Bowman & McMahan, 2007).

The main advantage of VR is that it can provide collaborative learning by emphasizing immersion, interaction, and communication. Immersion is a feeling of being physically present in a non-physical world. Interaction means that users can see a change in activity on the screen as a result of their input (e.g. a gesture) and respond to a new activity instantly and quickly. Imagination means that a VR environment triggers the human mind’s capacity to imagine in a creative sense, non-existent things. VR has two important features, namely representational fidelity and immediacy of control. Representational fidelity is the degree of realism provided by the rendered 3D images and scenes (Dalgarno et al., 2002). It also refers to the connectedness and continuity of the stimuli experienced (Witmer & Singer 1998). Meanwhile, the immediacy of control refers to the ability to change the view position or direction, giving the impression of smooth movement through the environment, and the ability to pick up, examine, and manipulate objects within the virtual environment (Dalgarno et al., 2002). Previous research has documented that there has been an increase in the use of VR in the field of education and training because it can stimulate interactivity and motivation thereby increasing PLE (Egiyi, 2022; Garris et al. 2002; Ott & Tavella 2009; Radianti et al., 2020; Sari et al., 2023; Zhang et al., 2017).

VR is able to help students explore computer-generated multimedia learning environments in real time. It can provide immersive learning environments in which learners are able to have an experience like real life where there is limited or no access to the real world (Freina & Ott, 2015; Radianti et al., 2020). One of the features of VR is that it’s capable of making learning environments highly immersive. Radianti et al. (2020) defined immersion as "the involvement of a user in a virtual environment during which his or her awareness of time and the real world often becomes disconnected, thus providing a sense of being in the task environment instead" (p. 2). The novelty of VR technology can bring excitement and fun to learning environments (Radianti et al., 2020). Previous studies have found that using new technology such as VR for learning can foster learners’ motivation in comparison to conventional learning materials. Jensen and Konradsen (2018) reviewed the use of immersive VR technologies, particularly for skill acquisition. They focused on immersion and presence and found that VR is useful for training cognitive skills that are related to spatial and visual knowledge, visual scanning, observational skills, psychomotor skills that involve head movement, and affective control of emotional response in stressful or difficult situations (Jensen & Konradsen, 2018). These findings are in line with other research that has also found that learning with VR can improve learning outcomes (Alhalabi, 2016). Zou (2019) argues that VR can help accountants in managing accounting systems and financial data and provide advice to top management for business development. VR is not only on its path to becoming a vital part of all businesses. It also provides a far-reaching and innovative experience to accounting education (Al-Gnbri, 2022; Chukuwani, 2022; Zou, 2019). In the context of accounting learning, Sari et al. (2023) have provided empirical evidence that VR can improve technology quality, technology accessibility, and self-efficacy. This research expands the research of Sari et al. (2023) with a more comprehensive model, namely by adding the variables task technology fit, reflective thinking, cognitive learning and perceived learning effectiveness.

In order to address a specific gap in previous research, this study develops a mediation model to analyze the mechanism or process of how the use of VR can increase perceived learning effectiveness. This study argues that VR technology has representational fidelity and immediacy of control features that are easy to use, thereby increasing technology accessibility. In addition, these two features are appropriate technology so that users perceive them as quality (Dalgarno et al., 2002; Sari et al., 2023; Zhang et al., 2017; Zou, 2019). The use of user-friendly and high-quality VR further increases the match between technological capabilities and the needs for carrying out an assignment or TTF. VR technology has ease of use features (the same as the technology accessibility variable attribute) and usefulness (the technology quality variable attribute) which can increase suitability for the user’s task (Pedram et al., 2020). Furthermore, TTF can improve reflective thinking in the learning process which requires higher-order thinking skills, as it involves students’ capacity to think, react and decide rationally (Jena 2016; Sari et al., 2023; Zhang et al., 2017). In addition, TTF can reduce cognitive overload because the presentation of information in VR is easy to understand but comprehensive (Jena, 2016; Buchner et al., 2022). Increasing reflective thinking and reducing cognitive overload further contribute positively to learning effectiveness.

2.3. Hypothesis development

2.3.1. The influence of VR on technology quality and technology accessibility

VR has features that include technology quality and technology accessibility (Salzman et al., 1999). Technology quality is described as “important,” “relevant,” “useful,” and “valuable,” so it is similar to “perceived usefulness” (Zhang et al., 2017). Meanwhile, technology accessibility is characterized by whether it is “convenient,” “controllable,” and “easy,” so it is the same as perceived ease of use.

