Theoretical Framework and Research Programme: Epistemic curiosity, conceptualized as the desire for new knowledge that motivates exploratory behavior, is widely recognized as a fundamental driver of academic achievement and psychological well-being (Kashdan et al., 2020; von Stumm et al., 2011). We adopt the multidimensional perspective (Litman & Pezzo 2007; Kashdan et al., 2018), distinguishing between Joyful Exploration—the pleasurable seeking of novelty—and Deprivation Sensitivity, the drive to resolve information gaps. Despite its documented importance, the environmental mechanisms that affect curiosity during the critical transition through lower secondary education remain significantly under-studied.

Our investigation integrates three complementary frameworks to elucidate the dynamics of engagement. Self-Determination Theory (SDT; Deci & Ryan, 2000) provides the foundational premise that intrinsic motivation is contingent upon the satisfaction of basic psychological needs, specifically Autonomy and Competence. However, in learner-centered curricula that prioritize high autonomy, the learner is inevitably confronted with the uncertainty of unstructured problems. Appraisal Theory (Silvia, 2006) refines this perspective by specifying that interest is not precipitated by novelty alone, but requires a secondary appraisal of Coping Potential—the subjective belief that one possesses the resources to understand and manage the incoming complexity.

This appraisal mechanism is critical when viewed through the lens of Hidi and Renninger’s (2006) Four-Phase Model of Interest Development. While established Individual Interest (Phases 3 and 4) is characterized by a self-sustaining resilience to challenge, the earlier phases of Situational Interest are fragile and environmentally dependent. We posit that for a student to successfully transition from environmentally triggered curiosity to a stable, self-regulating disposition, they must possess the specific competence to navigate failure without ego-threat.

Analytical Strategy: From Exploration to Confirmation To address the complexity of the data while ensuring generalizability, we follow the Model Generation strategy as defined by Kline (2023). This framework acknowledges that in the absence of a fully established theory, an initial model is often tentative and requires respecification based on empirical fit (e.g., modification indices, AIC/BIC).

However, Kline (2023) explicitly cautions that such data-driven modifications increase the risk of overfitting. Consequently, any model derived from this exploratory phase must be treated as a hypothesis requiring replication on an independent sample.

Accordingly, the research follows a multi-phase design:

• Model Generation (Current Study - 2025): Application of the strategy to identify the optimal structural architecture. The resulting model (M2025) is formalized as a new hypothesis.

• Confirmatory Testing (2026): M2025 will be strictly tested against the subsequent data collection, fulfilling the requirement for cross-validation on an independent time-point.

Crucially, this design allows for iterative expansion: upon successful confirmation of the M2025 parameters, the 2026 analysis will incorporate additional constructs to generate a new, more comprehensive hypothesis (M2026), continuing the cycle of inquiry.

Ecological Setting: The research was conducted within the ScioSchools network, a system of approximately 20 schools in the Czech Republic. This setting provides a unique ecological niche for investigating epistemic curiosity due to three specific methodological affordances.

First, the network acts as a "High-Autonomy" Baseline. Unlike traditional systems where curiosity is often suppressed by high-stakes testing and rigid grading, ScioSchools operate without numerical grades, prioritizing student self-direction and project-based learning. By controlling for these external suppressors, we can isolate the subtle intra-psychological mechanisms that sustain engagement when extrinsic coercion is absent.
Second, the environment is structurally stable yet diverse. While all schools share a unified pedagogical philosophy emphasizing psychological safety and "preparation for an unpredictable world," they exhibit naturally occurring variability in classroom organization.
Finally, access to the network allows for reliable longitudinal tracking. This enables the analysis of developmental trajectories, moving beyond cross-sectional snapshots to observe how specific environmental factors actively shape epistemic curiosity over time. 

