Problem Design Characteristics in School-Based PBL: A PRISMA-Informed Review of Korean K-12 Cases

Abstract

Problem-based learning (PBL) relies on the problem as the generative trigger for inquiry, collaboration, and assessment, yet school-based reports often provide limited guidance on how problems are actually designed. This PRISMA-informed review analyzed 24 PBL cases published in Korea Citation Index (KCI) journals (2020–2025) to characterize enacted problem design in Korean K–12 settings. Using a literature-grounded framework covering authenticity, cognitive demand, collaboration, and pedagogical alignment, five expert coders conducted calibration and consensus coding of 12 elements on a 0–3 rubric. The findings showed that Perspective Integration, Analysis Requirements, and Assessment Opportunities were most strongly represented, whereas Curriculum Integration and Cognitive Conflict were less frequently advanced. Correlation and triad analyses further indicated recurring design patterns centered on collaboration and authenticity, while cognitively destabilizing configurations were rare. These results suggest that Korean K–12 PBL tends to emphasize socially supported, contextually meaningful inquiry. By contrast, the rarely observed cognitive-demand triad suggests a tendency to avoid problems requiring strong conceptual destabilization and deep analysis. These findings identify context-bounded design patterns and offer practical guidance for designing PBL problems that better balance socio-cognitive support, conceptual challenge, and curricular coherence.

1. Introduction

In problem-based learning (PBL), the problem is not a peripheral task but the central organizer of learning. It initiates inquiry, structures collaboration, directs self-directed learning, and creates opportunities for assessment (Hmelo-Silver, 2004; Savery, 2006; Schmidt, 1983, 1993). Because the quality and form of the problem fundamentally shape how PBL unfolds, understanding how problems are actually designed is essential to interpreting both the strengths and variability of PBL outcomes. Yet published school-based PBL reports often provide limited and uneven descriptions of problem design, making it difficult to determine which problem features are emphasized in practice and which are neglected. Systematic analysis of enacted problem design is therefore needed to clarify how PBL problems are conceptualized and operationalized in school-based settings.

While numerous studies have documented the effectiveness of PBL across various educational contexts (Hmelo-Silver, 2004; Savery, 2006), less attention has been paid to systematically analyzing the core element that drives PBL, the problem itself. Problems serve as the generative engine of inquiry in PBL, yet their design characteristics remain underexamined. Without a clear understanding of how problems are structured across key dimensions such as authenticity, cognitive complexity, collaborative demands, and pedagogical alignment, educators lack a systematic framework for designing and evaluating PBL experiences. This gap is particularly significant in K–12 contexts, where problem design must balance curricular constraints, developmental appropriateness, and authentic inquiry. Therefore, this study addresses a critical need by analyzing how PBL problems are conceptualized and designed in school-based settings, providing empirical evidence to inform more effective problem design practices.

The concept of “problem” in PBL is multifaceted, encompassing various dimensions such as cognitive, social, and contextual elements. PBL is characterized by its learner-centered approach, where students engage with complex, open-ended problems that require critical thinking and collaborative problem-solving (Palwe, 2022; Cui & Han, 2017). This approach is often compared to other pedagogical methods, such as Project-Based Learning (PjBL) and Challenge-Based Learning (CBL), which also emphasize active learning and real-world applications (Vilalta-Perdomo et al., 2022; Nizami et al., 2023). While PjBL and CBL typically center on extended projects or broad real-world challenges, PBL is defined by the problem as the generative trigger that precedes instruction and structures the learning process itself. In a prototypical PBL sequence, learners first encounter a problem scenario, engage in small-group inquiry, identify what they need to learn, pursue self-directed learning, and then return to report, synthesize, and assess their understanding. This sequence makes the problem a precondition for learning rather than a task appended after instruction, and it underwrites the present study’s coding emphasis on the problem artifact itself. By distinguishing PBL from related pedagogies in this way, the study establishes a clearer conceptual boundary for analyzing problem design in school-based cases.

Many implementation difficulties in school-based PBL may stem not only from teacher capacity or contextual constraints, but also from the quality of the problems themselves. When problems are weakly authentic, insufficiently challenging, poorly aligned with curricular goals, or lacking clear collaborative affordances, they are less likely to sustain meaningful inquiry and collaborative knowledge construction in PBL (Hmelo-Silver, 2004; Loyens et al., 2015; Savery, 2006). In such cases, teachers may struggle to facilitate inquiry, students may participate unevenly or engage only superficially, and assessment may become unclear or disconnected from learning goals (Alves et al., 2016; Kokotsaki et al., 2016). From this perspective, challenges commonly attributed to PBL implementation—including facilitation difficulties, uneven participation, and ambiguous assessment—can be understood, at least in part, as consequences of underdeveloped problem design. Reframing implementation issues in this way strengthens the rationale for systematic analysis of problem characteristics in school-based PBL.

Given the significance of problem framing in PBL, further investigation into this topic is warranted. Understanding how educators perceive and define problems can lead to more effective teaching strategies that enhance student engagement and learning outcomes. Moreover, studying the implications of problem selection and presentation on academic achievement can inform policy and practice in teacher education, ensuring that future educators are equipped with the skills necessary to implement PBL effectively (Susilawati et al., 2018; Naviri et al., 2021). As the educational landscape continues to evolve, the exploration of the meaning of “problem” in PBL will contribute to the development of more responsive and impactful teaching methodologies that prepare students for the complexities of the modern world.

This study employs a structured review and content analysis of published school-based PBL cases, rather than a survey of teachers’ beliefs. We focus on KCI-indexed Korean K–12 empirical reports (2020–2025) because they provide a coherent curricular and policy context and allow close inspection of school-implemented PBL problem artifacts. Specifically, the study examines how PBL problems are distributed across twelve design elements spanning four dimensions—Real-World Authenticity, Cognitive Engagement, Collaborative Learning, and Pedagogical Alignment—and identifies bivariate and triadic co-occurrence patterns that may inform future problem design. Frequency patterns indicate which features are emphasized or underutilized in current practice, whereas correlation and triad analyses reveal recurring design constellations in school-based PBL problem design. To address this aim, the study was guided by the following research questions:

RQ1: In KCI-indexed Korean K–12 school-based PBL case reports (2020–2025), how are PBL problems distributed across twelve design elements spanning authenticity, cognitive engagement, collaboration, and pedagogical alignment?