VR technology has been widely used in various fields including medicine, engineering, construction, business, education, and entertainment (Radianti et al., 2020; Sari et al., 2023; Zhang et al., 2017; Zou, 2019). The VR system is able to immerse users into a 3D virtual environment where they can interact in real time with virtual objects. With these features, VR technology can improve user performance in various fields (Radianti et al., 2020; Zhang et al., 2017). Shiratuddin and Sulbaran (2006) and Umair et al. (2022) found that VR can help study construction in courses such as engineering and building design. Lawson et al. (2015) also found that VR enables better studies in physics. With two main features of VR, namely representational fidelity and immediacy of control, an appropriate set of learning tasks needs to be designed. With appropriate technology supporting their tasks in VR, learners will find the technology useful and easy to use (Dalgarno et al., 2002; Sari et al. 2023; Zhang et al., 2017). Thus, the use of VR can improve technology quality and technology accessibility. Based on TTF theory, VR technology may best be deployed to support individuals and facilitate the completion of tasks so that users perceive it as having good quality. An experimental study by Chapoulie (2014) focused on two types of VR interface, namely a six degrees of freedom (6DOF) joystick and finger-tracking systems. In the experiment, users found it easy to make manipulations and gestures. Chapoulie (2014) provided empirical evidence that VR is easy to use or meets the criteria for technology accessibility. Sari et al. (2023) have provided empirical evidence that the use of VR has a positive effect on technology quality and technology accessibility. Therefore, we hypothesize:

H1: The use of virtual reality has a positive effect on technology quality.

H2: The use of virtual reality has a positive effect on technology accessibility.

2.3.2. The influence of VR use on task-technology fit

A new kind of information technology such as VR will only be used if its function supports the user’s tasks and can improve learning outcomes. Therefore, it is necessary to design an appropriate set of learning tasks with support for those tasks, such as VR, that users perceive as useful and easy to use (Dalgarno et al., 2002; Radianti et al., 2020; Zhang et al., 2017).

The ability of an information system to support a task can be explained by the TTF model (Strong et al., 2006). TTF implies matching between the capabilities of the technology and the demands of the task (Goodhue & Thompson, 1995). There is an argument that in order for technology to have a positive impact on individual performance, the technology must fit the performance, system characteristics, and task characteristics in a TTF model (McGill&Klobas, 2009). Empirical evidence has shown that VR features have improved task-technology fit (Jung et al., 2023; Radianti et al., 2020; Zhang et al., 2017).

H3: The use of virtual reality has a positive effect on task-technology fit.

2.3.3. The influence of task-technology fit on technology quality and technology accessibility

According to TTF theory, if technology provides a good fit with the tasks, users should perceive the technology as being useful and easy to use to complete the task (Zhang et al., 2017). Zakaria and Daud (2014) argued that TTF is a predictor of technology quality. Previous studies have found that there is a significant positive relationship between TTF and technology quality (Dishaw et al., 2002; Zhang et al., 2017). Other empirical evidence has shown that technology quality is more dependent on the technology’s fit with the task than the workplace environment (Klopping & McKinney, 2004). In addition, Klopping and McKinney (2004) and Zhang et al. (2017) provided empirical evidence that TTF can increase technology accessibility. TTF will influence user beliefs about the consequences of use, namely technology quality and accessibility.

H4: Task-technology fit has a positive effect on technology quality.

H5: Task-technology fit has a positive effect on technology accessibility.

2.3.4. The influence of technology quality, technology accessibility, and task-technology fit on reflective thinking and perceived learning effectiveness

There is an argument that instructional implementation of technology, and not technology itself, determines learning outcomes (Webster & Hackley, 1997; Zhang et al., 2017). When the technology is right for solving learning tasks, it can encourage reflective thinking and reduce cognitive overload thereby increasing PLE. Reflection is an important factor for someone who wants to become an effective lifelong learner and an effective problem solver. Reflective thinking has characteristics including actively monitoring, evaluating, and modifying one’s thinking and comparing it to expert models, peers, and prior experience (Zhang et al., 2017). VR can offer several functions to support reflective thinking. There is a theoretical framework that good quality and easy-to-use technology can improve reflective thinking in four ways (Zhang et al., 2017): (a) process displays, showing learners explicitly what they are doing to solve a task or learn a concept; (b) process prompts, prompting students to explain and evaluate what they did before, during, or after problem-solving acts; (c) process models, focusing on the process that an expert would use in order to think about or solve specific problems; and (d) a forum for reflective social discourse, indicating that reflection can also be a social activity and can be influenced by a community.

The literature states that technology quality and technology accessibility are antecedents of learning behavior such as reflective thinking which in turn has an impact on PLE (Radianti et al., 2020; Sari et al., 2023; Zhang et al., 2017). In using VR, PLE depends on the quality and accessibility of the technology used which is referred to as the model of technology-mediated learning (TML) (Salzman et al., 1999; Sari et al., 2023; Sharda et al. 2004; Wan et al., 2007).

