Enhancing Project-Based Learning in Under-Resourced Ghanaian Senior High Schools Through a Pedagogical Agent Persona

Abstract

This study explores the design and evaluation of a pedagogical agent persona (PAP) to enhance project-based learning (PBL) in Ghanaian senior high schools. Using a design-based research approach, a multimodal human-like PAP was developed and tested with 15 students and two teachers. Data from interviews, observations and usability scales revealed high learner engagement and satisfaction, with the Learner Attenuated System Paced model scoring highest (SUS = 92). Findings suggest that a PAP can support instructional equity, autonomy, and engagement in large-class settings. The study offers practical insights for scalable learner-centred technology integration in under-resourced educational contexts.

Keywords

pedagogical agents project-based learning pacing pedagogical agent persona usability instructional pacing educational technology Ghana equity in education

Project-based learning (PBL) has gained increasing attention in education as an instructional approach that fosters student engagement, critical thinking, and collaboration by immersing learners in authentic real-world tasks (Song et al., 2025). In contrast to traditional teacher-centred methods, PBL shifts the responsibility of learning to students by positioning them as active investigators and creators while the teacher serves as a facilitator (Beem, 2021; Ramamurthy et al., 2023).

One emerging technology with strong potential to enhance PBL is the Pedagogical Agent Persona (PAP). A PAP is a type of pedagogical agent which is an interactive computer-generated character designed to deliver guidance, feedback and scaffolding in a human-like manner during the learning process (Lin et al., 2020; Pérez-Marín, 2021; Schroeder et al., 2025). Pedagogical Agent Personas (PAPs) may appear as animated avatars, virtual tutors, or character-based guides, capable of simulating social cues such as gaze, facial expressions and gestures. For example, a PAP in a visual arts class might demonstrate brush techniques through animated movements, provide instant feedback on a student’s project, and encourage reflection through prompts. By embedding such support directly into multimedia learning environments, PAPs can personalize the PBL experience and extend the teacher’s capacity in large or resource-constrained classrooms (Li et al., 2022; Ryong et al., 2023).

In Ghana, the context for this study, class sizes are typically large; thus, making it difficult for teachers to monitor each student’s progress in project-based learning (Aidoo, 2023; Essel et al., 2022, 2024). Recent national education statistics compiled by the Ghana Statistical Service (2023), drawing on EMIS administrative data, indicate that senior high school class sizes in Ghana often average 50–60 learners per classroom, with some urban senior high schools exceeding this figure. This contrasts with the sub-Saharan African regional average of about 44 students per secondary-level class and a global average of roughly 26 (UNESCO Institute for Statistics UIS, 2024). These figures illustrate the structural challenge of overcrowded classrooms in Ghana and highlight the pedagogical value of scalable digital supports such as pedagogical agent personas, which can extend teacher presence and promote equitable participation in large-class contexts. Additionally, most teachers in Ghana are used to traditional teaching methods where students are passive learners (Boakye et al., 2024). This is in contrast to project-based learning, which puts students in charge of their own learning and requires teachers to act as facilitators rather than knowledge holders and controllers (Beem, 2021; Ramamurthy et al., 2023).

Pedagogical practices in most senior high schools in Ghana are currently teacher-centred, with teachers acting as ‘knowledge disseminators’ (Boakye et al., 2024; Douglass, 2020; Woods, 2020). This method of teaching does not enable students to develop the necessary skills to function effectively in a modern society. In any case, senior high school students in Ghana are examined at the end of their 3-year programme by the West African Examinations Council (WAEC) in both the core subjects and the electives with visual art students always having a project work component in their electives (Annan-Brew, 2020; Appah et al., 2021).

PAP technology offers an alternative to the current teacher-centred practices (Mathews, 2021). This technology involves a new class of learning technologies known as artificial intelligent-instructional systems, which incorporate simulations, expert systems, fuzzy logic, and intelligent agents. The agents appear as pedagogical characters in a multimedia package and can provide guidance, feedback, and tutoring to students as they work through the courseware (Mathews, 2021; Ryong et al., 2023; Wu et al., 2021). With the rising popularity of e-learning, a human-like pedagogical agent (PA) that shows eye gazing and gestures is frequently used in online multimedia lessons to boost social connection and improve learning in e-learning settings (Li et al., 2022). Further, the PAP has attained popularity in recent years as a technology-supported instructional innovation that utilises the classroom period for students to actively engage in interactive learning tasks, which incorporate scaffolding and individualised feedback from the agent. As such, the importance of designing a pedagogical agent that can suit a PBL environment is very crucial, as the research conducted by some scholars also showed that most of the higher education students preferred constructive type of practices such as using technology applications in their work (Elfeky & Elbyaly, 2023; Žerovnik & Nančovska Šerbec, 2021). Pedagogical agents can demonstrate human-like intelligent behaviour such as reasoning, problem-solving and understanding natural language (Kadam & Vaidya, 2021; Laureano-Cruces et al., 2022). These digital characters are capable of providing real-time, contextualized, and individualized tutoring to students. Research has consistently shown that the use of pedagogical agents leads to significant improvement in retention, engagement, and the overall learning experience (Lin et al., 2020; Li et al., 2022; Wang et al., 2022).

Despite the abundance of perspectives and computational models in the literature on agents and multi-agent systems, little is known about how pedagogical agents are designed and developed to enhance project-based learning in different pacing conditions, specifically at the Senior High School level. This reality has prompted many scholars to emphasize a dearth of understanding on successful pedagogical agent design, particularly in the literature on pedagogical agents for project-based learning (Boss & Krauss, 2022; Kumar, 2021; Pan et al., 2021) and that is a gap in the literature that this research study aims to explore.

Hence, this study explores the design, development, and evaluation of a PAP tailored to different pacing conditions: System Paced, Learner Paced, and Learner Attenuated System Paced, to enhance PBL in drawing among students in Kumasi Senior High Technical School. By integrating human-like instructional support into visual arts project work, the PAP aims to increase learner autonomy, engagement, and skill mastery while ensuring equitable access to quality guidance regardless of class size.

Literature Review

The Role of Pedagogical Agent Personas in Education

In educational technology, Pedagogical Agent Personas (PAPs) are often presented as animated, human-like characters, though alternative designs can range from animals to abstract forms. Such agents may communicate with humans using a variety of modalities including speech, gesture, and facial expressions (Alisa et al., 2020; Pérez-Marín, 2021; Weber et al., 2021). In the field of technology in education, PAPs are also called ‘learning companions’ because they can accompany students in the learning process, like a real teacher or tutor.

Students can be engaged in active and reflective learning with the help of such agents (Grivokostopoulou et al., 2020; Lin et al., 2020). Sustained interaction between learners and PAPs fosters familiarity and rapport, which scholars emphasize as central to learning engagement. They see PAPs as ‘situated and embodied’ in their research. (Lin et al., 2020; McCallen & Johnson, 2020; Schroeder et al., 2025).

Building on this, Alisa et al. (2020) define PAPs as characters that actively facilitate learning, while Veletsianos & Russell (2013) distinguishes between teachable agents and conversational agents. While teachable agents are fictional characters whom students can train to perform tasks, conversational agents are capable of carrying on dialogues with learners. In his discussion of teachable agents, Blair (2007) highlights their shared visual representations and capacity to serve as an example for effective learner behaviour. A thorough analysis of the design elements and pedagogical agents’ effects is given by Martha & Santoso (2019), who also identifies the many types of design and their beneficial effects on student behaviour and learning.