RQ2: What bivariate (pairwise) and multielement (advanced-level triad) co-occurrence patterns link these elements within this corpus?

2. Literature Review

2.1. Features of PBL

Problem-Based Learning (PBL) is an instructional method rooted in constructivist learning theory that emphasizes student-centered learning through the experience of solving complex, real-world problems, aiming to develop flexible knowledge, problem-solving skills, self-directed learning, collaboration, and intrinsic motivation. One of the core characteristics of PBL is the activation of prior knowledge, which encourages students to connect new information with existing knowledge, enhancing learning and retention through the reorganization of students’ knowledge structures and promotion of epistemic curiosity by discussing relevant problems in small groups (Schmidt, 1983, 1993). PBL is characterized by collaborative and self-directed learning, where students work in collaborative groups to identify learning needs and engage in self-directed learning to acquire the necessary knowledge, thereby developing effective collaboration skills and autonomy in learning (Hmelo-Silver, 2004; Loyens et al., 2020). The typical PBL process consists of three phases: initial discussion to identify learning issues, self-study to explore these issues, and a reporting phase where findings are shared and evaluated, with teachers guiding the process, particularly in the first and third phases (Loyens et al., 2020).

PBL is consistently reported to significantly enhance motivation and engagement in learning. Research across engineering (Lara-Bercial et al., 2024; Neves & Ribeiro, 2024), language learning and computer science (Al-Bahadli et al., 2023; Kuo et al., 2021; Gu et al., 2025), and K–12 education (Gómez-Pablos et al., 2017; H. Park & Scanlon, 2024) demonstrates that PBL substantially increases students’ motivation, engagement, and interest in learning compared to traditional teaching methods. Meta-analyses and systematic reviews confirm these benefits across diverse educational levels (Bate et al., 2014; Kokotsaki et al., 2016). Qualitative feedback from students across educational contexts indicates that PBL is perceived as enjoyable and meaningful, with students appreciating opportunities to apply knowledge, collaborate, and develop practical skills (Gómez-Pablos et al., 2017; Neves & Ribeiro, 2024; H. Park & Scanlon, 2024).

PBL has been shown to be effective in developing flexible understanding and lifelong learning skills, facilitating conceptual change and improving knowledge retention over time, with students in PBL settings often outperforming those in traditional lecture-based settings in terms of understanding and applying concepts (Capon & Kuhn, 2010; Loyens et al., 2015; Hmelo-Silver, 2004). However, successful implementation of PBL requires adequate teacher training, resources, and strategies to ensure equitable participation and manage group dynamics, with particular emphasis on how ‘problems’ are understood and presented, as this significantly influences the quality and educational outcomes (Alves et al., 2016; Kokotsaki et al., 2016).

2.2. Features of Problem in PBL

In Problem-Based Learning (PBL), the term “problem” encompasses a fundamentally different conceptual framework from traditional educational contexts. Rather than representing a simple question with a predetermined answer, a problem in PBL refers to a complex, multifaceted real-world issue that serves as the central catalyst for student learning and knowledge construction (Hmelo-Silver, 2004; Savery, 2006). These problems are deliberately designed to be open-ended and authentic, reflecting the inherent complexity and ambiguity found in real-world situations where multiple solutions, perspectives, and approaches are not only possible but expected (Ali, 2019).

The pedagogical function of problems in PBL extends beyond mere content delivery to encompass the development of critical cognitive and metacognitive skills. Problems act as learning triggers that prompt students to identify knowledge gaps, engage in self-directed inquiry, and construct understanding through collaborative exploration (Gallagher et al., 1992). This process transforms students from passive recipients of information into active constructors of knowledge, fostering the development of critical thinking, problem-solving abilities, communication skills, and lifelong learning competencies (Ali, 2019; Jaganathan et al., 2024; Akçay, 2009). Furthermore, engagement with complex problems facilitates personal and social development as students navigate challenges collaboratively, experience diverse perspectives, and develop resilience in the face of uncertainty (Takahashi & Saito, 2013).

Through a systematic review of theoretical and empirical literature on PBL problem design, we identified four overarching dimensions that capture the essential characteristics of problems in PBL. This framework constitutes the central theoretical contribution of the present study. The four dimensions were selected based on four criteria: comprehensiveness, theoretical grounding, empirical support, and applicability to school-based PBL reports. These dimensions emerged from synthesizing key frameworks and empirical findings across multiple studies. The extraction process involved analyzing seminal theoretical works (Barrows, 1986; Hmelo-Silver, 2004; Savery, 2006; Schmidt, 1983) and empirical studies examining PBL problem characteristics across diverse educational contexts. We identified recurring themes regarding what makes problems effective in PBL settings.

Authenticity and Contextual Relevance emerged from studies emphasizing real-world connection and situatedness (Hmelo-Silver, 2004; Loyens et al., 2020). Multiple studies demonstrated that authentic, contextualized problems enhance motivation and transfer (Capon & Kuhn, 2010; Nariman & Chrispeels, 2016).

Cognitive Complexity and Learning Outcomes was derived from research on the cognitive demands and conceptual development facilitated by PBL problems (Loyens et al., 2015; Schmidt, 1993). Studies consistently showed that problems must present sufficient complexity and cognitive conflict to promote deep learning (Gallagher et al., 1992).

Social and Collaborative Structure reflects the extensive literature on collaborative knowledge construction in PBL (Bate et al., 2014; Hendarwati et al., 2021). Research demonstrates that effective problems inherently promote interdependence and diverse perspectives (Hmelo-Silver, 2004).

Pedagogical Design and Implementation synthesizes findings on curriculum integration, skill development, and assessment alignment (Savery, 2006; Kokotsaki et al., 2016). Studies emphasized that problems must align with learning objectives and support comprehensive assessment (LaForce et al., 2017).

