The present research focuses on a five-week collaborative project period within a larger PBLL presentation project through which students aim to develop content knowledge about their topic while simultaneously using English as the TL to improve their language skills. The theoretical foundation of this study draws on Vygotsky's (1978) socio- constructivist perspective, Bandura's (1986) social cognitive theory, and Kolb's (1984) experiential learning cycle. Beginning with Vygotsky (1978), who asserts that knowledge is not acquired in isolation but co-constructed through social interaction, peer collaboration and scaffolding are positioned as central processes through which learners deepen understanding. Building on Vygotsky's perspective, Swain and Lapkin (2001) argue that learners engage in collaborative dialogue using the TL, negotiating meaning and resolving linguistic problems in ways that support language development. Extending this focus, Bandura (1986) emphasizes self-regulation and goal setting, which are particularly relevant in multi-week group projects where successive tasks culminate in a final outcome; within this process, collaboration also contributes to the development of self-efficacy and confidence. Lastly, Kolb (1984) describes learning as a cyclical process through which learners construct knowledge by engaging in experience, reflection, conceptual understanding, and continued experimentation. Kolb and Kolb (2005) further emphasize the role of social interaction, suggesting that experiential learning is strengthened when learners collaborate to reflect, conceptualize, and experiment together. Collectively, these perspectives highlight how collaboration, interaction, and reflection support the co-construction of knowledge and ongoing language development in EFL contexts.

CL builds on these foundations by emphasizing partnership through structured social interaction. While definitions of CL vary, Dillenbourg (1999) defines it as "a situation in which particular forms of interaction among people are expected to occur, which would trigger learning mechanisms" (p. 5). This definition aligns with cooperative learning, which more specifically focuses on small-group structures that promote interdependence, personal accountability, participation, interpersonal skills, and reflection through structured tasks (Johnson & Johnson, 1987). While CL takes many forms, it is particularly relevant in EFL contexts, where it is widely recognized as an effective approach to second language acquisition (Jeong, 2019; Liu & Lan, 2016; Oxford, 1997; Putimasurai et al., 2024; Swain & Lapkin, 2001) and has been shown to function across both in-person and online environments (Nipp & Palenque, 2017). Within second language (L2) contexts, this interactional process may also involve collaborative dialogue, where learners jointly negotiate meaning, solve linguistic problems, and support one another's language development through communication in the TL (Swain & Lapkin, 2001). At the same time, effective collaboration depends on active participation, feedback, and reflection. Issues such as social loafing or free-riding (Strong & Anderson, 1990), uneven participation, task dominance, and conflict (Pauli et al., 2008; Storch, 2002; Tuckman, 1965) can disrupt group processes, highlighting the need for structured task design and appropriate assessment to support meaningful collaboration and accountability (Johnson & Johnson, 1987; Thomas, 2000).

Effective student collaboration can be achieved through multiple instructional methods; however, Project-Based Learning (PBL) provides one of the most effective frameworks for sustained cooperative engagement. PBL is an instructional approach through which students engage in a sustained, inquiry-driven investigation of authentic and complex real-world questions or problems that culminate in a finished, presentable product (Blumenfeld et al., 1991; Thomas, 2000). According to Krajcik and Shin (2014), there are six distinct features of PBL, which include choosing a primary focus of inquiry, establishing learning objectives, engaging students in structured activities, encouraging collaboration, incorporating technological scaffolds, and producing meaningful project outcomes. When these six features are implemented in a systematic and meaningful way, PBL has been shown to promote deeper learning, critical thinking, motivation, and the development of transferable skills such as problem-solving and teamwork, particularly when supported by appropriate scaffolding, feedback, and the effective integration of supporting technologies (Guo, P. et al., 2020; Guo, Q. et al., 2024; Lee & Kim, 2021; Liu & Lan, 2016; Putimasurai et al., 2024; Thomas, 2000). However, when these features are insufficiently structured or when planning, scaffolding, and technology integration are not effectively aligned, the benefits of PBL may be diminished, leading to challenges in collaboration, participation, and overall learning outcomes.