The educational study of PAPs gained momentum in the late 1990s (Clarebout et al., 2002), marking a shift from earlier research that had focused more on technological than on pedagogical considerations. According to Martha & Santoso (2019), each passing year is different in how pedagogical agents grow. Researchers asserted that pedagogical agents served as mentors, motivators, facilitators, navigators, and collaborative assistants in the educational process between 2007 and 2017. Research on pedagogical agents has up to this point focused on the agents’ appearance, their verbal and nonverbal interactions with students, and how they give feedback to enhance student performance, expand learning opportunities, inspire students, and boost engagement. More realistic and human-like educational agents can be made due to developments in human-computer interface technology.

Based on their first work from 2005, Kim & Baylor (2016) examined the state of the research on the functions and capabilities of pedagogical agents. They discovered that students need two things from an agent: the capacity to motivate them (the agent should be polite and friendly) and outstanding teaching abilities (the agent should be knowledgeable).

According to one of the principles of pedagogical agent designs of Moreno & Mayer (2007), known as the principle of modalities, the design of pedagogical agents has five forms: text, voice, 2-D character, 3-D character, and human (Clarebout & Heidig, 2012; Moreno & Mayer, 2007). In a systematic review by Martha & Santoso (2019), the study revealed that spoken text is a more effective means of communication for educational agents than text or voice alone. According to the study, text is the form that is most frequently used (72%), with 31 publications using text together with other forms and five papers using text alone. 3-D characters are the second most often utilized format (52%). Schroeder & Adesope’s (2013) review suggests that a 3-D agent form that is more realistic can be more accurate and efficient than a 2-D agent form. Two studies employ exclusively 3-D characters, whereas 24 use 3-D characters in conjunction with other forms as an instructional tool.

The PAP approach from previous studies has several advantages, such as better opportunities for students to interact with learning materials, more positive viewpoints, greater motivation to learn science, higher levels of student satisfaction, and higher scores in science examination for high school students. A lot of studies, however, have focused on high school students studying other subjects like science, chemistry, and mathematics (Piedrahita-Ospina & Moreno-Cadavid, 2019) rather than studio-based visual arts courses like PBL in picture making. Thus, additional investigation of the design and development of a PAP and its pacing influence on picture-making students’ knowledge and skill is necessary for the context of project-based learning in visual art education.

Potential of Pedagogical Agent Persona in Providing Quality PBL

PAPs offer several potential benefits for Project Based Learning. These include scaffolding learning, enhancing engagement, promoting self-directed learning and addressing large class sizes. Furthermore, the United Nations Sustainable Development goal 4 identifies Project-Based Learning (PBL) as a valuable tool for promoting the United Nations Sustainable Development Goal 4 (SDG 4) on Quality Education, as it encourages active and inclusive learning (Pérez-Sánchez & Sanchis, 2021). SDG 4 aims at ensuring inclusive and equitable quality education and promoting lifelong learning opportunities for all. By fostering critical thinking, creativity, and collaboration, Project-Based Learning (PBL) directly contributes to sub-targets such as 4.1 (universal free and equitable education), 4.4 (relevant skills for employment), and 4.7 (education for sustainable development). Research shows that PBL enhances creative thinking and real-world problem-solving skills by aligning with the competencies emphasized in SDG 4 (Prahani et al., 2025). Moreover, it facilitates the development of transversal competencies and sustainability awareness in engineering and vocational education (Llopis-Albert et al., 2024). Therefore, in integrating PBL-aligned assessment practices, PAPs not only represent instructional innovation but also advances global development goals (Hazard, 2023). As such, a PAP can provide additional support and guidance to students throughout the project by personalizing the learning experience which also helps students to develop relevant skills such as critical thinking and problem-solving for employment (Lluch et al., 2020).

PAPs can provide personalized guidance and support throughout the project stages by acting as a virtual coach or mentor thereby providing scaffolding in learning (Bundsgaard, 2018; Jin, 2010). Bundsgaard (2018) and Jin (2010) highlight the significance of technology in delivering personalized guidance and support in project-based learning. Both Bundsgaard’s Practice Scaffolding Interactive Platform (PracSIP) and Jin’s pre-programming analysis tutors employed technologies to scaffold learning. PracSIP simulates a professional practice, while Jin’s tutors offer online cognitive support. PAPs can also increase student engagement with PBL through interactive features, humour, and storytelling thereby enhancing engagement. PAPs can also guide students in setting goals, monitoring progress, and reflecting on their learning. As such, PAPs promote self-directed learning. Moreover, PAPs can offer individual support to students by alleviating the strain on teachers in large classrooms (Reategui et al., 2012). Schroeder & Adesope (2015) highlights the potential for pedagogical agents to facilitate learning in various subjects, while Reategui et al. (2012) demonstrates their ability to improve the efficiency of instructional conditions. These findings collectively suggest that PAPs can offer individual support to students and potentially alleviate the strain on teachers in large classrooms.

Method

Research Design

This study employed a design-based research approach that combined the Design Science Research Process Model (DSRPM) by Vaishnavi & Kuechler (2008) with the Analysis, Design, Development, Implementation, and Evaluation (ADDIE) framework. This integration ensured a systematic process for identifying instructional challenges, designing the PAP, and evaluating its effectiveness under three pacing conditions: System Paced (SP), Learner Paced (LP), and Learner Attenuated System Paced (LASP). Each pacing configuration was defined by specific learner control parameters, ranging from no control (SP) to full control over playback (LASP), with LP allowing for segmented learner-driven navigation. Both the ADDIE and DSRMP models have five phases. The framework for developing the model is shown in Figure 1 below.

Figure 1.

The Persona Creation Framework for Developing the PAP Module

By integrating these models, the researchers sought to gain from a thorough framework that directs every step of the research process, from problem identification to solution implementation. Since both models have been successfully employed in disciplines like education and information systems, this combination can be especially helpful in these areas (Bannan-Ritland, 2003).

Pedagogical Agent Persona Creation Process

Analysis (ADDIE) and Problem Identification (DSR)

The development of the Pedagogical Agent Persona (PAP) began with the analysis stage, where both primary and secondary data were gathered to identify the instructional needs in project-based learning (PBL) for senior high schools. Primary data came from teacher and student interviews as well as classroom observations. This provided insights into instructional practices and challenges in PBL delivery. Secondary data were drawn from a review of relevant literature, which highlighted the need for multimodal communication, flexible pacing, and culturally relevant visual representation. The problem was identified as the lack of effective strategies for teaching PBL in visual arts.

Design (ADDIE) and Suggestion (DSR)

Based on the analysis, the researchers designed the PAP through an iterative creative process. Initial thumbnail sketches of facial gestures and expressions were produced using Adobe Photoshop. These sketches were refined into comprehensive illustrations (see Figure 2), aligning with the principle that design evolves through iterative investigation and refinement to achieve its intended purpose (Jalil et al., 2024). Expert feedback from an educational technology specialist and an animation artist guided improvement to ensure that the design addressed both instructional and aesthetic needs.

Figure 2.