The technology quality and technology accessibility in the VR feature are designed to influence learning behavior such as reflective thinking and cognitive overload. VR’s unique features alone are not enough to facilitate learning, thinking, and understanding. If the technology can support one or more of these four ways (Zhang et al., 2017), then reflective thinking can be improved. VR technology that has good quality and accessibility and is suited to what is needed to complete tasks will be able to improve reflective thinking (Jung et al., 2023; Radianti et al., 2020; Zhang et al., 2017).

The main goal of the learning process is to gain knowledge and increase the capability to take effective action. However, in learning research, it is difficult to measure knowledge and capabilities; indeed, only actions and performance resulting from the learning process can be observed and measured (Alavi & Leidner, 2001). Reflective thinking requires higher-order thinking skills as it involves students’ capacity to think, react, and decide rationally (Zhang et al., 2017). Lee et al. (2010) provided empirical evidence showing that desktop VR can improve user’s reflective thinking and learning outcomes. Virtual learning environments (VLEs) with the use of VR can encourage reflection, accommodate student needs, increase self-confidence, improve readiness to learn, and improve academic performance (Jena, 2016).

Empirical evidence shows that reflective thinking is predictive of PLE if the learning objectives are aligned closely to the assessment tasks (Lee et al., 2010; Makransky & Lilleholt, 2018; Sari et al., 2023; Zhang et al., 2017). Khalid et al. (2016) and Hsieh et al. (2014) provided empirical evidence showing that it is important for students to master reflective thinking skills, because it will help to improve their learning outcomes.

H6: Technology quality has a positive effect on reflective thinking.

H7: Technology accessibility has a positive effect on reflective thinking.

H8: Task-technology fit has a positive effect on reflective thinking.

H9: Reflective thinking has a positive effect on perceived learning effectiveness.

2.3.5. The influence of technology quality, technology accessibility, and task-technology fit on cognitive overload and perceived learning effectiveness

Cognitive overload is a learning load where the amount or method of presenting information exceeds the carrying capacity of a person’s working memory (Buchner et al., 2022). According to TTF theory, it is important to consider cognitive overload when giving learning instructions. This is because the human cognitive architecture consists of a sensory register, a working memory with limited capacity, and a long-term memory with unlimited storage size (Sweller, 1988; Sweller et al., 1998, 2019). When compared to other technologies, VR seems to be less cognitively demanding and also leads to higher performance (Jena, 2016). Pedram et al. (2020) have provided empirical evidence that VR technology has ease of use features and usefulness which can make it suitable for the user’s tasks. These features can then reduce cognitive overload (Makransky et al., 2019). In a systematic review, Buchner et al. (2022) provided empirical evidence that the majority of previous studies report lower cognitive load with higher performance when compared to more traditional conditions such as display-based or paper-based instruction.

H10: Technology quality has a negative effect on cognitive overload.

H11: Technology accessibility has a negative effect on cognitive overload.

H12: Task-technology fit has a negative effect on cognitive overload.

H13: Cognitive overload has a negative effect on perceived learning effectiveness.

Figure 1 presents the research model used in this research:

Figure 1. Research model.

3. Research method

3.1. Data collection

This study used a purposive sampling method to select VR experiment participants, namely accounting students who had studied Islamic accounting as criteria. Based on the criteria, there were 199 students who participated in this study. They were asked to complete an assignment including data extraction and analysis as well as a case study using VR.

The average age of respondents is 20.27 years with a standard deviation of 2.04 years. There were 102 female respondents (51.25%) and 97 male respondents (48.75%). As many as 92.96% of respondents stated that they had no work experience. Based on experience using VR, the majority of respondents, namely 146 people (73.37%) stated that they had never used VR. Only 53 participants or 26.63% had experience using VR in learning. Based on the demographic data, it can be concluded that the characteristics of the respondent sample are relatively homogeneous.

This study uses an Islamic/Sharia accounting learning experiment using virtual reality-based behavioral simulation (VR-BS) technology. The VR-BS method was developed in two stages: first, the development of VR-BS, in the context of sharia financial literacy; second, testing the effectiveness of VR-BS. The method used to develop the VR system was the waterfall stage model as used in previous VR research (Sari et al., 2023), which is described in Figure 2. There were five stages in its development which are described as follows:

Figure 2. Waterfall stages.

Figure 3. PLS-SEM structural model results.

1. Communication

   At this stage, information about the user’s needs was obtained. Those needs were considered when developing the application.

2. Planning

   At this stage, the narratives and scenarios were created in the VR application being developed.