Expanding on the foundation of PBL, PBLL is an approach that aims to combine the effectiveness of PBL and Communicative Language Teaching (CLT). In the realm of second language education, CLT is widely regarded as an accepted method effective in enhancing engagement and motivation, and improving students language proficiency (Zou, 2025). CLT, like PBL, is a student-centered approach focused on authenticity. Beckett and Slater (2005) stress the importance of a project framework that combines academic content and higher-order skills in tandem with English language learning for effective acquisition. Thus, PBLL combines the effectiveness of a sustained, collaborative, content-focused project with authentic communicative language development. Farouck (2016) outlines a framework for PBLL based on CLT principles that consists of a seven-phase project which was adapted for the present research. The seven phases outlined include: 1) Designing the PBLL curriculum, 2) Introduction of the project tasks, 3) Selection of driving questions, 4) Introducing ICT tools, 5) Developing knowledge and skills to present the content, 6) Developing and revising presentations, and 7) Presentation of the final product (Farouck, 2016). Within these PBLL processes, ICT tools frequently support communication, coordination, drafting, revision, and presentation development throughout the collaborative process (Jeong, 2019; Liu & Lan, 2016). In Korean university EFL settings, students commonly rely on communication platforms such as KakaoTalk alongside shared presentation software, word-processing tools, and multimedia resources to coordinate tasks and exchange feedback both inside and outside the classroom. Within these environments, ICT functions less as the focus of instruction and more as a facilitative mechanism supporting interaction, organization, scaffolding, and collaborative project development over time. To date, research exploring students' perceptions of PBLL has primarily focused on post-project results (Farouck, 2016; Franzese & Irvin, 2025), leaving a gap in understanding how students' perceptions of collaboration and learning evolve throughout a multi-week PBLL project.

A key principle of PBLL is sustained collaboration within a multi-week inquiry-driven project; therefore, how groups form, develop, and organize roles is central to project success. Prior research indicates that several key factors exist that predict the effectiveness and success of a group, such as selection of group members, size of the group, background knowledge, and experience (Klug & Bagrow, 2016), but effective collaboration also requires time for members to become acquainted and develop a cohesive working relationship. Roles, whether assigned or emerging naturally, are not fixed but shift in response to group dynamics and individual contributions. In some cases, leadership may be shared across members rather than given to one individual, particularly when group members have varied skill sets or expertise (Carson et al., 2007; Denis et al., 2001). As such, both role formation and group development should be understood as fluid and evolving processes.

PBLL projects are typically guided by a structured, task-based framework, yet students retain autonomy in how collaboration unfolds; as a result, group development becomes an integral part of the project process. Group development refers to the stages groups pass through as they form and work toward shared goals. Tuckman (1965) originally identified these stages as Forming, Storming, Norming, and Performing, with Adjourning later added (Tuckman & Jensen, 1977). According to Tuckman (1965), Forming marks initial group organization, while Storming involves negotiating roles and managing conflict. Norming reflects the development of cohesion and shared expectations, followed by Performing, where groups work collaboratively toward task completion. Adjourning signals the conclusion of the project and reflection on the experience (Tuckman, 1965; Tuckman & Jensen, 1977). Although often presented linearly, these stages are not fixed, as groups may revisit earlier phases as tasks evolve (Dörnyei & Murphey, 2003). In multi-week PBLL projects, where tasks are built progressively, groups may cycle through stages multiple times to support the completion of successive tasks. As roles and responsibilities shift, ongoing reflection allows groups to adjust their approach and maintain effective collaboration throughout the project.

During the group development process, student roles become central to how tasks are completed. While some research supports explicit role assignment and rotation to promote accountability and balanced participation (Cohen, 1994; Johnson & Johnson, 1987), others suggest roles should emerge naturally, as rigid assignment may constrain group dynamics (Storch, 2002). In practice, roles are rarely fixed; instead, they shift in response to group needs and individual contributions, with students often taking on multiple responsibilities over time. Within more-collaborative roles, students may act as facilitators or leaders who guide the group's progress, encouragers who support group cohesion, or recorders who organize and synthesize ideas (Johnson & Johnson, 1987). At the same time, less- collaborative roles may emerge that disrupt group functioning. For example, free-riders or disengaged members contribute minimally while still benefiting from group outcomes, often reflecting patterns of social loafing (Albanese & Van Fleet, 1985; Latané et al., 1979; Strong & Anderson, 1990), whereas "lone wolves" may work independently and resist collaboration despite strong individual performance (Barr et al., 2005). These contrasting roles illustrate how student participation exists along a continuum, where contributions can either support or hinder the development of effective collaboration.

In addition to individual roles, other researchers have examined the patterns of interactions that emerge between group members. Specifically, Storch (2002) identifies four interaction patterns: Collaborative, Dominant/Dominant, Dominant/Passive, and Expert/Novice. According to Storch (2002), a Collaborative pattern is characterized by mutual engagement and shared contributions among group members. A Dominant/Dominant pattern occurs when members attempt to lead but fail to engage with each other's ideas, while a Dominant/Passive pattern reflects an imbalanced dynamic in which one member directs the task with limited input from others. The Expert/Novice pattern, in contrast, involves more knowledgeable members supporting others through scaffolding, which can facilitate learning (Storch, 2002). Importantly, these interaction patterns are not static and may shift as groups develop over time. More specifically, interaction patterns may also evolve throughout different stages of a multi-week PBLL project as task demands, leadership responsibilities, and collaborative relationships change over time. Viewed alongside student roles, these patterns further highlight the dynamic and evolving nature of collaboration as a function of group development in PBLL contexts, highlighting the need to examine not only outcomes but also how collaborative processes and student perceptions develop throughout a multi-week project.