These four dimensions were selected because they: (a) comprehensively capture the multifaceted nature of PBL problems identified in the literature, (b) are theoretically grounded in constructivist learning principles, (c) have empirical support across diverse educational contexts, and (d) provide actionable categories for analyzing and designing PBL problems. The dimensions are described as follows:

2.2.1. Authenticity and Contextual Relevance

This dimension draws on the long-standing PBL emphasis on authentic, contextually grounded problems, as described by Hmelo-Silver (2004) and Savery (2006), and is further informed by design perspectives that stress meaningful real-world situatedness. Effective PBL problems are fundamentally grounded in authentic situations that reflect genuine societal, professional, or personal challenges (Edwards & Hammer, 2007). The authenticity criterion encompasses several key aspects: problems should be derived from real-world scenarios that students might encounter in their future professional or personal lives, they must be relevant to contemporary societal issues and challenges, and they should be meaningful within students’ current experiential framework while extending their understanding (Nariman & Chrispeels, 2016). This dimension also includes the degree to which problems encourage practical application of theoretical knowledge and facilitate the transfer of learning from academic to real-world contexts (Cahyaningsih et al., 2025).

2.2.2. Cognitive Complexity and Learning Outcomes

This dimension operationalizes the inquiry and reasoning components of Hung’s 3C3R model, as well as the mechanisms of conceptual change described by Loyens et al. (2015). Problems in PBL should possess sufficient cognitive complexity to challenge students’ existing knowledge structures and promote deep learning. This includes the capacity to generate cognitive conflict that disrupts students’ preconceptions and encourages critical analysis and reasoning (Smith et al., 2023). The learning outcomes dimension encompasses the problem’s ability to facilitate active knowledge construction rather than passive information absorption, support the integration of prior knowledge with new learning experiences, and promote the development of both conceptual understanding and procedural skills (Fauzi et al., 2025; Magdalena et al., 2024). Additionally, this dimension considers the extent to which problems develop metacognitive awareness and self-regulated learning capabilities.

2.2.3. Social and Collaborative Structure

This dimension reflects the social-constructivist foundation of PBL emphasized by Hmelo-Silver (2004) and related studies on collaborative knowledge construction in inquiry-based settings. Effective PBL problems inherently require collaborative engagement and cannot be adequately addressed through individual effort alone (Yew & Schmidt, 2012). The collaborative structure encompasses the problem’s capacity to necessitate group discussion, collective inquiry, and distributed problem-solving approaches (Mezak & Pejić Papak, 2019). This dimension also considers how problems facilitate peer interaction, knowledge sharing, and the development of communication and teamwork skills. Furthermore, it addresses the extent to which problems create opportunities for students to assume different roles, contribute diverse perspectives, and engage in constructive discourse that enhances both individual and collective learning outcomes.

2.2.4. Pedagogical Design and Implementation

This dimension synthesizes design principles concerning alignment, facilitation, and assessment found in Savery (2006), constructive alignment approaches, and school-based PBL adaptation studies. It encompasses the careful selection and structuring of problems to align with specific learning objectives and curricular goals (Piñeiro & Vásquez Ortiz, 2019). The pedagogical dimension includes the problem’s capacity to support teacher facilitation rather than direct instruction, allowing educators to guide learning processes while maintaining student agency and inquiry-driven exploration (Mezak & Pejić Papak, 2019). This dimension also considers the degree of openness and structure within problems, ensuring they are sufficiently open-ended to promote genuine inquiry while providing adequate scaffolding for productive learning experiences (Nariman & Chrispeels, 2016; Fitriati et al., 2023). Additionally, it addresses curricular integration, examining how problems support both subject-specific learning goals and cross-disciplinary connections (Cahyaningsih et al., 2025; Chatila & Malaeb, 2025), as well as their capacity to facilitate progressive skill development and accommodate diverse learning needs (Khoirunnisa et al., 2023).

This classification framework reveals that problems in PBL function as multi-dimensional pedagogical tools that simultaneously serve multiple educational purposes. They must achieve a delicate balance between authenticity and instructional effectiveness, complexity and accessibility, individual growth and collaborative learning, and disciplinary depth and interdisciplinary breadth (Utama & Muhammadi, 2023). The effectiveness of problems in PBL contexts depends not only on their individual characteristics within each dimension but also on the synergistic interaction between dimensions. This framework provides educators and researchers with a comprehensive lens for analyzing, designing, and evaluating problems to maximize their potential for promoting meaningful learning experiences and developing the complex competencies required for success in contemporary professional and social contexts.

2.3. Prior Reviews of PBL and Literature Gap

Several systematic reviews have examined PBL implementation and effectiveness. Kokotsaki et al. (2016) reviewed the PBL literature and identified benefits including enhanced student engagement and knowledge retention, while noting implementation challenges. Loyens et al. (2020) synthesized research on PBL’s psychological mechanisms, emphasizing the role of self-directed learning and intrinsic motivation. More recently, Senyah (2024) conducted an integrative review of K–12 teachers’ strategies and challenges in adapting PBL, revealing tensions between standardized curricula and authentic problem-based inquiry.

However, these reviews have primarily focused on PBL outcomes, implementation processes, and pedagogical strategies—not on the systematic analysis of problem characteristics themselves. While scholars acknowledge that “the problem” is the generative engine of PBL (Hmelo-Silver, 2004; Savery, 2006), empirical research systematically analyzing how problems are designed across multiple dimensions remains scarce, particularly in K–12 contexts. Most existing studies examine PBL implementation holistically without disaggregating the specific design features of problems.

At the same time, the broader PBL literature offers several design frameworks that clarify what makes problems educationally productive. Rather than treating these as competing models, this study draws on them as complementary foundations for a unified analytic lens. As summarized in

Table 1. How prior problem-design frameworks informed the present four-dimension model.