3. Modeling

   The design process was carried out by creating assets that would be used in the VR application.

4. Construction

   The construction process involved arranging the assets into an environment that described events according to the storyline.

5. Deployment

   After these processes, the VR-BS was ready to be given to the users.

Using the waterfall stages method, the VR-BS application has been successfully developed through a series of software engineering processes, starting from the communication definition stage to the implementation stage. The VR software is named Aplikasi Media Edukasi Akuntansi dan Literasi Halal. This application presents a virtual environment for market and accounting simulations, in which there are several assets, such as fruit shops, vegetable shops, toy shops, conventional banks and Islamic banks. Participants use several equipment such as headsets, VR boxes, and VR remotes. Users are required to enter the virtual market to spend money according to sharia guidelines. They spend their money according to their budget by choosing goods and transactions that are in accordance with sharia principles. Next, participants simulate accounting records in journals, ledgers, trial balances and financial reports of these transactions. After using VR, participants are expected to understand the basic concepts of Islamic financial transactions and accounting systems for Sharia entities. All the participants received written and verbal information prior to their participation, explaining the purpose of the study and the experiment. Before interacting with VR, the participants completed a pre-assessment and demographic questionnaire. In the final stage, participants also filled out a post-questionnaire after finishing using VR.

3.2. Measurements

As mentioned previously, VR is defined as “the sum of the hardware and software systems that seek to perfect an all-inclusive, sensory illusion of being present in another environment” (Biocca & Delaney, 1995; Radianti et al., 2020; Zhang et al., 2017). VR measurement is based on instruments developed by Swanson (1987) and Zhang et al. (2017). Technology quality is the extent to which users believe that a particular technology is relevant and useful for their work and will improve their performance outcomes. Technology accessibility is the extent to which users believe that using a particular technology is convenient and controllable and that it will reduce the effort they have to make. The measurements of technology quality and technology accessibility in this research were constructed based on instruments developed by Swanson (1987) and Zhang et al. (2017). Task-technology fit implies matching technological capabilities with the demands of a task (Goodhue & Thompson, 1995). The task technology suitability scale in this research was developed based on the instrument developed by Goodhue and Thompson (1995). Reflective thinking involves actively monitoring, evaluating, and modifying one’s thinking and comparing it using expert models, peers, and previous experience (Webster & Hackley, 1997). Technology must improve user learning behavior, which is a determinant of learning outcomes. The reflective thinking suitability scale in this research is based on the instrument developed by Zhang et al. (2017). Cognitive overload is a learning load where the amount or method of presenting information exceeds the carrying capacity of a person’s working memory. The cognitive-load scale in this study was developed based on the cognitive-load scale proposed by Gerjets et al. (2009) and Leppink et al. (2013). Perceived learning effectiveness (PLE) is a learning result. Sharda et al. (2004) classified learning outcomes into three groups: psychomotor outcomes, cognitive outcomes, and effective outcomes. Effective outcomes include student perceptions of satisfaction, attitudes, and appreciation for the learning experience (Sharda et al., 2004). The PLE scale in this research is based on the instrument developed by Zhang et al. (2017). The questionnaire instrument is presented in the appendix.

3.3. Data analysis

Data analysis in this research uses partial least squares-structural equation modeling (PLS-SEM) method. The software used is Warp PLS 8.0. PLS-SEM is suitable for use as a data analysis method in this study because the research model is relatively complex with several latent variables, namely exogenous, endogenous and mediating (Hair et al., 2022). By using PLS, hypothesis testing results can be obtained simultaneously while minimizing measurement and structural errors (Hair et al., 2022). PLS-SEM analysis was performed with two stages: (1) evaluation of the measurement model or outer model which explains the relationship between latent variables and their indicators; and (2) evaluation of the structural model or inner model which explains the relationship between latent variables/constructs.

4. Results

This section presents the results of the outer model including reliability and validity. Table 1 presents the composite reliability and Cronbach alpha results for all variables. Construct reliability has been met, namely exceeding 0.70 for all variables (Hair et al., 2022). Convergent validity has also been met with average variance extracted (AVE) above 0.50 (Hair et al., 2022; Kock, 2020).

Table 1. Reliability and validity test results.

Variables	Composite reliability	Cronbach alpha	AVE
Virtual reality	0.951	0.939	0.735
Technology quality	0.930	0.905	0.728
Technology accessibility	0.930	0.887	0.815
Task-technology fit	0.937	0.899	0.832
Cognitive overload	0.735	0.712	0.501
Reflective thinking	0.930	0.900	0.769
Perceived learning effectiveness	0.972	0.969	0.728

The convergent validity results are presented in Table 2. Loading factors for all indicators are above 0.70 and significant so they meet the convergent validity criteria (Hair et al., 2022).