Effectively measuring collaboration can be challenging, and approaches vary depending on project objectives and assessment methods. Most commonly, the efficacy of PBL has been examined through pre- and post-tests (Thomas, 2000). While these provide useful insight, additional methods have been employed to capture the CL process more fully. In a systematic review of 76 PBL-focused empirical studies, Guo et al. (2020) found that research relied on a wide range of tools, including questionnaires, rubrics, tests, interviews, observations, self-reflection journals, and artifacts. Moreover, studies measuring knowledge and skill acquisition, perceptions of collaboration, and student engagement all drew from these overlapping methods, with engagement examined less frequently but through similar instruments. Despite this range of approaches, most studies relied on intermittent or limited data collection points rather than tracking students' experiences consistently over time (Guo et al., 2020). These prior results indicate that much of the existing research has emphasized outcome-oriented measurement rather than process-based tracking of collaboration over time.

Across East Asian EFL contexts, research examining collaboration and ICT-supported PBLL has consistently reported positive outcomes related to communication, engagement, and perceived learning; however, these studies have primarily relied on pre- and post-project measurements or retrospective reflections. For instance, Liu and Lan (2016) used pre- and post-tests in Taiwan to examine motivation and ICT use, while Shin (2019) investigated Korean university students over a 10-week project, reporting gains in communication, critical thinking, and problem-solving, yet relying exclusively on pre- and post-testing. Additionally, studies in Korean EFL contexts have incorporated questionnaires and interviews (Jeong, 2019) or pre- and post-questionnaires (Lee & Kim, 2021), while others have used single post-project surveys to examine student perceptions, including both benefits and challenges such as language barriers and fairness concerns (Lee & Osman, 2020). Although these studies consistently highlight the benefits of CL, their reliance on limited or retrospective data collection may overlook how collaboration develops during the process and may be subject to recall bias (Schwarz, 1999). In contrast, Watanabe and Swain (2007) demonstrate the value of a multi-phase design, showing how tracking interaction across stages can provide a more nuanced understanding of collaborative dynamics, particularly in relation to scaffolding between learners of differing proficiency levels. Building on this, there remains a need for research that captures collaboration as it unfolds; therefore, the present study tracks students' perceptions and satisfaction on a weekly basis to better understand how collaborative experiences develop throughout a multi-week PBLL project.

As noted earlier, research examining roles and relationships in group work is well established (Storch, 2002), yet few studies track how these roles shift over the course of a project. The present study addresses this gap by examining changes in student roles throughout a multi-week PBLL project. To do so, seven roles were identified: Leader, Independent Worker, First-Follower, Group Collaborator, Passenger, Disengaged, and Absentee. Roles such as Leader, First-Follower, and Group Collaborator reflect commonly identified collaborative functions (Johnson & Johnson, 1987), with the First-Follower supporting and reinforcing group direction and the Group Collaborator encompassing functions such as organizing, summarizing, and maintaining progress. At the same time, roles were intentionally left flexible to allow students to interpret and enact them in response to task demands. In contrast, other roles reflect less- collaborative patterns drawn from prior research. The Independent Worker aligns with the concept of the "lone wolf" (Barr et al., 2005), referring to a member who works separately from the group, while Passenger and Disengaged correspond to forms of social loafing and free-riding, representing minimal contribution or limited participation despite shared outcomes (Albanese & Van Fleet, 1985; Latané et al., 1979). These roles were framed using more neutral terminology to encourage accurate self- and peer-identification without attaching overtly negative labels. By tracking these roles alongside students' perceptions over time, this study seeks to capture how collaboration evolves throughout the project, providing a more detailed account of group development and learning processes in PBLL contexts.

This study took place at two campuses of the same private university in South Korea during the fall semester of 2024. Participants were enrolled in four sections of a prerequisite, three-credit general English course primarily taken by first-year university students. Each researcher taught two sections, with classes meeting twice weekly for 90 minutes. As part of the university's graduation requirements, students completed a two-course English sequence designed to support four-skill language development. The second course placed greater emphasis on public speaking and presentation skills, supported by academic writing. To ensure consistency across sections, the researchers collaboratively designed the course curriculum, assessment criteria, and a PBLL group presentation project that served as the central instructional task.

The project was developed prior to the semester to align with course objectives and ensure consistent implementation across campuses. The final design was an eight-week instructional sequence, adapted from Farouck's (2016) PBLL framework and organized into a six-phase structure that guided students from initial task orientation through preparation, development, and final presentation delivery. Within this broader eight-week sequence, the present study specifically focuses on the five-week project implementation phase, during which students actively engaged in collaborative task development and production. Moreover, rather than focusing solely on the final product, the project was structured as a process-oriented sequence of tasks that emphasized scaffolded collaboration, sustained engagement, and repeated group interaction. As a result, formative feedback was provided throughout the process, and students were provided with opportunities to negotiate roles, make collaborative decisions, and reflect on their individual and group contributions over time. This process-oriented design formed the basis for examining students' J week-by-week engagement with the project.