Sketches of Different Facial Gestures of the Agent Used in the Study

Development Phase (ADDIE) and Development (DSR)

In the development stage, the refined sketches were animated in Filmora to simulate natural blinking, mouth movements, and expressive gestures. Emotion was further conveyed through facial expressions such as smiling. The principal author recorded the instructional content with his phone recorder and later transferred it into the Filmora software to deliver voice in the instructional content, as research confirms that human narration fosters learner engagement by conveying prosody and emotional nuance (Ryu & Ke, 2018).

The PAP was embedded into a multimedia slide-based environment with three pacing configurations: System Paced, Learner Paced, and Learner Attenuated System Paced. Expert feedback prompted adjustments, including increasing the voice audibility and adding extra gestures to enhance social presence (see Table 1) (Figure 3).

Table 1.

Suggestions made by experts and improvements made afterwards

Position of expert	Suggestions from experts	Revisions made by the researchers
Lecturer in educational technology	The voice of the PAP is not audible enough	The voice volume was increased
Animation artist	Add extra gestures to PAP	Extra gestures were added to make the eye blink and the mouth move

Source: Field Survey, 2024

Figure 3.

Image of the Revised PAP Used in the Study

Implementation Stage (ADDIE)

At this stage, the revised PAP was introduced in the visual arts classroom for PBL on perspective drawing. The PAP acted as a virtual teacher, providing animated guidance through gestures, slides, and symbols. This scaffolding helped students to understand and execute their projects more effectively. Research by Mohammad (2019) similarly found that multimedia tools can aid teaching by making abstract content more accessible. Students were also oriented on how to interact with the PAP before the lessons commenced (see Figure 4).

Figure 4.

Learners Interacting With the PAP on Their Projects

Evaluation Phase (ADDIE) and Evaluation (DSR)

The PAP’s effectiveness was evaluated using the System Usability Scale (SUS) to measure usability, satisfaction, and user-friendliness across the three pacing modes. The SUS is widely recognized as a reliable instrument for usability testing across various contexts and sample sizes (Bangor et al., 2009; Martins et al., 2015; Setemen et al., 2019).

Conclusion (DSR)

Conclusions were drawn on the PAP’s usefulness for teaching PBL in visual arts. The evaluation pointed to its potential to enhance learner engagement and understanding. Findings were linked to the objectives of the study, with recommendations for future research focussing on refining multimedia artefacts to further support PBL delivery.

Participants

Participants comprised fifteen (15) second-year Visual Art students (average age 17.6 years) and two Visual Art teachers from Kumasi Senior High Technical School. Purposive sampling ensured that all participants had prior exposure to project-based learning in visual arts as verified through curriculum records, teacher confirmation, and preliminary interviews. Two experienced teachers, each with over 5 years of experience in PBL facilitation, were included to provide informed pedagogical perspectives. The 15 participants that consented to full participation were randomly placed, 5 students in each of the pacing conditions.

Ethical approval for the study was obtained from the Ghana Education Service (Reference Number: 14/90/148), along with written informed consent from all participants and guardians, where applicable. Verbal and written assent was also secured for participants under 18.

Data Collection

Observation

In research, observation is a useful technique for comprehending complicated problems (Weston et al., 2021). According to Speldewinde (2022), the degree of researcher participation in education settings varies. This may have an impact on the advantages and restrictions of the study. Observation was used at the analysis stage and implementation stage. The principal author was a participant observer and he actively immersed himself as a facilitator while observing and recording data about the activeness, enthusiasm, attention, exploration, usability and quality of work of learners. The goal was to gain a deeper insider perspective during the implementation of the prototype.

Questionnaire

A questionnaire is an important tool for the collection of data from potential respondents (Kumar, 2017). With a variety of response options and measurement scales available, questionnaires are an essential instrument for gathering data in research. In this research, closed-ended questionnaires were used at the design stage. Also, the System Usability Scale (SUS) was administered to assess students’ perceptions of the ease of use and effectiveness of the developed PAP. The SUS is a validated instrument originally developed by Brooke (1996) with high reliability (average Cronbach’s alpha >0.85). The scale includes 10 statements rated on a five-point Likert scale ranging from Strongly Disagree (1) to Strongly Agree (5), with scores converted to a 0–100 scale by multiplying each score by 2.5. Students responded using the numerical scale along with the corresponding verbal labels (Strongly Agree, Agree, Neutral, etc.). The SUS was administered in English, as it is the primary language of instruction in the school (Table 2).

Table 2.

System Usability Scale

No.	System usability score	1	2	3	4	5
1	I think that I would like to use this system frequently
2	I found the system unnecessarily complex
3	I thought the system was easy to use
4	I think that I would need the support of a technical person to be able to use this system
5	I found the various functions in this system were well integrated
6	I thought there was too much inconsistency in this system
7	I would imagine that most people would learn to use this system very quickly
8	I found the system very cumbersome to use
9	I felt very confident using the system
10	I needed to learn a lot of things before I could get going with this system

Source: Brooke (1996).

Semi-Structured Interview

During the analyses stage of the research, semi-structured interviews were conducted in both English and Twi to accommodate participant language preferences. The interviews explored participants’ views on current PBL methods, their expectations for the pedagogical agent, and their perceived usability of the platform. Interview guides were developed based on the research questions and validated by experts in educational technology.

Experts were also interviewed on their views on the PAP platform and suggested changes were made in the learning material. Interviews were audio-recorded with participant consent, transcribed verbatim, and analysed thematically using NVivo software. Observational notes were also taken to capture non-verbal cues. Each interview lasted between 20 and 45 minutes, with an average duration of 30 minutes. To ensure credibility, member checking was conducted by allowing participants to review and verify the accuracy of the transcripts. Table 3 outlines the data collection procedure, detailing the specific sources, techniques, instruments, analysis methods, and the exact timing of each in relation to the PAP intervention.

Table 3.

Data Collection Procedure

Steps	Source of data	Collection technique	Instrument	Data analysis technique	Timing of administration/Use
Analysis (ADDIE) and problem awareness (DSR)	-Primary data (teachers and students)	Semi-structured interviews	Interview guide	Thematic coding (Qualitative)	Conducted individually before the design phase to establish a baseline understanding of current PBL practices and challenges
Analysis (ADDIE) and problem awareness (DSR)	-Secondary data literature (Books, theses, article)	Semi-structured interviews	Interview guide	Thematic coding (Qualitative)
Design (ADDIE) and suggestion (DSR)	-Sketches and idea development, students’ responses, expert suggestions	Semi-structured interviews	Interview guide	Descriptive qualitative	Conducted during design refinement, prior to prototype finalization, to validate design decisions and gather expert input
Development	-Lesson notes, development of sketches into learning video	Expert review and feedback	–	Descriptive qualitative	Occurred prior to the intervention; expert feedback was integrated into the PAP module before implementation
Implementation and evaluation	-Observational data Students’ (SUS) usability ratings of the PAP platform	Observation and SUS questionnaire	Observation checklist, system usability scale (SUS)	Descriptive quantitative	Continuous observation was conducted throughout the intervention sessions. The SUS was administered immediately after each pacing condition session to capture participants’ post-interaction perceptions
Conclusion (DSR)	-Judgment on the effectiveness and outcomes of the PAP platform	Researcher reflection and synthesis	–	Qualitative	Conducted after completing all data collection and analysis phases to draw overall conclusions about the usefulness of PAP.