Table 2. Loading factors.

		VR	TQ	TA	TTF	CO	RT	PLE	p-value
VR	VR_1	0.766	−0.186	0.226	0.087	−0.037	0.259	−0.268	<.001
	VR_2	0.894	0.049	0.057	−0.037	0.024	0.038	−0.051	<.001
	VR_3	0.880	0.029	−0.030	−0.147	0.065	−0.020	0.189	<.001
	VR_4	0.891	0.030	−0.044	0.125	0.140	−0.161	−0.035	<.001
	VR_5	0.909	0.063	0.059	0.132	0.206	−0.196	0.069	<.001
	VR_6	0.842	−0.122	−0.174	0.031	−0.162	0.118	0.098	<.001
	VR_7	0.809	0.114	−0.081	−0.199	−0.279	0.010	−0.037	<.001
TQ	TQ1	0.027	0.773	−0.378	0.253	0.034	−0.083	−0.178	<.001
	TQ2	0.016	0.861	−0.137	−0.208	0.053	−0.309	0.127	<.001
	TQ3	−0.041	0.910	0.075	−0.025	−0.054	0.133	−0.026	<.001
	TQ4	−0.171	0.905	0.027	0.099	0.008	0.215	−0.119	<.001
	TQ5	0.108	0.706	0.102	0.104	0.035	0.029	−0.234	<.001
	TQ6	0.196	0.808	0.392	−0.104	−0.037	0.019	0.198	<.001
TA	TA1	−0.235	0.237	0.881	−0.169	0.011	−0.274	0.064	<.001
	TA2	0.079	−0.326	0.924	0.114	−0.040	0.002	0.106	<.001
	TA3	0.149	0.102	0.903	0.048	0.030	0.265	−0.171	<.001
TTF	TTF1	−0.075	0.027	−0.077	0.919	0.013	0.195	−0.190	<.001
	TTF2	0.186	−0.282	0.168	0.888	−0.005	0.020	0.191	<.001
	TTF3	−0.103	0.242	−0.084	0.929	−0.009	−0.211	0.005	<.001
CO	CO1	−0.124	0.169	1.178	−0.382	0.633	−0.713	0.255	<.001
	CO2	0.124	−0.169	−1.178	0.382	0.633	0.713	−0.255	<.001
RT	RT1	−0.100	−0.263	−0.008	0.209	0.002	0.854	−0.241	<.001
	RT2	0.122	−0.479	−0.240	0.369	−0.019	0.909	0.029	<.001
	RT3	0.075	0.535	0.265	−0.450	0.027	0.841	−0.088	<.001
	RT4	−0.097	0.232	0.003	−0.150	−0.008	0.902	0.281	<.001
PLE	PLE1	−0.085	0.169	−0.121	0.122	0.067	0.051	0.860	<.001
	PLE2	0.032	0.056	−0.056	0.018	0.046	0.079	0.889	<.001
	PLE3	−0.019	0.024	0.306	−0.136	−0.056	−0.254	0.908	<.001
	PLE4	0.064	−0.142	0.091	−0.205	−0.030	0.212	0.895	<.001
	PLE5	0.037	0.147	0.147	−0.196	0.010	0.026	0.897	<.001
	PLE6	−0.110	0.198	0.189	0.032	−0.067	−0.286	0.874	<.001
	PLE7	0.033	0.297	−0.385	0.084	−0.077	0.089	0.818	<.001
	PLE8	0.153	0.004	−0.328	0.198	−0.048	−0.060	0.825	<.001
	PLE9	−0.188	0.768	−0.297	0.280	0.219	−0.572	0.709	<.001
	PLE10	−0.050	−0.211	0.096	−0.098	0.008	−0.029	0.867	<.001
	PLE11	−0.072	−0.305	−0.017	−0.087	0.052	0.459	0.849	<.001
	PLE12	−0.044	−0.322	0.018	0.061	−0.073	−0.014	0.860	<.001
	PLE13	0.230	−0.567	0.255	0.010	−0.016	0.224	0.822	<.001

Table 3 presents the results of discriminant validity by comparing the square root of AVE with the correlation between constructs (Hair et al., 2022). The square root of AVE in the diagonal column exceeds the correlation between constructs so that the discriminant validity criteria have been met.

Table 3. Discriminant validity test results.