Participants and Procedure Data were collected via a two-wave longitudinal design (Wave 1: May 2024; Wave 2: May 2025) within the ScioSchools network. The sample comprised adolescents spanning Grades 4–9 (N = 305; nfemale = 146, nmale = 159). The mean age was 12.5 years (SD = 1.37). Participation was voluntary and conditional on informed parental consent. Measures We mapped the relationship between the pedagogical environment and learner outcomes using 5-point Likert scales. Internal consistency was verified using McDonald’s Omega (ω). Epistemic Curiosity: We utilized the Joyous Exploration subscale of the Five-Dimensional Curiosity Scale (Kashdan et al., 2020), operationalized as the intrinsic drive to seek novel information. Perceived School Pedagogy: A self-developed inventory was designed to capture the specific affordances of the ScioSchool curriculum. We modeled a single latent factor integrating Task Design, Divergence Support, and Student Autonomy. Strong inter-factor covariance (r = .63–.76) confirmed that these elements function as a cohesive 'autonomy-supportive' system. Positive Error Orientation (Mediator): Assessed via a self-developed scale measuring the cognitive capacity to reframe failure as information, serving as a direct proxy for the school’s "learning from mistakes" competency. Controls: Family Support, General Well-being (WHO-5; Bech, 2004), and Age. Data Analysis We employed Structural Equation Modeling (SEM) using the lavaan package in R. To account for potential non-normality and attrition, we utilized the MLR estimator in conjunction with Full Information Maximum Likelihood. The analysis utilized a "competing models" protocol to ensure the validity of the findings. We formally compared three hypothetical structures: the Saturated Model, the Partial Mediation Model, and the Full Mediation Model. Models were compared using the Akaike Information Criterion (AIC) and Chi-Square difference tests to identify the most parsimonious fit. To rule out alternative explanations, we conducted a series of stress tests: Discriminant Validity: We verified that Positive Error Orientation and Curiosity are empirically distinct constructs. Reverse Causality: We tested an alternative model where Curiosity precedes Error Orientation. The poor fit of this reverse model confirmed the directionality of our findings. Common Method Bias: We controlled for General Well-being (WHO-5) to ensure that the observed relationships were not artifacts of students’ general mood or response tendencies. Independence Check: The low Intraclass Correlation Coefficient ($ICC < .05$) indicated negligible school-level clustering, confirming that the results reflect individual student development rather than classroom-specific effects. Finally, we employed Multi-Group SEM to test whether this psychological mechanism functions universally or varies by gender.

Structural Equation Modeling (SEM) identified Positive Error Orientation (PEO) as the sole mediator of environmental influence. Direct paths from School Pedagogy and Family Support were non-significant (Full Mediation), indicating that autonomy support promotes engagement only when channeled through this specific competence. Additionally, the analysis revealed a "broken inertia" effect: baseline curiosity (2024) failed to predict outcome curiosity (2025) after controlling for the mediator, challenging trait-stability models. Multi-group analysis demonstrated a gender asymmetry: while the mechanism was universal, the path coefficient from PEO to Curiosity was significantly stronger for girls (β=.87) than for boys (β =.72). Integrating Self-Determination Theory (Deci & Ryan, 2000) with Silvia’s (2006) Appraisal Theory, we posit that PEO bridges the gap between environmental affordances and cognitive needs. While the curriculum provides Autonomy (via novelty and choice), PEO supplies the requisite Coping Potential—the subjective appraisal that the resulting complexity is manageable. Without the capacity to reframe failure as information, the uncertainty inherent in autonomous learning triggers a low-coping appraisal. This precipitates a shift from Interest to Anxiety, effectively thwarting the need for Competence. Thus, PEO functions as the affective gatekeeper, determining whether the freedom to explore is perceived as a masterable challenge or an unmanageable threat. Consequently, to ensure that learner autonomy effectively translates into epistemic engagement, curriculum design must focus on establishing "error-positive" learning environments. The objective is to equip students with the competence to own and metabolize mistakes into forward-pointing datapoints. This reframing proves even more critical for female students, serving as the decisive lever for sustained engagement. Ultimately, such curricular architecture nurtures epistemic curiosity, positioning it not only as a primary driver of intrinsic motivation but as a fundamental proxy for openness to the world.

Bech, P. (2004). Clinical psychometrics. Wiley-Blackwell. Hidi, S., & Renninger, K. A. (2006). The Four-Phase Model of Interest Development. Educational Psychologist, 41(2), 111–127. https://doi.org/10.1207/s15326985ep4102_4 Kashdan, T. B., Disabato, D. J., Goodman, F. R., & McKnight, P. E. (2020). The Five-Dimensional Curiosity Scale Revised (5DCR): Briefer subscales while separating overt and covert social curiosity. Personality and Individual Differences, 157, Article 109836. https://doi.org/10.1016/j.paid.2020.109836 Kline, R. B. (2023). Principles and practice of structural equation modeling (5th ed.). The Guilford Press. Litman, J. A., & Pezzo, M. V. (2007). Dimensionality of interpersonal curiosity. Personality and Individual Differences, 43(6), 1448–1459. https://doi.org/10.1016/j.paid.2007.04.021 Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78. https://doi.org/10.1037/0003-066X.55.1.68 von Stumm, S.; Hell, B.; Chamorro-Premuzic. (2011) T. The hungry mind: Intellectual curiosity is the third pillar of academic performance. Perspect. Psychol. Sci., 6, 574–588. https://doi.org/10.1177/1745691611421204 Silvia, P. J. (2006). Exploring the psychology of interest. Oxford University Press.