Framework/Source	Core Contribution to PBL	Primary Dimension(s)	Related Element(s)
3C3R model (Hung, 2006, 2009)	Integration of content, context, connection, and inquiry processes	Real-world authenticity; Cognitive engagement	Context authenticity; connection to student lives; complexity structure; cognitive conflict level; analysis requirements
Effective PBL case design principles (Ertmer & Simons, 2006; Savery, 2006)	Goal alignment, prior-knowledge activation, and appropriate openness	Cognitive engagement; Pedagogical alignment	Complexity structure; cognitive conflict level; analysis requirements; curriculum integration; skill development range
Authentic learning framework (Herrington & Oliver, 2000)	Real-world relevance, situated tasks, and meaningful constraints	Real-world authenticity	Context authenticity; societal relevance; connection to student lives
Cognitive alignment (Biggs, 1996)	Alignment among outcomes, tasks, and assessment	Pedagogical alignment	Curriculum integration; skill development range; assessment opportunities
Social-constructivist PBL tradition (Hmelo-Silver, 2004; Schmidt, 1983; Savery, 2006)	Collaborative inquiry, knowledge sharing, and perspective taking	Collaborative learning	Group interaction; knowledge construction; perspective integration

, these frameworks collectively informed the development of the present four-dimension, twelve-element model by highlighting authenticity, cognitive demand, collaboration, and alignment.

Taken together, these frameworks suggest that effective PBL problems must be understood not through a single design criterion but through an integrated configuration of authenticity, cognitive demand, collaboration, and pedagogical alignment. Yet, enacted school-based PBL reports are rarely examined through such a unified, problem-centered analytic lens. This is the key literature gap addressed in the present study.

This gap is consequential: without understanding which problem characteristics are prevalent or absent in current practice, and how these characteristics relate to one another, educators lack evidence-based guidance for problem design. For instance, do current PBL problems emphasize authenticity at the expense of cognitive conflict? Do problems with strong collaborative structures also demonstrate high curriculum integration? These questions remain unanswered.

The present study addresses this gap by conducting a systematic analysis of PBL problems in school-based settings, examining their distribution across twelve design elements and exploring relationships among these elements. By focusing specifically on problem characteristics rather than broader implementation factors, this study provides targeted insights to inform more effective problem design practices.

3. Research Method

This systematic review was conducted and reported in accordance with the PRISMA 2020 Statement (Page et al., 2021). The review was not registered, and no publicly accessible protocol is available; therefore, there were no protocol amendments to report. Because this review aimed to characterize problem-design features in published school-based PBL reports rather than estimate intervention effects, we did not perform study risk-of-bias, reporting-bias, or certainty-of-evidence assessments.

3.1. Research Design

This study employed a qualitative content analysis to investigate the characteristics of problems used in PBL, applying a structured analytical framework. The purpose of the analysis was to identify the levels of Real-World Authenticity, Cognitive Engagement, Collaborative Learning, and Pedagogical Alignment embedded within PBL problems and to evaluate their qualitative attributes across multiple dimensions. By systematically classifying and interpreting the problem features, this study aimed to derive insights into how PBL problems are designed and implemented in authentic educational practice settings. The research design was carefully crafted to ensure a comprehensive and rigorous examination of PBL problem characteristics, enabling the identification of key patterns and trends in problem design and application.

3.2. Data Collection

For data collection, a total of 530 academic articles focusing on PBL, published over the past five years (2020–2025) in journals indexed in the Korea Citation Index (KCI), were systematically reviewed. This study specifically focused on the Korea Citation Index (KCI) database to examine PBL problem design within the Korean educational context.

This focus is justified by several considerations. First, Korea has experienced rapid adoption of PBL across K–12 settings over the past decade, driven by educational reforms emphasizing student-centered learning and 21st-century competencies. Analyzing Korean cases provides insight into how PBL problems are conceptualized in an educational system shaped by a centralized national curriculum and, particularly in secondary education, by a high-stakes assessment culture linked to competitive college admission, including the College Scholastic Ability Test (CSAT). These structural conditions may influence problem design by increasing the importance of curriculum alignment, assessment feasibility, instructional efficiency, and academic performance demands. Second, examining context-specific implementation allows for more coherent analysis, as problem design choices reflect not only pedagogical principles but also cultural and systemic factors. While this focus necessarily limits generalizability beyond the Korean context, it enables a deeper understanding of how PBL problems are designed within a specific educational ecosystem. Future research could extend this analysis to other national contexts to identify cross-cultural patterns and context-specific variations in problem design. The scope included empirical studies that documented the application of PBL within actual school settings, encompassing both elementary and secondary education. Studies that focused exclusively on theoretical discussions without practical implementation were excluded.

3.3. Search Strategy and Screening

From the initial pool of 530 articles retrieved from KCI using the English keywords “problem-based learning,” “PBL,” and “project-based learning” (along with their Korean equivalents “문제기반학습” and “프로젝트학습”) published between 2020 and 2025, studies were screened in three stages. The final database search was completed in October 2025. First, titles and abstracts were reviewed to identify studies conducted in school-based settings (n = 156 after exclusion of higher education and theoretical studies). Second, full-text review applied the inclusion criteria, retaining studies that provided detailed problem descriptions with authentic classroom implementation (n = 67). Finally, studies meeting all inclusion criteria and containing sufficient information to code across all twelve design elements were selected, yielding the final sample of 24 studies. The overall study selection process following the PRISMA framework is presented in Figure 1.

Specific inclusion criteria for study selection were established as follows. First, Authentic Classroom Implementation required that studies involved real students in actual school settings, with at least one complete PBL instructional cycle implemented during regular class time or school-based extracurricular activities. Second, PBL Fidelity criteria required that the following core procedures be textually verifiable: (a) problem presentation as a prerequisite, (b) learning issue identification through small group discussions, (c) self-directed learning (exploration and information gathering), and (d) sharing/reporting and reflection phases (including tutor/teacher facilitation). Third, Problem Explicitness required that the reported PBL case clearly specify the following four elements in the main text, appendices, or lesson plans:

• Initial scenario/context: a background situation, commission, or challenge task;
• Constraints and resources: time, materials, tools, or norms;
• Stakeholders/perspectives: relevant actors, multiple viewpoints, or requirements;
• Output and process requirements: expected deliverables, process criteria, or assessment expectations.