	VR	TQ	TA	TTF	CO	RT	PLE
VR	0.857	0.688	0.603	0.622	−0.210	0.640	0.659
TQ	0.688	0.853	0.773	0.809	−0.353	0.851	0.744
TA	0.603	0.773	0.903	0.792	−0.426	0.811	0.799
TTF	0.622	0.809	0.792	0.912	−0.400	0.798	0.753
CO	−0.210	−0.353	−0.426	−0.400	0.633	−0.347	−0.393
RT	0.640	0.851	0.811	0.798	−0.347	0.877	0.811
PLE	0.659	0.744	0.799	0.753	−0.393	0.811	0.853

Table 4 presents the results of the goodness of fit analysis which shows that all indicator criteria have been met. Thus, it can be concluded that the model fits the empirical data. Additionally, the results analyze Stone-Geisser’s Q² to measure the model’s out-of-sample predictive power or relevance. The results in Table 4 show that all Q² values are larger than zero so that they indicate predictive relevance in terms of the effectiveness of VR in improving learning effectiveness.

Table 4. Model fit indices.

Fit indicators	Results	p-values	Criteria
Average Path Coefficient (APC)	0.368	P < .001	P < .05
Average R-Square (ARS)	0.569	P < .001	P < .05
Average Adjusted R-Square (AARS)	0.565	P < .001	P < .05
Average Full Collinearity VIF (AFVIF)	3.639		≤ 5.0
Tenenhaus GoF (GoF)	0.638		≥ 0.10
Statistical suppression ratio (SSR)	1.000		≥ 0.70
Q-squared (Q^2):
Technology quality	0.715		> 0.00
Technology accessibility	0.658
Task technology fit	0.394
Cognitive overload	0.209
Reflective thinking	0.803
Perceived leaning effectiveness	0.697

Figure 3 presents the results of testing the PLS-SEM structural model.

Note: * significant at 0.05 alpha.

Table 5 presents a summary of the results of the hypothesis testing in the form of path coefficient and significance. As shown in Table 5, aside from H10, all the hypotheses were supported. The use of VR has a positive effect on technology quality, technology accessibility, and task-technology fit respectively with standardized coefficients of 0.298, 0.196 and 0.628 (significant at alpha 0.01) meaning that H1, H2, and H3 are supported. These findings support previous study that the two main features of VR, namely representational fidelity and immediacy of control, can improve technology quality and technology accessibility (Chapoulie, 2014; Zhang et al., 2017). This study’s empirical evidence also supports previous research that showed that VR features can improve task-technology fit (Jung et al., 2023; Radianti et al., 2020; Zhang et al., 2017).

Table 5. Path coefficients and p-values results.

Hypothesized paths	Coefficient	p-value	Effect size	Conclusion
VR → TQ	0.298	<.001	0.206	H1 is supported
VR → TA	0.196	.002	0.119	H2 is supported
VR → TTF	0.628	<.001	0.394	H3 is supported
TTF → TQ	0.628	<.001	0.512	H4 is supported
TTF → TA	0.677	<.001	0.539	H5 is supported
TQ → RT	0.494	<.001	0.424	H6 is supported
TA → RT	0.298	<.001	0.244	H7 is supported
TTF → RT	0.170	.007	0.136	H8 is supported
RT → PLE	0.763	<.001	0.627	H9 is supported
TQ → CO	0.081	.124	0.031	H10 is not supported
TA → CO	−0.294	<.001	0.130	H11 is supported
TTF → CO	−0.113	.053	0.047	H12 is supported
CO → PLE	−0.148	.016	0.068	H13 is supported

The PLS-SEM results in Table 5 also show that task-technology fit has a positive effect on technology quality and technology accessibility (coefficient: 0.628 and it is significant at alpha 5%). The empirical evidence from this study supports the TTF theory that posits that, if a technology such as VR provides a good fit with the tasks, users will perceive that the technology is useful meaning that it has quality technology attributes and is easy to use.

The test results show that technology quality, technology accessibility, and task-technology fit in using VR can increase reflective thinking (coefficients of 0.494, 0.298, and 0.170 respectively and they are statistically significant) meaning that H6, H7, and H8 are supported. Reflective thinking can increase PLE as indicated by a path coefficient of 0.763 and it is significant at alpha 0.01 meaning that H9 is supported. This empirical evidence is consistent with the findings of Zhang et al. (2017) and Makransky and Lilleholt (2018).

The test results also support the TTF theory that technology accessibility and TTF can reduce cognitive overload; this is indicated by coefficients of −0.294 and −0.113. However, H10, which states that technology quality can reduce cognitive overload, is not supported by a coefficient of 0.081 and is not statistically significant. Finally, the PLS-SEM test results show that cognitive overload has a negative effect on PLE, indicated by a coefficient of -0.148 and it is significant. This result is different from Makransky et al. (2019) which shows that the use of VR can actually increase cognitive overload thereby reducing learning outcomes. Overall, the results of this study demonstrate the importance of reducing cognitive overload to improve learning outcomes.