Table 1 provides an overview of the weekly instructional focus and collaborative activities that structured the five- week process examined in this study. Following an initial orientation week, participants were randomly assigned to groups of three or four. The project then unfolded across a five-week, process-oriented instructional sequence. Each week, groups worked toward structured goals, progressing through topic selection, planning, drafting, revision, and rehearsal. Instruction was supported through guided activities and structured group work during twice-weekly class sessions. Groups were also encouraged to continue working outside of class as needed. As a result, patterns of coordination and communication varied across groups and individuals. These in- and out-of-class interactions were captured through weekly reflection surveys. Additionally, ICT-supported tools such as tablets or laptops, messaging platforms (e.g., KakaoTalk), shared documents (e.g., Microsoft Word, HanCom, Google Docs), and presentation software (e.g., Microsoft PowerPoint, Google Slides) were used to support communication, task management, and collaborative production. At the end of each week, students completed a structured reflection survey capturing their perceptions of role adoption, group collaboration, and learning at that stage of the project.

Convenience sampling was used, with participants drawn from four sections of the same general English course taught by the two researchers across two campuses, each instructor responsible for two sections, initially totaling 81 students. Students were informed of the study procedures at the start of the semester and could withdraw at any time. To be included in the final dataset, participants were required to complete a background survey and all five weekly reflection reports. One participant was excluded after self-identifying as a native English speaker rather than a Korean EFL learner. After removing disqualifying cases, the final sample consisted of 49 students, representing a retention rate of 60.49%. This attrition rate reflects natural participant drop-off across the five-week data collection period and is further addressed in the discussion of this study's limitations.

The sample included more female students (n=30; 61.22%) than male students (n=19; 38.78%), with an average age of 20.53 years (range: 19-25), consistent with the course's primary enrollment of first-year university students. Participants represented a range of academic majors, most commonly Nursing (n=14; 28.57%) and Foreign Languages (n=14; 28.57%), with additional representation from Engineering (n=6; 12.24%), Law (n=5; 10.2%) Communication Design (n=2; 4.08%), and Public Policy (n=2; 4.08%). Majors consisting of only one participant were grouped together (n=6; 12.24%). Participants' self-reported English proficiency was also collected to provide a more nuanced account of language backgrounds. While students were placed into course sections at the advanced level based on institutional placement tests, they reported perceived proficiency using the Common European Framework of Reference for Languages (CEFR). The most frequent level was B2 (n=15; 30.61%), followed by equal proportions of C1 (n=13; 26.53%) and B1 (n=13; 26.53%). Levels of A2 (n=7; 14.29%) and C2 (n=1; 2.04%) were also represented. Overall, these self-reported levels indicate a sample largely composed of upper- intermediate learners with some variation in perceived proficiency. These participant characteristics provide context for interpreting students' experiences of collaboration, role adoption, and perceived learning throughout the PBLL project.

To capture students' evolving perceptions of learning, collaboration, and role adoption throughout the process- oriented PBLL group presentation project, this study employed a researcher-designed measurement instrument, the Weekly Reflection Progress Report. The development of the Weekly Reflection Progress Report was guided by the theoretical frameworks of socio-constructivism, social cognitive theory, and experiential learning, ensuring that survey constructs aligned with key dimensions of interaction, self-regulation, reflection, and collaborative knowledge construction in PBLL contexts. In line with these frameworks, the instrument was designed as a repeated, process- oriented measure intended to capture changes in student experience across time rather than single-point evaluation. As such, the instrument was administered at the end of each instructional week across the five-week project sequence described in Section 3.1, allowing repeated documentation of students' experiences as they progressed through different stages of the group task. Rather than relying on a single pre- or post-project measure, the instrument was designed to track week-by-week variation in satisfaction, perceived learning, and collaborative engagement. In line with this process-oriented focus, the survey used a structured format that included five-point Likert-scale items, multiple-choice questions, and open-ended reflection prompts. This design allowed the collection of both descriptive trends over time and contextualized insights into students' experiences, keeping the instrument closely aligned with the study's research questions on roles, collaboration, and learning development.