Compiled by researchers.

Source: Field Data, 2024

Semi-structured interviews with students and teachers during the analysis phase established the baseline pedagogical context before design work began while expert interviews took place during design refinement. Observations were conducted during each PAP-facilitated PBL session, and the System Usability Scale (SUS) was administered immediately after the intervention in each pacing condition. This structure ensured that each instrument captured data at the most contextually relevant point in the research process (Kumar, 2017; Weston et al., 2021).

Data Analysis

The study employed a mixed-methods approach to ensure triangulation and a comprehensive evaluation of the PAP’s impact. Qualitative data from semi-structured interviews with students and teachers were transcribed and thematically analysed using NVivo. Coding focused on recurring patterns related to perceptions of project-based learning (PBL), preferred characteristics of the PAP, and challenges with traditional instructional methods. This thematic analysis provided nuanced insights into participant experiences, which directly informed the iterative design and development of the PAP.

Quantitative data were analysed descriptively to complement the qualitative findings. System Usability Scale (SUS) responses were examined to determine usability scores across the three pacing conditions; System Paced, Learner Paced, and Learner Attenuated System Paced (Bangor et al., 2009; Martins et al., 2015; Setemen et al., 2019). Mean scores, percentages, and usability ratings were computed to assess effectiveness and user satisfaction. Observation checklists were also summarized using frequency distributions to capture patterns of student engagement, time management and enthusiasm during implementation. Together, these analyses ensured both depth and breadth in understanding the usability and instructional impact of the PAP.

Results

This section presents the empirical findings from the design, development, and implementation of a PAP aimed at enhancing project-based learning (PBL) in Ghanaian senior high schools.

Pedagogical Limitations in Current PBL Implementation

Qualitative data obtained through semi-structured interviews with two experienced visual art instructors revealed a pedagogical mismatch between traditional instructional practices and the requirements of effective PBL. Teachers identified significant challenges in monitoring individual progress and facilitating differentiated instruction in large class settings. Both educators consistently described current methodologies which are the lecture, discussion, and teacher-led demonstration as rigid, teacher-centred, and insufficient for fostering independent exploration, especially during studio-based projects.

One instructor noted that time constraints and students’ staggered arrival patterns often led to fragmented learning experiences, with key demonstration steps being missed by latecomers. The absence of visual references in syllabi and instructional materials was further cited as a limitation to student comprehension and autonomy.

These findings establish a compelling rationale for the development of a PAP that can function as a virtual facilitator to aid students internalize complex tasks without direct teacher intervention.

Based on the outcome of the interviews conducted with the two teachers, the identified issues and challenges are presented in the Table 4 below.

Table 4.

Problems with Project-Based Learning in the Traditional Classroom

Teaching methods	Problems identified
Demonstration	Not all students have the ability to clearly observe every step demonstrated by the teacher
	Due to time constraints, the teacher may not have the opportunity to reiterate the concepts being taught
	Students lack access to the procedures demonstrated by the teacher once the teacher has concluded the class
Discussion	As the discussions progress, students will be required to employ their imagination
Lecture	Student collaboration, inquiry and hands-on exploration is inhibited

Source: Interviews with teachers

Learner-Driven Design Features of the PAP

To guide the development of a learner-centred Pedagogical Agent Persona (PAP), students (n = 15) were surveyed and interviewed about their preferences for digital learning aids within the context of project-based learning (PBL). The findings revealed a strong consensus on the value of multimedia-enhanced instruction. All respondents (100%) indicated a preference for instructional materials that integrated slides with accompanying notes, illustrative images, and audio narration. This unanimous response underscores the importance students place on visual and auditory supports, which aligns with Mayer’s (2005) assertion that multimedia resources significantly enhance learning by combining visual and auditory stimuli. In addition, a substantial majority (93.3%) of participants endorsed the inclusion of human-like animation in the PAP design. This highlights the potential of the PAP to stimulate learner engagement and support comprehension, consistent with Jonassen’s (2011) argument on the role of cognitive scaffolds such as worked examples and simulations in promoting effective PBL experiences. Similarly, Lafifa and Rosana (2023) affirm the relevance of the PAP as an instructional medium for fostering critical thinking and strengthening digital literacy skills. Together, these findings provide an empirical foundation for tailoring the PAP design to student preferences, ensuring both pedagogical relevance and learner engagement (Table 5).

Table 5.

Descriptive Statistics Analysis of Appropriate Pedagogical Agent Persona Characteristics Preferred by Students in PBL

PAP characteristics	Agree		Disagree
PAP characteristics	N	%	N	%
Human-like animation	14	93.3	1	3.3
Slides with notes	15	100	0	0
Pictures	15	100	0	0
Audio	15	100	0	0

Source: Field Survey, 2024

Students emphasized the value of visual aids for clarifying abstract or technical aspects of drawing projects. Furthermore, interview responses indicated a need for design features that mirror traditional teaching methods while incorporating additional flexibility and interactivity. Concerns were raised regarding infrastructure challenges, such as unreliable electricity, suggesting that offline access or minimal-resource versions of the PAP may enhance future scalability.

Research on project-based learning (PBL) has demonstrated its positive impact on students’ content knowledge, skills development, motivation, and engagement (Megayanti et al., 2020). However, implementing PBL can be challenging, especially in terms of time management, monitoring, and assessment (Meng et al., 2023). PBL has shown particular effectiveness for students with learning difficulties, as it enhances their academic performance, motivation, and ability to work in groups (Filippatou & Kaldi, 2010).

Evidence of Enhanced Learner Engagement During Implementation

During the implementation phase, the revised PAP prototype was deployed in a real-time instructional setting, and student were systematically documented using an observation checklist. Indicators of engagement were measured across cognitive, and emotional dimensions. The results revealed consistently high levels of participation (see Table 6). A large majority of students (90%) were observed actively engaging with project tasks, while 95% maintained sustained focus throughout the lesson. Time management was effectively demonstrated by 85% of learners, and 80% displayed observable enthusiasm, indicating a strong emotional investment in the learning activity. Furthermore, 75% of students effectively used the PAP to complete their assigned tasks, reflecting adaptability and confidence in technology-mediated learning. Collectively, these observations provide compelling evidence of the PAP’s capacity to enhance learner engagement during PBL implementation.

Table 6.

Summary of observational activeness during implementation of the revised prototype

Activity indicator	Percentage (%)	Categorization
Actively engages with project tasks	90	Very active
Maintains focused attention	95	Very active
Utilizes technology tools effectively	75	Active
Demonstrates independent exploration	90	Very active
Produces high-quality work	80	Active
Manages time effectively	85	Active
Demonstrates enthusiasm	80	Active
Total	85	Very active

Source: Field Survey, 2024

The above table suggests that the activeness in the improved prototype is practical. It can be seen from students’ activities with an average percentage of 85% in the category of very active students.

The mean activity engagement rate was 85%, categorized as ‘very active’, confirming the PAP’s ability to elicit high levels of independent, sustained involvement in a studio-based visual arts task. This finding supports the PAP’s effectiveness in promoting the ‘learn-by-doing’ ethos of PBL.