This study provides empirical evidence supporting 12 of the 13 hypotheses. There is one non-significant finding, namely the hypothesis (H10) regarding the effect of technology quality on cognitive overload which is not supported. The effect of effect of technology quality on cognitive overload is not statistically significant. There are various possible causes for non-significant findings, including participants using VR for the first time in learning. Furthermore, in a systematic review, Buchner et al. (2022) shows that the results of previous research provide conflicting empirical evidence about the influence of the use of technology in education (such as VR) on cognitive overload.

The results of testing the TTF mediation effect hypothesis are presented in Table 6 in the form of indirect and direct effects. As shown in Table 6, the results of the indirect effect test show the significance of the mediating effect of the TTF variable except for the number 3 structural path. These results show support for the TTF theory regarding the mechanism or process of how the use of VR can improve learning outcomes. This study provides empirical evidence that shows that the use of VR can increase task-technology fit which in turn increases user perceptions of technology quality and accessibility. Furthermore, there is an increase in the reflective thinking of its users and a decrease in cognitive overload, thereby increasing PLE.

Table 6. Results of testing the mediating effect of TTF.

Structural paths	Coefficient	p-value	Conclusion
Panel A: Indirect effect
1. VR → TTF → CO → PLE	−0.148	.016	Significant
2. VR → TTF → RT → PLE	0.512	<.001	Significant
3. VR → TTF → TQ → CO → PLE	−0.093	.067	Not Significant
4. VR → TTF → TQ → RT→ PLE	0.322	<.001	Significant
5. VR → TTF → TA → CO → PLE	−0.104	<.043	Significant
6. VR → TTF → TA → RT→ PLE	0.312	<.001	Significant
Panel B: Direct effect
VR → PLE	0.247	<.001	Partial mediation

The test results show that the direct effect of using VR on PLE is significant. This shows that the TTF mediation effect is partial mediation which occurs when the indirect effect and direct effect are both significant (Hair et al., 2022; Nitzl, 2016). Partial mediation implies that there are mediating variables other than those being examined by this study.

The results of this study provide practical significance that can be used in real-life settings. VR can be used as an innovation in accounting learning because it can increase TTF and reflective thinking and reduce cognitive overload thereby increasing the effectiveness of accounting learning. This technology-based learning is important for prospective accountants to remain competitive and relevant in the business profession (Al-Gnbri, 2022; Chukuwani, 2022; Tavares et al., 2023).

The use of VR can improve and transform traditional accounting education in the learning process, namely by students switching from textbook-based classroom situations to virtual financial transactions and applying these accounting principles in a real life. The results of this study show that students can learn financial literacy and accounting systems in an interactive and fun way and have a picture of the real-life situation from the occurrence of transactions to the process of preparing financial reports. The results of this study provide empirical evidence that VR technology can improve learning outcomes in the accounting field. VR-based accounting learning can provide immersion, interactivity and realism compared to traditional accounting learning.

5. Conclusion

The results of this study provide empirical evidence that supports the TTF theory, namely that the use of VR can provide a level of conformity between the requirements of a person’s task and the ability of technology to support that person’s skills when performing the task (Zhang et al., 2017). TTF can increase perceived usefulness and perceived ease of use so that it prompts the user’s reflective thinking and reduces cognitive overload which then results in PLE. The empirical evidence demonstrates the mediating role of the TTF variable in the relationship between VR use, behavioral learning, and learning outcomes.

The main contribution of this study is by expanding previous VR research, including Zhang et al. (2017) and Sari et al. (2023) by providing empirical evidence of the mechanism or process of how the use of VR can increase PLE using a mediation model. Apart from that, this study also contributes by providing empirical evidence about the role of learning using VR in reducing cognitive overload within the framework of the TTF theory. This study also contributes by analyzing technology-based learning in the accounting field which is still very limited in previous VR research. The empirical research on technology-based accounting learning makes an important contribution because the accounting profession is currently affected by technological change meaning that it must adapt in order to remain relevant in its future work (Jackson et al., 2022; Tavares et al., 2023).

There are several limitations to this study. Research on the effect of using VR in the context of accounting learning is still relatively new, so this study should be considered to be exploratory in nature. Second, the test results show a form of partial mediation, so there are other mediating variables that have not been included in this research model. Future research could consider satisfaction, technology readiness, and attitude as potential mediating variables, which have not been included in this study, in the relationship between VR use and PLE (Jena, 2016; Hadi et al., 2021; Jung et al., 2023). Despite its limitations, this study has sought to address the lack of empirical research on the use of VR in accounting and provide empirical evidence about the role of TTF as a mediating variable in the relationship between VR use and PLE.