The five Weekly Reflection Progress Reports were organized into four parts. First, Part 1 collected basic participant identifiers for administrative purposes, including tracking survey completion, organizing weekly responses, and determining eligibility for inclusion in the final dataset. Next, Part 2 focused on students' individual weekly experiences. It included six Likert-scale items assessing personal satisfaction and perceived English skill development. These consisted of one overall satisfaction item (e.g., "Please rate your overall personal satisfaction level with this week's progress, including work conducted both in and out of the classroom"), one item measuring perceived overall English improvement, and four items addressing development across the four language skills through the Perceived Language Skill Development scale (e.g., "Please rate whether you believe your English skills have improved in the following areas based on this week's progress, including work conducted both in and out of the classroom: Reading, Listening, Writing, Speaking"). Part 2 also included multiple-choice items assessing the proportion of English and Korean used during weekly group work (e.g., "Please indicate how much English you used during this week's progress"), as well as one open-ended prompt inviting students to describe their participation and experiences in the week.

The following parts of the survey aimed to understand students' roles within the group and their overall collaborative efforts. In Part 3, group roles were examined. Students identified their primary and secondary roles for the week from a predefined set (i.e., Leader, Independent Worker, First-Follower, Group Collaborator, Passenger, Disengaged, Absentee) and reported their perceptions of peers' roles. Two open-ended questions asked students to explain their role choices and their perceptions of peers' roles, providing additional context for analyses of role development and negotiation. Finally, Part 4 addressed team collaboration and group functioning. It included seven Likert-scale items drawn from the Group Collaboration Evaluation scale (e.g., "Please rate your overall satisfaction with your group's progress this week, including work conducted both in and out of the classroom") and the Perceived Group Learning Benefit scale (e.g., "Please rate whether your group contributed to improving your English during this week's progress"). Part 4 also included multiple-choice items capturing group-level language use and one open-ended prompt asking students to reflect on their feelings about their collaborative experience. Cronbach's alpha tests indicated acceptable internal reliability for the three primary scales, with α=.816 for the Perceived Language Skill Development scale, α=.818 for the Group Collaboration Evaluation scale, and α=.821 for the Perceived Group Learning Benefit scale. Collectively, these reports provided the primary data source for analyzing week-by-week changes in students' roles, collaboration, and perceived learning throughout the PBLL project. Finally, the full set of Weekly Reflection Progress Report items included in Parts 2-4 are provided in the Appendix.

Data collection for this study was conducted using the Weekly Reflection Progress Report, administered as five identical surveys at the end of each instructional week. The two researchers collaboratively developed the instrument to align with the study's research questions and ensure consistent coverage of role adoption, collaboration, and perceived learning. After drafting the surveys, a bilingual professor, fluent in English and Korean with over fifteen years of experience, reviewed the instrument, resulting in minor revisions and Korean translations to improve clarity. Each weekly survey was distributed through the university's Learning Management System (LMS) or shared via QR code in the in-class PowerPoint materials. Students were instructed to complete the report by the start of the following week's class, with nearly all participants choosing to complete the surveys outside class. Once collected, responses were automatically recorded through Google Forms, and the first author organized the dataset by removing incomplete cases, assigning randomized participant codes (e.g., S1-S49), and compiling each week's responses into separate Excel sheets for analysis across the five-week project timeline.

Survey responses were analyzed using a descriptive and inductive approach consistent with the process-oriented aims of the study. Multiple-choice and Likert-scale responses were examined descriptively in Excel to identify frequencies, means, and week-by-week patterns across the five surveys. Following Kaur et al. (2018), these summaries documented trends in students' reported satisfaction, perceived learning, collaboration, and role adoption, providing a foundation for qualitative interpretation. Internal reliability for grouped Likert-scale items was assessed using Cronbach's alpha to support interpretation of these trends. Open-ended responses were then analyzed to contextualize and elaborate on the observed patterns. Qualitative analysis followed an inductive thematic approach (Maykut & Morehouse, 1994; Merriam, 1998) and built on prior work demonstrating the value of long-response survey data (Jansen, 2010). Given the size and repeated nature of the dataset, coding was conducted iteratively alongside ongoing familiarization, consistent with Saldana's (2021) cycle-based approach for large qualitative datasets. To enhance trustworthiness, strategies from Lincoln and Guba (1985) were applied, including iterative review, triangulation of themes across all five weeks, and maintenance of an audit trail documenting coding decisions and theme development. In addition, as the instructor- researcher was embedded within the learning context, the study acknowledges the positionality of insider research. Following Griffith (1998), such positioning can enhance interpretive depth when approached reflexively and transparently, particularly in classroom-based inquiry. Lastly, representative quotations were selected to illustrate key themes and preserve the authenticity of student voices. Collectively, these procedures support the credibility, dependability, and confirmability of the findings.

To examine how students' experiences evolved across the five-week PBLL project, learners reported weekly on their satisfaction, perceptions of collaboration, perceived language skill development, perceived group learning benefits, and English language use during group work. Building on the role-based participation patterns described in Section 4.1, this section shifts focus from how students participated to how they evaluated those experiences over time. Figure 2 presents these measures using weekly descriptive means, allowing for comparison of descriptive patterns across the project timeline.