Exclusion criteria eliminated studies that contained only declarative statements such as “PBL was applied,” theoretical/conceptual papers that performed only literature reviews, and opinion pieces, columns, or book reviews. Through this systematic selection process, a final dataset of 24 PBL cases was identified, ensuring a representative and diverse sample of PBL problem scenarios across various educational levels and subject areas. All selected cases provide concrete examples of PBL problem situations that were designed and applied within authentic educational environments. The data collection process was meticulously executed to accurately reflect the actual implementation of PBL in educational settings. The final sample comprised 24 studies (Chang, 2024; Ha, 2024; Han et al., 2022; Im et al., 2022; D. Kim, 2021; M. Kim, 2023; D. Kim et al., 2023; J. Kim & Lee, 2022; S. S. Kim & Lee, 2022; S. J. Kim & Lee, 2024; S. Lee, 2020; Y. W. Lee, 2020; J. Lee, 2021; H. W. Lee, 2022; S. H. Lee et al., 2020; A. Lee & Jeong, 2024; Y. J. Lee & Kim, 2021; Lim, 2023; C. S. Park, 2020; H. Park, 2020; Y. K. Park et al., 2023; Shin et al., 2023; Song & Kim, 2022; Weon & Nam, 2023).

3.4. Analytical Framework

To analyze the characteristics of problems used in PBL, an analytical framework was developed based on prior theoretical research and empirical studies on PBL problem design. The twelve sub-dimensions within these four overarching dimensions were operationalized based on specific criteria derived from the literature. For instance, Context Authenticity assesses the extent to which problems reflect real-life situations, ranging from hypothetical scenarios (Level 1) to problems grounded in actual community or professional contexts (Level 3). Societal Relevance evaluates connection to contemporary issues, with higher levels representing problems addressing pressing social, environmental, or economic challenges. Connection to Student Lives measures personal relevance and relatability to students’ immediate experiences.

Similarly, Cognitive Conflict Level captures the degree to which problems challenge existing knowledge frameworks. Level 1 indicates minimal conceptual disruption, while Level 3 represents problems that fundamentally destabilize prior conceptions and require conceptual reorganization. Perspective Integration ranges from single-perspective problems (Level 1) to those explicitly requiring synthesis of multiple disciplinary or stakeholder viewpoints (Level 3).

The decision to use a three-level rating scale (Basic, Intermediate, Advanced) rather than a five-level or continuous scale was made for both theoretical and practical reasons. Theoretically, educational research on learning progressions suggests that three levels adequately capture meaningful developmental distinctions without over-differentiating (Anderson & Krathwohl, 2001). Practically, pilot coding revealed that finer distinctions (e.g., five levels) reduced inter-rater reliability, as coders struggled to differentiate between adjacent levels (e.g., Level 3 vs. Level 4) based on the level of detail typically provided in published descriptions. The three-level scale provided sufficient granularity to detect meaningful variation while maintaining coding reliability (ICC > 0.85 across all dimensions). Additionally, Level 0 was included to capture instances where an element was essentially absent, providing a complete range from absent to advanced implementation. The framework synthesizes insights from the literature on effective problem construction and integrates dimensions critical to ensuring high-quality PBL experiences.

The framework consists of four overarching dimensions, each encompassing several subcategories:

• Real-World Authenticity

• Context Authenticity: The extent to which the problem reflects real-life situations and challenges;
• Societal Relevance: The problem’s connection to contemporary issues and its potential impact on society;
• Connection to Student Lives: The problem’s relatability and relevance to students’ personal experiences and interests.

• Cognitive Engagement

• Complexity Structure: The level of complexity and the presence of multiple interrelated elements within the problem;
• Cognitive Conflict Level: The degree to which the problem challenges students’ existing knowledge and stimulates critical thinking;
• Analysis Requirements: The depth and breadth of analysis required to effectively address the problem.

• Collaborative Learning

• Group Interaction: The extent to which the problem promotes collaborative learning and teamwork;
• Knowledge Construction: The problem’s potential to facilitate the co-construction of knowledge through peer interaction;
• Perspective Integration: The problem’s ability to elicit diverse perspectives and encourage the integration of different viewpoints.

• Pedagogical Alignment

• Curriculum Integration: The problem’s alignment with curricular objectives and learning outcomes;
• Skill Development Range: The variety of skills and competencies that the problem aims to develop;
• Assessment Opportunities: The problem’s potential for formative and summative assessment of student learning.

Each dimension classifies problems across three levels—Basic, Intermediate, and Advanced—based on criteria such as authenticity, complexity, social relevance, and depth of knowledge construction. This framework enables systematic evaluation of each PBL problem’s design features, providing a structured approach to analyzing the qualitative attributes of PBL problems and capturing their multifaceted nature. We used a four-point ordinal rubric (0–3) to balance conceptual resolution with coding reliability under limited reporting detail in published case descriptions.

Table 2. Analytical framework and scoring rubric (excerpt).

Dimension	Element	Level 0	Level 1	Level 2	Level 3
Real-world Authenticity	Context authenticity
	Societal relevance
	Connection to students’ lives
Cognitive complexity and learning outcomes	Complexity structure
	Cognitive conflict level
	Analysis requirements
Collaborative learning	Group interaction
	Knowledge construction
	Perspective integration
Pedagogical design	Curriculum integration
	Skill development range
	Assessment opportunities

Note. (Level 0) indicates that the element was absent or not evident in the problem description. (Level 1) represents minimal implementation, Intermediate (Level 2) indicates moderate implementation, and Advanced (Level 3) represents comprehensive implementation of each element.

presents an excerpted version of the analytical rubric used in coding. The full rubric, including detailed descriptors for all 12 elements across Levels 0–3, is provided in Appendix A.

3.5. Data Analysis

We summarized each element with counts for Levels 0–3 and means (0–3). Pairwise associations among the twelve ordinal variables were estimated using Spearman’s rank correlation coefficient (ρ). To examine higher-order co-occurrence, we identified triads (three elements simultaneously at Level 3).