Author contributions

Conceptualization: DR, RCS, SW; Data collection: DR, MU, LMW; Formal analysis: DR, RCS, SW; Methodology: DR, RCS, SW; Project administration: DR, MU, LMW; Software: DR, RCS, MU, LMW; Validation: DR, MU, LMW; Writing – original draft: DR, RCS, SW, MU, LMW; Writing – review & editing: DR, RCS, SW; All authors agree to be accountable for all aspects of the work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Additional information

Funding

This study was funded by Indonesian Collaborative Research (Riset Kolaborasi Indonesia), with Activity Implementation Assignment Letter (Surat Penugasan Pelaksanaan Kegiatan) Universitas Diponegoro, Number 391-15/UN7.D2/PP/V/2023 dated 30 May 2023. The author would like to thank for this funding.

Notes on contributors

Dwi Ratmono

Dwi Ratmono is a lecturer at the Accounting Department, Universitas Diponegoro, Indonesia. He has published papers in international journals with research interests in accounting education, accounting systems and digitalization, occupational fraud, sustainability and corporate governance. Currently, he is the Head of the Accountant Professional Education Study Program, Universitas Diponegoro.

Ratna Candra Sari

Ratna Candra Sari is a lecturer at the Faculty of Economics and Business, Yogyakarta State University, Indonesia. Apart from being active in writing books and publishing in reputable international journals, she is also active in educational activities on financial literacy. Virtual reality- based educational media have been widely produced and obtained patents.

Sony Warsono

Sony Warsono is a lecturer at the Accounting Department, Gadjah Mada University, Indonesia. He has published various papers in international journals with expertise in accounting information systems, accounting education, information technology systems and business ethics. Currently, he is the Head of the Accountant Professional Education Study Program, Universitas Gadjah Mada.

Muhammad Ubaidillah

Muhammad Ubaidillah is a student of Professional Accountant Education, Faculty of Economics and Business, Universitas Diponegoro. He is also a lecturer at the Vocational School, Universitas Diponegoro with expertise in accounting education, class instruction, financial reporting and Shariah accounting.

Luqman Mulki Wibowo

Luqman Mulki Wibowo is a student of Professional Accountant Education, Faculty of Economics and Business, Diponegoro University. He has expertise as a consultant on accounting information systems, financial reporting and taxation.

Appendix.

Questionnaire

Table

Variables	Indicators
Virtual reality	The realism of the 3D images motivates me to learn. The smooth changes of images make learning more motivating and learning. The realism of the 3D images helps to enhance my understanding. The ability to change the view position of the 3D objects allows me to learn better. The ability to change the view position of the 3D objects makes learning more motivating and interesting. The ability to manipulate the objects (e.g: pick up, cut, change the size) within the virtual environment makes learning more motivating and interesting. The ability to manipulate the objects in real time helps to enhance my understanding.
Technology quality	Using the virtual reality-based e-learning system will enhance my perceived learning effectiveness. Using the virtual reality-based e-learning system will give me greater control over learning. I will find the virtual reality-based e-learning system to be useful in my learning. Using the virtual reality-based e-learning system will enable me to improve my performance in study. Using the virtual reality-based e-learning system will enable me to increase my productivity in study. Using the virtual reality-based e-learning system will make it easier to study.
Technology accessibility	Interacting with the virtual reality-based e-learning system will not require a lot of my mental effort. I will find the virtual reality-based e-learning system to be ease to use. My interaction with the virtual reality-based e-learning system will be clear and understandable.
Task-technology fit	I think that using e-learning with virtual reality will be well suited for the way I like to improve my achievement. E-learning with virtual reality will be a good medium to provide the way I like to improve my achievement. I think that using E-learning with virtual reality will be a good way to improve my achievement.
Reflective thinking	I will be able to reflect on how I learn. I will be able to link new knowledge with my previous knowledge and experiences when learning through virtual reality. I will be able to become a better learner when learning through virtual reality. I will be able to reflect on my own understanding when learning through virtual reality.
Cognitive overload	Using VR is boring Interacting with VR-based teaching doesn’t require much mental effort on my part
Perceived learning effectiveness	I will be more interested to learn the topics. I will learn a lot of factual information in the topic. I will gain a good understanding of the basic concepts of the materials. I will learn to identify the main and important issues of the topics. I will be interested and stimulated to learn more. I will be able to summarize and conclude what I learn. The learning activities will be meaningful. What I learn, I can apply in real context. With VR, this activity really increased my understanding of the formulas discussed With VR, instructions and/or explanations during activities, are very unclear With VR, instructions and/or explanations in learning are very ineffective With VR, this activity really increased my understanding of the topics covered With VR, this activity really increased my understanding of concepts and definitions