As shown in Figure 2, students' overall satisfaction and perceptions of group collaboration remained relatively high across the project, while self-reported English language use followed a lower and more stable trajectory. Satisfaction varied within a narrow range (3.84-3.96), showing a slight increase from Week 1 through Week 4 before a small decline in the final week. This pattern aligns with the project timeline outlined in Section 3.1 (Table 1), in which early stages emphasized group formation and planning, followed by progressively more demanding coordination, production, and rehearsal tasks. In contrast, collaboration ratings showed greater fluctuation than satisfaction, beginning at their highest point in Week 1, decreasing during Weeks 2 and 3 as task demands intensified and responsibilities became more distributed, and then rising again in Weeks 4 and 5 as groups worked to integrate materials and prepare final submissions. At the same time, perceived English language use remained below the midpoint of the rescaled metric across all five weeks, indicating that students reported using English less than half of the time on average. English use was relatively higher during early planning and discussion stages and lowest during Week 4, when efficiency-driven revision and coordination were most prominent. Taken together, these patterns suggest that while evaluations of group work remained generally positive, language choice within groups shifted in response to task demands and coordination pressures. It is important to note that these observations are based on descriptive means and do not indicate statistically significant changes across weeks.

Further examination of Figure 2 suggests a gradual upward pattern in perceived language skill development and group learning benefits over time. Self-reported gains across reading, writing, speaking, and listening showed a steady increase from Week 1 (3.43) to Week 5 (4.02), representing the strongest upward trend among the measured variables. A related pattern is evident in perceived group learning benefits, which showed a slight decrease from Week 1 (3.97) to Week 2 (3.90) during the initial stages of group formation and role negotiation, before increasing consistently through Weeks 3-5 and reaching the highest level in the final week (4.14). Although these trajectories differ in their early stages, both appear to reflect an overall upward movement across the project. This may suggest that students' evaluations of learning became more positive as they progressed through successive stages of collaboration, with later weeks reflecting not only immediate experiences but also their growing familiarity with the tasks and expectations of the project.

While Figure 2 captures these patterns at a descriptive level, students' open-ended reflections help explain how and why these trends developed over time. In particular, qualitative responses provide additional context for the patterns observed in satisfaction and collaboration. Early reflections reveal mixed emotional responses that align with the relatively high but unsettled ratings observed in Week 1 of Figure 2. Many students described anxiety, pressure, and self-criticism during initial presentations, often linked to time constraints and limited confidence using English. For example, one student noted, "I felt very pressured and nervous due to the 5-minute time limit... I turned to speaking in Korean instead of English" (S1, Week 1). As the project progressed, these early reactions did not lead to consistently higher satisfaction scores. Instead, later reflections point to a mix of pressure and accomplishment as task demands increased. One learner described the final presentation as demanding but ultimately "giv[ing] me a sense of accomplishment" (S13, Week 5). Taken together, these accounts suggest that satisfaction was shaped less by immediate comfort and more by students' ability to work through increasingly complex collaborative tasks.

A similar pattern is evident in reflections on collaboration, mirroring the fluctuation observed in the descriptive trends. Early responses emphasized cohesion and shared engagement as groups established working routines and negotiated roles. One student described how this process created a sense of unity, explaining that "in this process, we worked as one, and I felt comfortable while I proposed some ideas" (S5, Week 1). Similar to learner satisfaction, later reflections on collaboration more frequently referred to the pressure associated with coordination, revision, and rehearsal tasks. Despite these challenges, students also described how continued cooperation helped groups manage these demands. For example, one student noted that although the workload was demanding, "I still feel happy when working with my peers" (S5, Week 5). Taken together, these accounts suggest that collaboration functioned both as a source of difficulty and as a resource that supported groups in completing the project.

Finally, reflections on learning align more closely with the steady upward trends observed in Figure 2. Rather than focusing on isolated weekly gains, students frequently described learning as developing through repeated participation in presentations and group interaction. Several responses noted that later presentations felt less burdensome due to prior experience, while others highlighted how continued collaboration increased confidence and ease of communication. At the same time, students' comments provide important context for the consistently low English-use ratings. While English was central to presentation outputs and prepared materials, many students reported relying on Korean during group coordination to maintain efficiency and clarity. As one student explained, "we mainly used Korean to understand each other's opinions clearly" (S43, Week 1). This contrast between reported learning gains and relatively low English use may suggest a distinction between perceived language development and actual language use during interaction. This pattern may reflect the role of the learner's first language (L1) as a facilitative tool for coordination and task management, even within contexts designed to promote L2 use. Given that language use was not a primary focus of the study design, these findings are interpreted as an emergent pattern rather than a central outcome. As a whole, the qualitative findings reinforce the descriptive trends by showing that perceived learning gains developed progressively across the project, even when English use during group interaction remained limited.