Usability Outcomes Across Pacing Conditions

To assess the perceived usability and instructional satisfaction associated with different temporal structures, participants were randomly assigned to one of three pacing models: System Paced (SP), Learner Paced (LP), and Learner Attenuated System Paced (LASP). The System Usability Scale (SUS) was administered post-intervention (Table 7).

Table 7.

Usability Outcomes Across Pacing Conditions

Pacing condition	Mean SUS score	Qualitative interpretation
SP	80.0	Good usability
LP	86.5	Excellent usability
LASP	92.0	Best imaginable usability

All three conditions achieved usability scores well above the accepted benchmark of 68 (Sauro, 2011), with the LASP model emerging as the most effective in balancing structure and learner control. This finding aligns with emerging theories on optimal learner-system pacing convergence (Goguey et al., 2021; Schroeder & Craig, 2017). A representative comment from the SP group illustrates this:

‘Though I couldn’t rewind or forward the video, the way the teacher explained with diagrams helped me understand everything’.

This suggests that even in less interactive modalities, a well-designed pedagogical agent can compensate for pacing constraints through clarity, scaffolding, and visual cues. The above comment synchronizes with the studies on ‘Interaction Pace and User Preferences’ by Goguey et al. (2021) which clearly identified that the overall pace of interaction is determined by the combined pace of the user and the system. A mismatch in pace could potentially hinder user preferences. However, in the research by Schroeder & Craig (2017), their results on the three pacing conditions indicated that pedagogical agents were perceived similarly, regardless of whether learners had any control over the pacing of their learning environment or not.

Tan et al. (2022) argue that this it is important to take into account individual learning styles and preferences and design courseware that can accommodate these differences since the design of blended courses can have a significant impact on the academic performance of students.

The graph below further gives a graphical representation to the usability scores of participants on the developed PAP (Figure 5).

Figure 5.

System Usability Scale Score

Discussion

Advancing Constructivist Learning through Agent-Supported PBL

The findings affirm that Pedagogical Agent Persona (PAPs) are well-positioned to enhance constructivist pedagogies, particularly in settings where instructional individualization is constrained by large class sizes and limited teacher resources. As theorized by Jonassen (2011), PBL environments require tools that support student-driven exploration, iterative learning, and collaboration. The PAP functioned not merely as a content presenter but as a situated instructional actor, capable of guiding learners through scaffolded, multimodal experiences in the absence of continuous teacher intervention.

This design approach aligns with sociocultural learning theory (Vygotsky, 1978), where the agent acts as a mediated scaffold, extending the learner’s zone of proximal development. The high engagement levels observed, particularly in tasks requiring independent application, suggest that well-designed agents can simulate core aspects of human mentorship in digital contexts.

Human-like Design and the Multimedia Principle

Students’ strong preference for human-like, visually rich agents supports longstanding research on multimedia learning (Mayer, 2005) and recent studies on embodiment in digital agents (Lane & Schroeder, 2022). By integrating voice, gesture, facial expression, and diagrammatic cues, the PAP leveraged dual-channel processing to reduce cognitive load and enhance retention, especially critical in visual art education, where process-based learning is central.

Furthermore, learners’ comments reveal that the agent’s ability to ‘show and tell’ clearly was a decisive factor in usability satisfaction. This aligns with findings by Schroeder & Adesope (2013) showing that 3D agents outperform static visuals in skill-oriented domains. In this study, the PAP became a multimodal tutor; blending narrative, demonstration, and affective cues to deepen learner understanding.

Instructional Pacing: The Promise of Attenuated Autonomy

A notable contribution of this study is its comparative analysis of pacing models, particularly the introduction of the Learner Attenuated System Paced (LASP) condition. While Learner Paced (LP) models often prioritize autonomy, they can inadvertently burden novice learners with the task of managing complex workflows (Goguey et al., 2021). Conversely, System Paced (SP) models may oversimplify and constrain learner agency.

The LASP condition, where learners operated within structured sequences but retained some control, offered an optimal balance. With a SUS mean score of 92, this group reported the highest satisfaction, suggesting that hybrid pacing designs may represent a best-practice model for agent-led instruction. This aligns with the work of Lajoie & Poitras (2017), who argue that successful instructional technologies blend system-driven scaffolding with learner-responsive feedback.

Addressing Global Equity and Sustainable Development Goals

The context of this study, involving an urban Ghanaian classroom with infrastructural limitations, makes its contribution particularly relevant to ongoing debates about equity in educational technology. The academic community continues to emphasize the need for scalable and inclusive educational interventions that support the global commitment to Sustainable Development Goal 4, which guarantees inclusive and equitable quality education and promotes lifelong learning opportunities for all (United Nations, 2015).

The PAP’s ability to provide accessible, repeatable, and differentiated instruction, despite logistical constraints, positions it as a viable strategy for closing equity gaps in under-resourced educational systems. While most PAP literature is situated in high-income, digitally saturated environments (e.g. Mathews, 2021; Pérez-Marín, 2021), this study indicates that meaningful agent-based innovation is possible in lower-middle-income settings, provided the design is sensitive to the local context.

Limitations and Directions for Future Research

This study offers valuable insights into the use of pedagogical agents in project-based learning (PBL), but several limitations also point to opportunities for future research. First, the study took place in a single urban Ghanaian senior high school with a small, purposively selected sample. While the findings are contextually rich, they may not apply universally. Future research should explore similar interventions across diverse regions, school types, and cultural settings to strengthen external validity. Second, the intervention was brief, so longer-term effects such as sustained motivation, knowledge retention, and transfer could not be assessed. Future studies should track learners over an academic term or year to evaluate deeper educational impacts. Third, unreliable electricity and limited access to devices affected implementation. These challenges reflect real barriers in many under-resourced settings. Further research should explore how PAPs can be designed for low-tech environments, including mobile-first or offline models. Fourth, the agent supported one-on-one learning but did not address peer collaboration, which is a key element of PBL. Future iterations could integrate collaborative tools or multi-agent systems to support group work and the co-construction of knowledge. Fourth, this study did not compare the PAP with other instructional methods or technologies. Comparative studies would help determine how PAPs perform against traditional teaching, videos, or AI-based chat interfaces. Fifth, while this study successfully evaluated learners’ perceptions of the system through the System Usability Scale (SUS) and qualitative reflections, it did not include a direct assessment of the learners’ perceptions of the pedagogical agent itself. Future studies could incorporate a validated agent-specific measure, such as the Agent Persona Instrument–Revised (API-R; Kim & Baylor, 2016), to complement the SUS. Using both instruments together would enable a more holistic understanding of learners’ cognitive, affective, and social responses to the pedagogical agent.

By addressing these areas, future research can strengthen the design and implementation of pedagogical agents and extend their relevance to broader educational contexts.

Implications for Practice, Research, and Design

From a practical perspective, the findings of this study demonstrate how pedagogical agent personas (PAPs) can transform classroom practice within under-resourced senior high schools. The integration of a PAP in project-based learning (PBL) environments offers a new instructional paradigm that enhances learner autonomy, engagement, and instructional equity. Teachers, who often face the challenge of managing large classes and providing individualized support, can employ the PAP as an instructional companion that extends their pedagogical reach. This allows them to focus on higher-level facilitation tasks while the agent provides scaffolding, demonstrations, and continuous feedback. The implication for practice is that teachers must evolve from being knowledge transmitters to becoming facilitators of inquiry and creativity, supported by technology. Effective professional development in digital pedagogy is therefore crucial to enable educators to design, implement, and evaluate PAP-based interventions meaningfully within learner-centred frameworks.