The extracted 24 PBL cases were systematically coded according to the criteria specified in the analytical framework. Each problem was analyzed across the four dimensions, and its characteristics were classified into the appropriate levels based on its alignment with the framework’s descriptors of authenticity, cognitive challenge, collaboration, and pedagogical implementation. The coding process involved a rigorous and iterative approach to ensure accurate and consistent application of the analytical framework to the PBL cases.

To enhance the credibility of the coding process, five independent coders, all holding doctoral degrees in education and possessing expertise in PBL, participated in the analysis. After an initial training phase to ensure shared understanding of the coding scheme, each coder independently rated the same subset of cases. During the calibration phase, inter-rater reliability for this subset was examined using intraclass correlation coefficients (ICC), which exceeded 0.85 across all dimensions. Discrepancies were addressed through multiple rounds of discussion until full consensus was achieved, ensuring a high level of consistency in the application of the coding criteria.

Following the calibration phase, the complete dataset of 24 PBL cases was coded. Over the two-month coding period, the coders met weekly to resolve ambiguities, discuss challenging cases, and ensure consistent application of the analytical framework. These meetings provided opportunities to compare coding decisions, share insights, clarify interpretations, and reach consensus on the classification of problem characteristics. The collaborative nature of the coding process enhanced the reliability and validity of the analysis.

Once the coding was complete, the final results were synthesized to identify common patterns, strengths, and areas for improvement in the PBL problems implemented in school contexts. Descriptive statistics and correlation analysis were calculated to provide an overview of the distribution of problem characteristics across the four dimensions and their respective levels.

Because the element ratings are ordinal, we used Spearman rank correlations to summarize monotonic associations without assuming interval properties. Triad analysis was added to detect higher-order co-occurrence at the advanced level (Level 3), operationalized as three elements simultaneously reaching Level 3 within the same case. This complements bivariate results by identifying recurrent multi-element configurations that can be interpreted as candidate “design patterns” in this corpus.

4. Results

This study coded 24 PBL articles on a 0–3 ordinal scale (0 = absent, 1 = basic, 2 = intermediate, 3 = advanced) across 12 problem-design elements, yielding a 24 × 12 matrix analysed through descriptive statistics, bivariate correlations, and multivariate (triad) procedures. The distribution of PBL design elements across levels, along with their mean scores and standard deviations, is presented in

Table 3. Distribution of PBL design elements.

Element	Level 0	Level 1	Level 2	Level 3	Mean	SD
Context Authenticity	0	1	12	11	2.42	0.58
Societal Relevance	0	6	3	15	2.38	0.88
Connection to Student Lives	0	5	13	6	2.04	0.69
Complexity Structure	0	2	16	6	2.17	0.56
Cognitive Conflict Level	0	5	17	2	1.88	0.54
Analysis Requirements	0	0	11	13	2.54	0.51
Group Interaction	1	3	6	14	2.38	0.88
Knowledge Construction	1	2	9	12	2.33	0.82
Perspective Integration	1	1	1	21	2.75	0.74
Curriculum Integration	0	15	0	9	1.75	0.99
Skill Development Range	0	1	10	13	2.50	0.59
Assessment Opportunities	0	3	5	16	2.54	0.72

Note. N = 24. Level 0 indicates that the element was absent or not evident in the problem description.

. Descriptively, Perspective Integration recorded the highest mean (M = 2.75), followed by Analysis Requirements (M = 2.54) and Assessment Opportunities (M = 2.54). Context Authenticity averaged 2.42, Societal Relevance 2.38, Group Interaction 2.38, and Knowledge Construction 2.33. In contrast, Curriculum Integration (M = 1.75) and Cognitive Conflict Level (M = 1.88) were lowest, indicating that practice tends to foreground authenticity, collaboration, and process assessment while comparatively under-emphasising high-intensity cognitive conflict and systematic cross-disciplinary integration. Distributional details reinforce this pattern: Context Authenticity reached Level 3 in 11 studies (46%) and Level 2 in 12; Perspective Integration reached Level 3 in 21 studies (88%); Curriculum Integration was Level 1 in 15 studies (63%); and Cognitive Conflict Level reached Level 3 in only 2 studies (8%). To facilitate visual comparison, Figure 2 presents the mean scores of all 12 design elements in descending order.

4.1. Relationships Among Individual Elements

Spearman rank correlations were computed to capture non-parametric associations.

Table 4. Medium-to-strong bivariate correlations (|ρ| ≥ 0.40).

Element 1	Element 2	ρ	p
Context Authenticity	Skill Development Range	0.44	0.031
Context Authenticity	Assessment Opportunities	0.42	0.041
Complexity Structure	Cognitive Conflict Level	0.50	0.013
Complexity Structure	Analysis Requirements	0.45	0.027
Cognitive Conflict Level	Analysis Requirements	0.42	0.041
Cognitive Conflict Level	Knowledge Construction	0.48	0.018
Cognitive Conflict Level	Perspective Integration	0.42	0.041
Cognitive Conflict Level	Curriculum Integration	0.51	0.011
Group Interaction	Knowledge Construction	0.63	0.001
Group Interaction	Perspective Integration	0.54	0.006
Perspective Integration	Assessment Opportunities	0.61	0.002
Perspective Integration	Skill Development Range	0.50	0.013
Assessment Opportunities	Cognitive Conflict Level	0.59	0.002

Note. N = 24. Spearman’s ρ (two-tailed). Only correlations with |ρ| ≥ 0.40 are displayed. All displayed correlations are statistically significant at p < 0.05. For interpretive reference, coefficients of 0.40–0.59 indicate moderate associations and 0.60–0.79 indicate strong associations.

lists all medium-to-strong coefficients (|ρ| ≥ 0.40). Bivariate associations were consistent and strong around collaboration. The Spearman coefficient for Group Interaction–Knowledge Construction was the largest (ρ = 0.63), and Group Interaction–Perspective Integration was also high (ρ = 0.54). Authenticity–competency–assessment linkages were clear: Context Authenticity correlated with Skill Development Range (ρ = 0.44) and with Assessment Opportunities (ρ = 0.42). The coupling of problem complexity and cognitive demand appeared in Complexity Structure–Cognitive Conflict Level (ρ = 0.50) and Complexity Structure–Analysis Requirements (ρ = 0.45), with Cognitive Conflict Level also linked to Analysis Requirements (ρ = 0.42). Cognitive Conflict Level related positively to Perspective Integration (ρ = 0.42) and to Curriculum Integration (ρ = 0.51), suggesting that conflict operates as a hub that pulls both perspective-taking and curricular expansion. Assessment Opportunities likewise showed hub-like properties, correlating with Cognitive Conflict Level (ρ = 0.59), Perspective Integration (ρ = 0.61), and Skill Development Range (ρ = 0.49).