To complement the role-based participation patterns and descriptive trends reported in Sections 4.1 and 4.2, this section examines students' open-ended reflections on their experiences during the PBLL project. Rather than tracking week-by-week changes, reflections were analyzed to identify dominant experiential themes that students most frequently emphasized. Each comment was assigned a single dominant theme based on its primary experiential focus, allowing frequency counts to reflect the relative salience of experiential categories across the dataset. A summary of the ten dominant themes, organized into three thematic domains, is presented in Table 5.

A closer examination of Table 5 suggests that students’ reflections clustered around ten dominant experiential themes across three domains. Within Project Design & Development, Content Development (n = 98) appeared most frequently, followed by Feedback & Revision (n = 43) and Collaborative Rehearsal (n = 39). In Performance & Language Demands, Presentation Anxiety (n = 80) was prominent, while Language Difficulty (n = 29) appeared less frequently but remained consistent across responses. Within Coordination & Structural Constraints, Time Pressure (n = 48) and Scheduling Challenges (n = 47) were most common, followed by Team Synergy (n = 44), ICT & Logistical Challenges (n = 31), and Uneven Participation (n = 30). Overall, these patterns reflect a mix of positive and challenging experiences, with students emphasizing both the collaborative processes that supported project development and the practical and affective pressures. These patterns also highlight how coordination demands, language difficulty, and ICT-mediated collaboration shaped how groups worked through the project.

Across the first thematic domain, Project Design & Development, students most often described the practical work of shaping, refining, and practicing their group presentations. Consistent with the task ownership and organizational patterns reported in Section 4.1, many reflections emphasized topic selection, scope control, and structural clarity during the early stages of the project. As one student noted, “There is no topic without limitations. Rather, the group should develop it” (S2, Week 1). As drafts developed, attention shifted toward feedback and revision, with students describing instructor guidance and collaborative editing, for example, “After receiving feedback from the professor, we revised it together” (S9, Week 2). In addition, collaborative rehearsal became increasingly visible in later weeks as the presentation date approached, with groups practicing repeatedly to improve timing and delivery, such as “We prepared visuals together and rehearsed several times to improve our delivery” (S8, Week 3). Taken together, these comments suggest that project development was experienced as a sustained and shared process, consistent with the collaborative patterns observed earlier rather than a series of isolated individual tasks.

The second domain, Performance & Language Demands, reflects challenges closely tied to the presentation-based nature of PBLL. Given that the final outcome required public-speaking delivery in English, presentation anxiety emerged as one of the most frequently reported themes, aligning with the satisfaction fluctuations and English-use patterns described earlier in Section 4.2. From Week 1 onward, students described nervousness in direct terms, such as "I was very nervous about presenting in English in front of many people" (S7, Week 1). Similar reactions continued into later weeks, with one student recalling, "My head turned white and nothing comes to my mind" (S38, Week 2). Closely connected to these emotional responses was the theme of language difficulty, with students noting, "It is still difficult for me to explain it in English" (S20, Week 1) and "We were trying [not] to speak in Korean and express ourselves in English, but that was difficult" (S18, Week 3). These reflections appear to suggest that performance pressure and language difficulty were consistently intertwined, helping explain the fluctuations in satisfaction and English-use patterns observed earlier.

Finally, the third domain, Coordination & Structural Constraints, captures the broader conditions that shaped how groups functioned throughout the five weeks. An initial theme that emerged from student reflections was time pressure, often linked to competing academic demands, as seen in comments such as "I was very busy because I had to take a total of 5 examinations this week" (S23, Week 5). Closely related were scheduling challenges, particularly difficulties meeting outside class, with students noting, "It was difficult to gather face-to-face because they all have different timetables" (S39, Week 1). Uneven participation also appeared as a recurring concern, as one student explained, "The tasks that one team member was supposed to handle naturally fell onto the rest of us" (S14, Week 2). At the same time, ICT and logistical issues shaped collaboration, including disruptions such as "All the data were blown away in the process of editing it into English" (S4, Week 3), while also supporting coordination through shared tools. Thus, while ICT-supported tools frequently facilitated communication and task management, these reflections also suggest that technological challenges, unfamiliar platforms, and occasional data loss introduced additional complexity into the collaborative process. Despite these constraints, many students emphasized team synergy, describing how groups worked "as one" (S5, Week 1). Taken together, these accounts suggest that coordination challenges were a persistent feature of group work but were often managed through shared effort and ICT-supported collaboration, consistent with the patterns of adaptation observed across the project.