The study also highlights that hybrid pacing models, such as the Learner Attenuated System Paced (LASP) condition, promote deeper learner engagement by striking a balance between autonomy and guided structure. In practical classroom settings, this means teachers and instructional designers should provide learners with some degree of control over pacing without sacrificing coherence and progression. Moreover, the PAP’s demonstrated capacity to provide consistent, individualized feedback supports its potential as a mechanism for ensuring equity in learning outcomes across large and diverse classrooms. At the institutional level, integrating PAP-supported learning aligns with Ghana’s educational technology policies, including the One-Teacher-One-Laptop initiative and the national agenda for achieving Sustainable Development Goal Four (SDG 4). Consequently, the study advocates for institutional investment in scalable, context-aware educational technologies that can enhance the accessibility and quality of project-based learning.

From a design and research standpoint, the study presents several imperatives for future work. Effective PAPs require culturally grounded humanization, where gestures, language, and affect resonate authentically with learners’ cognitive and emotional expectations. Hybrid pacing models like LASP should be further explored across disciplines and learning contexts, especially where learner variability and task complexity converge. Additionally, teachers should not merely be end-users of agent-based tools, but active co-designers who shape the agent’s instructional behaviour to fit their local classroom realities. Future research should examine long-term learning outcomes, transfer effects, and collaborative learning gains associated with PAP integration, particularly in group-based PBL contexts. These directions are essential for refining both the pedagogical design and sustainable deployment of PAPs as integral components of equitable, learner-centred education.

Theoretical Contribution and Novelty

This study advances theoretical understanding in three areas. Firstly, it operationalizes Vygotskian scaffolding in a digital format, demonstrating that agents can extend instructional reach in skill-based subjects such as visual arts. Secondly, it offers a framework for learner-attenuated pacing, which has not been extensively explored in the literature on pedagogical agents. Thirdly, it contributes to the globalization of PAP research, broadening the field’s empirical base beyond high-income contexts.

Conclusion

This study suggests that pedagogical agent personas (PAPs) can enhance project-based learning (PBL) by supporting learner engagement and autonomy, particularly in resource-constrained educational environments. The findings revealed that the Learner Attenuated System Paced (LASP) condition, where learners exercised partial control within a guided digital environment, resulted in the highest usability and satisfaction outcomes. This indicates that adaptive pacing models that combine learner agency with system scaffolding can sustain engagement and optimize instructional interactions.

The successful implementation of a multimodal, human-like PAP also demonstrated its capacity to supplement teacher instruction, extend teacher presence in large-class settings, and provide equitable access to personalized learning support. These outcomes underscore the transformative role of contextually designed instructional technologies in strengthening participation and engagement in PBL environments. While the present findings are context-specific, they offer a credible foundation for future studies to examine how the sustained use of PAPs may influence broader outcomes, such as critical thinking, collaboration, and lifelong learning.

Footnotes

ORCID iDs

Akwasi Adomako Boakye

Harry Barton Essel

Akosua Tachie-Menson

Francis Kofi Nimo Nunoo

Ama Amponsah Dwamena

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data supporting the findings of this study are not publicly available due to institutional restrictions.*

Appendix

References

Aidoo

(2023). Teacher educators experience adopting problem-based learning in science education. Education Sciences, 13(11), 1113. https://doi.org/10.3390/educsci13111113

Alisa

A. K.

Arkhangelskaya

Rudenko-Morgun

Lopanova

(2020). Pedagogical agents in teaching language: Types and implementation opportunities. IJAEDU-international E-Journal of Advances in Education, 5(15), Article 275285. https://doi.org/10.18768/IJAEDU.593875

Annan-Brew

(2020). Differential item functioning of west African senior school certificate examination in core subjects in southern Ghana. University of Cape Coast. https://doi.org/10.19044/esj.2020.v16n16p188

Appah

Tokede

Tunder-Francis

A. A.

Afolabi

Anifowose

(2021). Environmental condition, study habit, and academic self-concept of public senior secondary schools' adolescents in preparation for the 2020 West African senior school certificate Examination-A case of ibadan southwest. Ethiopian Journal of Environmental Studies and Management, 14(1): 47–58. https://doi.org/10.11v14i1.4

Bangor

Kortum

P. T.

Miller

J. T.

(2009). Determining what individual SUS scores mean: Adding an adjective rating scale. Journal of Usability Studies Archive, 4(1), 114–123.

Bannan-Ritland

(2003). The role of design in research: The integrative learning design framework. Educational Researcher, 32(1), 21–24. https://doi.org/10.3102/0013189X032001021

Beem

H. R.

(2021). Exploring the role of project-based learning in building self-efficacy in first-year African engineering students. In 2021 ASEE virtual annual conference content access. American Society for Engineering Education, 2021. https://peer.asee.org/37161.

Blair

K. P.

(2007). Pedagogical agents for learning by teaching: Teachable agents. Educational Technology, 47(1), 56–61. https://doi.org/10./stable/44429380

Boakye

A. A.

Essel

H. B.

Appiah

Anane-Antwi

(2024). Effectiveness of interactive multimedia-based learning in teaching art history as A component of general knowledge in art for senior high school students. Journal of Science and Technology (Ghana), 42(2), 140–156. https://doi.org/10.4314/just.v42i2.9

10.

Boss

Krauss

(2022). Reinventing project-based learning: Your field guide to real-world projects in the digital age. International Society for Technology in Education.

11.

Brooke

(1996). SUS: A quick and dirty usability scale. In Jordan

P. W.

, (Eds.), Usability evaluation in industry (pp. 189–194): Taylor & Francis.

12.

Bundsgaard

(2018). Using technology to scaffold progressive teaching In: Promoting Language and STEAM As Human Rights in Education. Springer, 109–124. https://doi.org/10.1007/978-981-13-2880-0_7

13.

Clarebout

Elen

Johnson

W. L.

Shaw

(2002). Animated pedagogical agents: An opportunity to be grasped? Journal of Educational Multimedia and Hypermedia, 11(3), 267–286.

14.

Clarebout

Heidig

(2012). Pedagogical agents. In Seel

N. M.

(Ed.), Encyclopedia of the sciences of learning (6, pp. 2567–2571): Springer.

15.

Douglass

S. M.

(2020). Relationship of teachers' transference and acceptance of information and communication technology professional development. Keiser University.

16.

Elfeky

A. I. M.

Elbyaly

M. Y. H.

(2023). The impact of project-based learning on the development of cognitive achievement in the course of applications in educational technology among students of the college of education at Najran university. European Chemical Bulletin, 12(5), 6643–6648. https://doi.org/10.31838/ecb/2023.12.si6.589

17.

Essel

H. B.

Vlachopoulos

Nunoo‐Mensah

Amankwa

J. O.

(2024). Exploring the impact of VoiceBots on multimedia programming education among Ghanaian university students. British Journal of Educational Technology, 56(1), 276–295. https://doi.org/10.1111/bjet.13504

18.

Essel

H. B.

Vlachopoulos

Tachie-Menson

Johnson

E. E.

Baah

P. K.

(2022). The impact of a virtual teaching assistant (chatbot) on students' learning in Ghanaian higher education. International Journal of Educational Technology in Higher Education, 19(1), 57. https://doi.org/10.1186/s41239-022-00362-6

19.