4.2. Multi-Element Patterns

We report triads as the proportion of cases in which all three elements simultaneously reached Level 3. Triad analysis (three elements simultaneously at Level 3) highlighted recurring multi-element constellations. The frequency and characteristics of advanced-level triads are summarized in

Table 5. Frequency of advanced-level triads.

Triad (All Elements at Level 3)	Studies	% of Sample	Common Attributes
Collaboration Triad (GI + KC + PI)	10	41.7	Interdependent teamwork, distributed knowledge building, multi-stakeholder integration; 8 of 10 also had advanced assessment.
Authenticity-Skill Triad (CA + Skill + Assess)	8	33.3	Authentic real-world tasks paired with integrated competence goals and process assessment.
Cognitive-Demand Triad (Comp + Conflict + Analysis)	2	8.3	Highly ill-structured problems, strong dissonance, knowledge creation; one also satisfied the Collaboration Triad.

Note. N = 24. GI = Group Interaction; KC = Knowledge Construction; PI = Perspective Integration; CA = Context Authenticity; Comp = Complexity Structure.

. The Collaboration Triad (Group Interaction + Knowledge Construction + Perspective Integration) appeared in 10 studies (41.7%); among these, 8 studies (80%) also reported Assessment Opportunities at Level 3, indicating that interdependent teamwork and distributed knowledge building are routinely coupled with portfolio/process assessment. The Authenticity–Skill Triad (Context Authenticity + Skill Development Range + Assessment Opportunities) occurred in 8 studies (33.3%). The Cognitive-Demand Triad (Complexity Structure + Cognitive Conflict Level + Analysis Requirements) was rare, appearing in 2 studies (8.3%), suggesting that designs requiring ill-structured problems, strong conceptual dissonance, and knowledge creation in tandem remain uncommon.

5. Discussion and Conclusions

The findings reported above reveal distinct patterns in how PBL problems are designed and implemented in Korean K–12 contexts, as documented in KCI-indexed empirical reports. This section interprets these patterns through the lens of PBL learning theory and existing problem-design frameworks, examines potential explanations for observed distributions and correlations, and proposes evidence-informed design implications.

Treating the problem as the generative engine of PBL, the analyses point to three mutually reinforcing design logics: (a) collaborative knowledge construction, (b) authenticity–competency–assessment integration, and (c) calibrated cognitive challenge. These logics align with canonical accounts of the PBL learning process—complex authentic problems that activate prior knowledge and (when appropriate) productive cognitive conflict, followed by small-group inquiry/self-directed learning and process-oriented assessment—while also clarifying which elements are most consistently foregrounded in the sampled Korean K–12, KCI-indexed reports (Barrows, 1986; Hmelo-Silver, 2004; Schmidt, 1983).

5.1. Interpretation of Problem Design Patterns

The distribution of PBL problems across design elements reveals both strengths and opportunities in current practice. The predominance of high-level Perspective Integration (M = 2.75) and Analysis Requirements (M = 2.54) suggests that Korean PBL implementations successfully create problems that demand multiple viewpoints and deep analytical thinking. This pattern suggests that many Korean K–12 PBL cases are designed to foreground perspective-taking and analytic engagement. However, the relatively low prevalence of advanced Cognitive Conflict (M = 1.88) indicates a potential limitation: problems may engage students in analysis without fundamentally challenging their existing conceptual frameworks. In broad terms, the corpus suggests that school-based PBL problems are more often designed to support manageable, socially mediated inquiry than to provoke disruptive conceptual reorganization. This distinction helps explain why collaboration and authenticity emerged strongly, whereas cognitive conflict remained comparatively weak.

This pattern suggests a “comfortable complexity” phenomenon, where problems are intellectually demanding yet conceptually safe—requiring effort and synthesis but rarely destabilizing prior knowledge in ways that catalyze genuine conceptual change (Loyens et al., 2015). From a learning theory perspective, this is consequential. Conceptual conflict and disequilibrium are central mechanisms for deep learning (Piaget, 1985); problems that avoid cognitive conflict may promote procedural engagement without transformative understanding.

The strong correlations between Context Authenticity and both Skill Development Range (ρ = 0.44) and Assessment Opportunities (ρ = 0.42) suggest that authentic, real-world contexts naturally couple with explicit competency goals and process-based assessment. This finding validates design principles emphasizing authentic contexts as catalysts for comprehensive learning experiences (Hmelo-Silver, 2004). Meanwhile, the robust associations among collaborative elements—Group Interaction with Knowledge Construction (ρ = 0.63) and Perspective Integration (ρ = 0.54)—indicate that collaborative structures tend to cluster together in practice. Conversely, the modest correlations between Cognitive Conflict and other elements suggest that conceptual destabilization may be an independent design consideration—not automatically achieved through authenticity or complexity alone.

The triad analysis further illuminates multi-element design constellations. The Collaboration Triad (Group Interaction + Knowledge Construction + Perspective Integration) appeared in 10 studies (41.7%), indicating that collaborative elements tend to cluster in the sampled reports. The Authenticity–Skill Triad (Context Authenticity + Skill Development Range + Assessment Opportunities) occurred in 8 studies (33.3%), suggesting that many designs couple real-world grounding with explicit competence goals and process assessment. By contrast, the Cognitive-Demand Triad (Complexity Structure + Cognitive Conflict Level + Analysis Requirements) was rare (2 studies; 8.3%), highlighting that designs simultaneously requiring ill-structured complexity, strong conceptual dissonance, and high analytical demands are uncommon in this corpus.