Over the course of this five-week PBLL group project, students' role selection remained a dynamic aspect of group development and a key factor in each group's overall functioning. As groups progressed through successive tasks, roles were continuously negotiated and adjusted, requiring students to revisit and move between stages of group development (Dörnyei & Murphey, 2003). Allowing students to select roles based on personal strengths and group synergy aligns with Storch (2002), who suggests that roles should emerge naturally through interaction. In this sense, the prevalence of collaborative and supportive roles observed in the results can be interpreted as broadly consistent with interaction patterns such as Collaborative and Expert/Novice, where mutual engagement and scaffolded support are emphasized (Storch, 2002). Additionally, as pointed out by Klug and Bagrow (2016), it is necessary that group members be flexible and willing to take on multiple roles for the successful outcome of a group project. In this study, students were given autonomy to determine roles through discussion and interaction, and results indicate that they were largely successful in negotiating responsibilities and expressed satisfaction with both their own and their peers' contributions. However, these findings should be interpreted cautiously, as self- and peer-reported data may reflect a tendency to emphasize more favorable or collaborative behaviors, particularly within multi-week PBLL projects that require students to assume roles aligned with their strengths at different stages of the project. At the same time, this process reinforces that group development is not static, but an ongoing process through which roles are adapted to meet evolving task demands and to build group cohesion (Dörnyei & Murphey, 2003; Tuckman, 1965). These findings suggest that, while guidance remains important, groups benefit from opportunities to develop organically through interaction, shared responsibility, and engagement in meaningful tasks. Considering these results, several pedagogical implications emerge regarding how PBLL projects can better support group development, role engagement, and learning outcomes.

The first implication is that students require explicit preparation in interpersonal and leadership skills to fully benefit from collaborative PBLL environments. While group development naturally supports cohesion, the findings suggest that students do not always enact leadership roles in ways that maximize group efficiency or satisfaction. Although leadership was clearly present, as reflected in peer reports and frequent student references, key components such as delegation and addressing imbalanced contributions were less evident. This aligns with prior work indicating that effective collaboration depends not only on participation, but also on the ability to guide, coordinate, and intervene when necessary (Klug & Bagrow, 2016). At the same time, students may be reluctant to confront peers due to sociocultural norms that emphasize harmony and avoidance of direct criticism (J. Y. Kim, 2016; M. Kim, 2017; Morita, 2004). As a result, leadership may be enacted in limited or indirect ways. In classroom practice, this suggests that instructors should incorporate brief, structured training or modeling activities early in the project, such as demonstrating how to assign tasks, provide constructive feedback, or address uneven participation within groups. To address this, instruction should focus on developing interpersonal competencies alongside task management, including assertiveness, conflict resolution, and equitable workload distribution. Framing leadership as a relational and facilitative role, rather than one based on authority, may help students more effectively support group functioning and monitor progress throughout the project.

The second implication is that structured role exposure, particularly through early role assignment and rotation, can support students understanding of collaborative responsibilities before transitioning to self-regulated role selection. Although students in this study demonstrated the ability to negotiate and adapt roles over time, satisfaction was influenced by perceived imbalances in contribution. Providing students with opportunities to experience different roles in the early stages of a project can help clarify expectations and develop a more concrete understanding of how each role contributes to group success (Cohen, 1994; Johnson & Johnson, 1987). Such structured experiences may also help mitigate issues such as social loafing or uneven participation, particularly among less active group members (Lee & Kim, 2021). For example, instructors might assign rotating roles (e.g., discussion leader, organizer, editor) during the first one or two weeks before gradually allowing groups to self-select roles. Importantly, this approach does not contradict the value of autonomy but rather supports it. As students gain experience and confidence, they can gradually transition to selecting and adapting roles based on task demands and group needs. In this way, role rotation functions as a developmental scaffold that strengthens both individual accountability and overall group cohesion within PBLL contexts.

The third implication is that effective PBLL design should integrate structured support, ICT-mediated collaboration, and a balanced approach to target language use. The findings suggest that collaborative learning is most effective when instructors move beyond a sole focus on final products and instead incorporate structured milestones and formative feedback throughout the project. Without such guidance, group work may be reduced to a task-oriented process rather than a meaningful collaborative experience. Consistent with prior research, ICT-mediated practices can further support collaboration by addressing logistical challenges and facilitating ongoing interaction, particularly through tools such as shared documents and communication platforms (Jeong, 2019; Lee et al., 2022; Putimasurai et al., 2024; Salovaara & Järvelä, 2003). In practice, this may involve requiring groups to use shared documents for drafting, messaging platforms for coordination, and periodic submission checkpoints to monitor progress and participation. At the same time, the results highlight a tension between students' perceived improvement in English and their dissatisfaction with actual language use. Although students reported increasing confidence and perceived skill development, English use during group interaction remained relatively limited, which may suggest a divergence between perceived learning and interactional language practices. While maximizing TL use is supported by theory (Ellis, 1994; Krashen, 1981), a more flexible approach appears necessary in multi-week PBLL contexts. Encouraging meaningful use of English while allowing strategic L1 use for task management may reduce anxiety and better align with students' experiences. As proficiency develops, greater emphasis can be placed on sustained target language use, supporting both language development and overall satisfaction.