Filippatou

Kaldi

(2010). The effectiveness of project-based learning on pupils with learning difficulties regarding academic performance, group work and motivation. International Journal of Special Education, 25, 17–26. https://files.eric.ed.gov/fulltext/EJ890562.pdf

20.

Ghana Statistical Service . (2023). National education report 2022/23: Ghana Statistical Service. https://statsghana.gov.gh

21.

Goguey

Gutwin

Chen

Suwanaposee

Cockburn

(2021). Interaction pace and user preferences. In: Proceedings of the 2021 CHI conference on human factors in computing systems, New York NY USA, May 8–13, 202, (pp. 1–14). Association for Computing Machinery. https://doi.org/10.1145/3411764.3445772

22.

Grivokostopoulou

Kovas

Perikos

(2020). The effectiveness of embodied pedagogical agents and their impact on students learning in virtual worlds. Applied Sciences, 10(5), 1739. https://doi.org/10.3390/app10051739

23.

Hazard

(2023). Project-based learning for sustainable development: An open educational research framework for design. In: Proceedings of the Asian conference on education 2023. The International Academic Forum. https://doi.org/10.22492/issn.2435-9467.2023.70

24.

Jalil

M. Z. A.

Razali

Rahman

K. A. A.

Rahim

M. B.

Abd Samad

(2024). Development and evaluation of an augmented reality chiller system simulator for TVET teaching. Journal of Technical Education and Training, 16(1), 40–55. https://penerbit.uthm.edu.my/ojs/index.php/JTET/article/view/16324

25.

Jin

S. A. A.

(2010). The effects of incorporating a virtual agent in a computer-aided test designed for stress management education: The mediating role of enjoyment. Computers in Human Behavior, 26(3), 443–451. https://doi.org/10.1016/j.chb.2009.12.003

26.

Jonassen

D. H.

(2011). Supporting problem solving in PBL. Interdisciplinary Journal of Problem-Based Learning, 5(2), 95–119. https://doi.org/10.7771/1541-5015.1256

27.

Kadam

Vaidya

(2021). Cognitive evaluation of machine learning agents. Cognitive Systems Research, 66(1), 100–121. https://doi.org/10.1016/j.cogsys.2020.11.003

28.

Kim

Baylor

A. L.

(2016). Research-based design of pedagogical agent roles: A review, progress, and recommendations. International Journal of Artificial Intelligence in Education, 26(1), 160–169. https://doi.org/10.1007/s40593-015-0055-y

29.

Kumar

J. A.

(2021). Educational chatbots for project-based learning: Investigating learning outcomes for a team-based design course. International Journal of Educational Technology in Higher Education, 18(1), 1–28. https://doi.org/10.1186/s41239-021-00302-w

30.

Kumar

(2017). Response styles of structured questions in business research. Asian Review of Social Sciences, 6(2), 23–28. https://doi.org/10.51983/arss-2017.6.2.1371

31.

Lafifa

Rosana

(2023). Development and validation of animation-based science learning media in the STEM-PBL model to improve students critical thinking and digital literacy. Jurnal Penelitian Pendidikan IPA, 9(9), 7445–7453. https://doi.org/10.29303/jppipa.v9i9.4448

32.

Lajoie

S. P.

Poitras

E. G.

(2017). Crossing disciplinary boundaries to improve technology-rich learning environments. Teachers College Record: The Voice of Scholarship in Education, 119(3), 1–30. https://doi.org/10.1177/016146811711900301

33.

Lane

H. C.

Schroeder

N. L.

(2022). Pedagogical agents In: The handbook on socially interactive agents: 20 years of research on embodied conversational agents, intelligent virtual agents and Social Robotics Volume 2: Interactivity, Platforms, Application, (pp. 307–330). Association for Computing Machinery. https://doi.org/10.1145/3563659.3563669

34.

Laureano-Cruces

A. L.

Sánchez-Guerrero

Ramírez-Rodríguez

Ramírez-Laureano

(2022). Intelligent interfaces: Pedagogical agents and virtual humans. International Journal of Intelligence Science, 12(03), 57–78. https://doi.org/10.4236/ijis.2022.123005

35.

Wang

Mayer

R. E.

Liu

(2022). Animated pedagogical agents enhance learning outcomes and brain activity during learning. Journal of Computer Assisted Learning, 38(3), 621–637. https://doi.org/10.1111/jcal.12634

36.

Lin

Ginns

Wang

Zhang

(2020). Using a pedagogical agent to deliver conversational style instruction: What benefits can you obtain? Computers & Education, 143, Article 103658. https://doi.org/10.1016/j.compedu.2019.103658

37.

Llopis-Albert

Rubio

Mata-Amela

Devece

Torner-Feltrer

M. E.

Zeng

(2024). Project-based learning methodology (PBL) for the acquisition of transversal competences (TCs) and integration of sustainable development goals (SDGs) in mechanical engineering subjects. Multidisciplinary Journal for Education, Social and Technological Sciences, 11(1), 34–47. https://doi.org/10.4995/muse.2024.21101

38.

Lluch

I. M.

López

J. F.

Bermell

S. G.

Alberola

S. C.

Sánchez

(2020). Project-based learning in the subject energy and sustainable development. https://doi.org/10.4995/inred2020.2020.11978

39.

Martha

A. S. D.

Santoso

H. B.

(2019). The design and impact of the pedagogical agent: A systematic literature review. The Journal of Educators Online, 16(1), 1. https://doi.org/10.9743/JEO.2019.16.1.8

40.

Martins

A. I.

Rosa

A. F.

Queirós

Silva

Rocha

N. P.

(2015). European Portuguese validation of the system usability scale (SUS). Procedia Computer Science, 67(2), 293–300. https://doi.org/10.1016/j.procs.2015.09.273

41.

Mathews

J. L.

(2021). The effect of pedagogical agent persona on performance, self-efficacy, and attitudes in adult learners in an online environment. Doctoral Dissertation, Western Kentucky University. https://digitalcommons.wku.edu/cgi/viewcontent.cgi?article=1200&context=diss

42.

Mayer

R. E.

(2005). Principles of multimedia learning based on social cues: Personalization, voice, and image principles. In Mayer

(Ed.), The Cambridge handbook of multimedia learning (pp. 201–212): Cambridge University Press. https://doi.org/10.1017/CBO9780511816819.014

43.

McCallen

L. S.

Johnson

H. L.

(2020). The role of institutional agents in promoting higher education success among first-generation college students at a public urban university. Journal of Diversity in Higher Education, 13(4), 320–332. https://doi.org/10.1037/dhe0000143

44.

Megayanti

Busono

Maknun

(2020). Project-based learning efficacy in vocational education: Literature review. IOP Conference Series: Materials Science and Engineering, 830(4), Article 042075. https://doi.org/10.1088/1757-899x/830/4/042075

45.

Meng

Dong

Roehrs

Luan

(2023). Tackle implementation challenges in project-based learning: A survey study of PBL e-learning platforms. Educational Technology Research and Development: ETR & D, 71(1), 1–29. https://doi.org/10.1007/s11423-023-10202-7

46.