In response to RQ1, the findings indicate that enacted Korean K–12 PBL problems are unevenly distributed across design elements: collaboration, perspective-taking, and assessment are frequently emphasized, whereas curriculum integration and cognitive conflict are less consistently advanced.

5.2. The Cognitive Conflict Design Tension: Balancing Ideals and Constraints

The relative scarcity of advanced Cognitive Conflict Level (M = 1.88; Level 3 in only 8% of cases) requires nuanced interpretation and is best understood as a design tension in school-based PBL rather than as a simple design shortcoming. On one hand, this pattern may reflect deliberate pedagogical calibration rather than design failure. K–12 teachers facing time constraints, diverse student readiness, and curricular pacing demands may intentionally prioritize manageable complexity over conceptual destabilization (Nariman & Chrispeels, 2016). Additionally, cognitive conflict may be developmentally inappropriate for certain grade levels or inconsistent with cultural norms emphasizing harmonious learning environments. From this perspective, lower cognitive conflict represents a pragmatic adaptation to contextual constraints.

However, extensive research demonstrates that conceptual change and durable transfer require cognitive conflict—the deliberate disruption and reconstruction of prior knowledge structures (Capon & Kuhn, 2010; Hmelo-Silver, 2004; Loyens et al., 2015). Problems that engage students in analysis without fundamentally challenging their conceptual frameworks may produce procedural competence without transformative understanding. Thus, while acknowledging the practical realities teachers face, we interpret cognitive conflict as a consequential design tension in PBL: insufficient conflict may limit deep conceptual change, whereas excessive or poorly calibrated conflict may undermine feasibility, developmental appropriateness, or classroom stability.

Our data cannot definitively distinguish whether low Cognitive Conflict represents a design limitation or a context-appropriate choice. Future research should investigate teachers’ decision-making processes regarding cognitive conflict in problem design, examining how they balance conceptual challenge with practical, developmental, and cultural considerations. Such research could clarify whether and when increasing cognitive conflict is beneficial across different educational contexts.

5.3. Design Implications

To address the tension between pedagogical ideals and practical constraints, we propose modular scaffolds for problem design. These may include problem-framing templates, graduated “conflict ladders,” peer-feedback protocols, and staged reflection prompts that help teachers increase conceptual challenge while maintaining cognitive and socio-emotional support. For example, in a sustainability-themed PBL unit, a conflict ladder might move from identifying local stakeholders (Level 1), to comparing conflicting evidence (Level 2), to proposing solutions that challenge taken-for-granted assumptions about convenience and responsibility (Level 3). These tools would allow teachers to calibrate conflict intensity based on context while preserving the essential role of disequilibrium in knowledge construction.

A second design tension emerged regarding curriculum integration. Our analysis indicates that as curricula advance toward broad thematic integration, connections to students’ lived experiences may attenuate. Consequently, interdisciplinary projects would benefit from preliminary anchor activities (e.g., field inquiries, stakeholder interviews) that connect rigorous disciplinary work to local contexts and meanings. For instance, before beginning a broader sustainability project, students might photograph examples of waste, recycling, or resource use in their own homes, school, or neighborhoods and use these observations to generate initial problem questions.

Collectively, these findings suggest that a powerful problem transcends mere authenticity or complexity; it must be simultaneously authentic, competency-explicit, conflict-calibrated, collaboration-dependent, and transparently assessed. In response to RQ2, the co-occurrence analyses show that problem-design elements do not combine randomly. Rather, they form recurring configurations centered on collaboration and authenticity, while cognitively demanding configurations remain relatively uncommon.

5.4. Limitations and Contextual Boundaries

These findings should be interpreted within the constraints of our sample and methodology. Our analysis is based on a bounded corpus of 24 PBL cases drawn exclusively from a single database, namely KCI-indexed Korean publications, representing school-based implementations in a specific educational context. The Korean educational system is characterized by strong national curriculum standards, centralized assessment, and a cultural emphasis on academic achievement—factors that may shape problem design in ways distinct from other national contexts. Therefore, while our findings illuminate problem design patterns in Korean K–12 PBL, they should not be assumed to represent global PBL practice.

Additionally, our data rely on published descriptions of PBL problems, which may reflect reporting bias (e.g., authors may emphasize certain design features over others) and may not fully capture the enacted complexity of problems as experienced by students in classrooms. Observational studies or analyses of actual problem materials could reveal nuances not evident in published descriptions. Relatedly, the correlational and triadic findings should be interpreted as exploratory pattern-detection results within a restricted corpus rather than as robust inferential evidence for general PBL design principles. Despite these limitations, our findings provide valuable insights into problem design patterns within a significant PBL implementation context and establish a methodology for analyzing problem characteristics that can be applied to other educational settings.

5.5. Conclusions

This review found that Korean K–12 school-based PBL problems were more likely to emphasize authenticity, collaboration, and process assessment than strong cognitive conflict or grounded interdisciplinarity. This design landscape helps explain both the motivational strengths often associated with PBL and the variability of its deeper conceptual effects. By conceptualizing problem framing as an integrated design ecosystem, educators and school leaders may be better able to design problems that more coherently connect authenticity, conceptual challenge, collaboration, and process-oriented assessment. In this sense, aligned school-level supports may help PBL move closer to a fuller design sequence in which authentic problems, productive cognitive conflict, collaborative knowledge construction, and assessment are more intentionally linked. More balanced problem design may better equip students to engage with the multifaceted civic, scientific, and professional challenges that contemporary education increasingly seeks to address.

This synthesis is limited by its focus on KCI-indexed Korean K–12 studies, the relatively small corpus of 24 cases, possible reporting bias in the primary articles, and the absence of common outcome measures that would permit meta-analytic modeling. Future research should examine how specific problem profiles relate to learning processes and outcomes, including through studies that experimentally vary conflict intensity, interdisciplinarity, or collaboration structure.