Mohammad

(2019). The development of “I-GEP” (Infographic of elements and principles of design) module as scaffolding in teaching and learning visual arts for secondary school [Master's Thesis, Universiti Teknologi Mara (UiTM)]. https://doi.org/10.13140/RG.2.2.11996.62084

47.

Moreno

Mayer

(2007). Interactive multimodal learning environments: Special issue on interactive learning environments: Contemporary issues and trends. Educational Psychology Review, 19(3), 309–326. https://doi.org/10.1007/s10648-007-9047-2

48.

Pan

Seow

P. S.

Shankararaman

Koh

(2021). An exploration into key roles in making project-based learning happen: Insights from a case study of a university. Journal of International Education in Business, 14(1), 109–129. https://doi.org/10.1108/jieb-02-2020-0018

49.

Pérez-Marín

(2021). A review of the practical applications of pedagogic conversational agents to be used in school and university classrooms. Digital, 1(1), 18–33. https://doi.org/10.3390/digital1010002

50.

Pérez-Sánchez

Sanchis

(2021). Sustainable development goals integrated in project-based learning in the mechanical engineering degree. In: Inted2021 Proceedings. IATED. https://doi.org/10.21125/INTED.2021.1502

51.

Piedrahita-Ospina

Moreno-Cadavid

(2019). Design science research for development and validation of a pedagogical agent for STEM education. In: 2019 XIV Latin American conference on learning technologies (LACLO) (pp. 129–133). IEEE. https://doi.org/10.1109/LACLO49268.2019.00031

52.

Prahani

B. K.

Trianggono

M. M.

Zahro

Siswono

Ashadi

Saphıra

H. V.

(2025). Effectiveness of digital project-based science learning in optimizing student’s creative thinking skills: Alignment with SDG 4 in higher education. Journal of Lifestyle and SDGs Review, 5(1), Article e03914. https://doi.org/10.47172/2965-730x.sdgsreview.v5.n01.pe03914

53.

Ramamurthy

Shafien

Nawi

N. S. M.

Sood

S. M. M.

Hussein

N. H.

(2023). Empowering students for the future: Project based learning and community building in the classroom. Insights journal: International English language teachers association. Journal, 6(1).

54.

Reategui

dos Santos

L. M. A.

Tarouco

(2012). Pedagogical agents and the efficiency of instructional conditions in educational applications. In Educational stages and interactive learning: From kindergarten to workplace training (pp. 121–133). IGI Global Scientific Publishing. https://doi.org/10.4018/978-1-4666-0137-6.ch008

55.

Ryong

Lee

(2023). Chatbot’s complementary motivation support in developing study plan of E-Learning English lecture. International Journal of Human–Computer Interaction, 40(10), 2641–2655. https://doi.org/10.1080/10447318.2022.2163786

56.

Ryu

(2018). Increasing persona effects: Does it matter the voice and appearance of animated pedagogical agent. Educational Technology International, 19(1), 61–91.

57.

Sauro

(2011). A practical guide to the system usability scale: Background, benchmarks & best practices. Measuring Usability LLC.

58.

Schroeder

N. L.

Adesope

O. O.

(2013). Do 3D pedagogical agents help students learn science? In Nettleton

Lennex

(Eds.), Cases on 3D technology application and integration in education (pp. 49–71): IGI Global Publishing. https://www.igi-global.com/chapter/pedagogical-agents-help-students-learn/74405

59.

Schroeder

N. L.

Adesope

O. O.

(2015). Impacts of pedagogical agent gender in an accessible learning environment. Educational Technology & Society, 18(4), 401–411. https://www.jstor.org/stable/jeductechsoci.18.4.401

60.

Schroeder

N. L.

Craig

S. D.

(2017). The effect of pacing on learners’ perceptions of pedagogical agents. Journal of Educational Computing Research, 55(7), 937–950. https://doi.org/10.1177/0735633116689790

61.

Schroeder

N. L.

Davis

R. O.

Yang

(2025). Designing and learning with pedagogical agents: An umbrella review. Journal of Educational Computing Research, 62(1), 1907–1936. https://doi.org/10.1177_07356331241288476

62.

Setemen

Erawati Dewi

Purnamawan

I. K.

(2019). PAON usability testing using system usability scale. Journal of Physics: Conference Series, 1165(2), Article 012009. https://doi.org/10.1088/1742-6596/1165/1/012009

63.

Song

Razali

A. B.

Jeyaraj

J. J.

(2025). How project-based learning improves college EFL learners’ critical thinking skills and reading comprehension ability: A case study. Language Teaching Research, Article 13621688251352275. https://doi.org/10.1177/13621688251352275

64.

Speldewinde

(2022). Where to stand? Researcher involvement in early education outdoor settings. Educational Research, 64(2), 208–223. https://doi.org/10.1080/00131881.2022.2064323

65.

Tan

J. Y.

Gervais

Koh

H. C.

(2022). A course level analysis of academic performance on adult learners. ASCILITE Publications. https://doi.org/10.14742/apubs.2019.278

66.

UNESCO Institute for Statistics (UIS) . (2024). World education indicators 2024: Class size and pupil–teacher ratios: UNESCO. Retrieved from: https://uis.unesco.org/en/topic/education-indicators

67.

United Nations . (2015). Transforming our world: The 2030 agenda for sustainable development (A/RES/70/1). United Nations. https://sdgs.un.org/2030agenda

68.

Vaishnavi

Kuechler

(2008). Design science research methods and patterns: Innovating information and communication technology: Auerbach Publications.

69.

Veletsianos

Russell

G. S.

(2013). Pedagogical agents. In Handbook of research on educational communications and technology (pp. 759–769): Springer New York. https://doi.org/10.1007/978-1-4614-3185-5_61

70.

Vygotsky

L. S.

(1978) In Cole

John-Steiner

Scribner

Souberman

(Eds.), Mind in society: The development of higher psychological processes: Harvard University Press.

71.

Wang

Gong

Jiang

(2022). Can affective pedagogical agent facilitate multimedia learning? Advances in Psychological Science, 30(7), 1524–1535. https://doi.org/10.3724/sp.j.1042.2022.01524

72.

Weber

Wambsganss

Rüttimann

Söllner

(2021). Pedagogical agents for interactive learning: A taxonomy of conversational agents in education In Forty-second international conference on information systems (pp. 1–17). researchgate.net.

73.

Weston

L. E.

Krein

S. L.

Harrod

(2021). Using observation to better understand the healthcare context. Qualitative Research in Medicine & Healthcare, 5(3), 9821. https://doi.org/10.4081/qrmh.2021.9821

74.

Woods

K. D.

(2020). Teacher technology efficacy: The relationship among generation, gender, and subject area of secondary teachers. https://digitalcommons.liberty.edu/cgi/viewcontent.cgi?article=3597&context=doctoral

75.

W.-C. V.

Lin

I.-T. D.

Marek

M. W.

Ou Yang

F.-C.

(2021). Analysis of English idiomatic learning behaviors of an audio-visual Mobile application. SAGE Open, 11(2), 21582440211016899. https://doi.org/10.1177/21582440211016899

76.

Žerovnik

Nančovska Šerbec

(2021). Project-based learning in higher education. In Vaz de Carvalho

Bauters

(Eds.), Technology supported active learning. Lecture notes in educational technology: Springer. https://doi.org/10.1007/978-981-16-2082-9_3