Journal of Intelligence doi: 10.3390/jintelligence14090216

Authors: Jiawei Xiong Cheng Tang Qidi Liu Feiming Li

The accurate measurement of individual cognitive abilities in educational settings increasingly relies on personalized assessment systems that adapt to the learner’s current standing on the measured ability. These systems generate and use assessment items continuously and in real-time, bypassing the field-test cycle that traditional calibration requires. Item banks therefore contain a growing share of items whose parameters must be predicted first, and each test form must be assembled in real time to match the learner’s current ability level and content needs for effective personalized instruction. This study proposes an Ising framework that addresses both requirements by predicting item parameters from text with calibrated uncertainty estimates and assembling personalized test forms within the time budget of the assessment loop. This framework is validated through a factorial simulation and an empirical study on state-level English Language Arts items across Grades 3 through 12. In the simulation, the proposed uncertainty-aware method achieves a mean test information function (TIF) gap of 0.226, compared with 1.171 for the conventional baseline that ignores prediction uncertainty. The proposed method also produces feasible forms across the full operational region, whereas the chance-constrained and robust baselines become infeasible at high prediction error combined with a high share of predicted items. In the empirical study, predictive intervals achieve coverage within ten percentage points of the nominal level in the pooled condition, and the uncertainty-aware assembly reduces the TIF gap relative to the baseline, with the largest gains at the extremes of the ability distribution where accurate measurement matters most for instructional placement decisions. By separating prediction uncertainty from true individual differences in ability, this framework preserves the interpretability of cognitive ability estimates and provides a foundation for personalized assessment systems that support differentiated instruction under continuous item bank growth.

Journal of Intelligence doi: 10.3390/jintelligence14090215

Authors: Zhen Qiang Hong Yi Zhuo Wang

Critical thinking (CT) is a core competence that schools are expected to cultivate, yet whether it reaches classrooms depends on teachers’ readiness to teach it. Using an explanatory sequential mixed methods design, we surveyed 312 primary English-as-a-foreign-language teachers in Shandong Province, China, across seven dimensions of readiness and interviewed five teachers. Latent profile analysis identified four profiles organized by two weakly related dimensions—the capacity teachers bring to CT teaching and the constraints they perceive around them: Empowered-Engaged (15%), Willing-but-Constrained (28%), Moderate-Unobstructed (20%), and Committed-but-Under-resourced (37%). Endorsement of CT was near-universal, so the profiles differed in capacity and circumstance rather than in commitment. The central finding is a disconnection between what teachers believe and what they report doing: Willing-but-Constrained teachers wanted further professional learning as much as their Empowered-Engaged peers did, yet reported teaching CT less often. The divergence held under methods that account for classification uncertainty and under alternative measurement specifications, though not under all of them, and profile membership was largely unrelated to teachers’ background characteristics. The interview findings help explain why positive beliefs about CT did not always correspond to classroom practice. Teachers reported limited CT-specific training, examination pressure, time constraints, and norms that discouraged questioning, despite generally recognizing the value of CT. This suggests that the main barriers to CT instruction may lie not in teachers’ attitudes toward CT, but in their opportunities, preparation, and working conditions for implementing it.

Journal of Intelligence doi: 10.3390/jintelligence14090214

Authors: Ibrahim Yikilmaz Lutfi Surucu Mustafa Bekmezci Bulent Cetinkaya Kemal Erogluer

Emotional intelligence is the capacity to understand, regulate, and manage one’s emotions, and it has significant effects on employees’ psychological adjustment, interpersonal relationships, and job performance. However, the specific psychological mechanisms by which emotional intelligence influences employees’ creative performance, and the conditions under which this relationship is strengthened, remain not fully understood. The aim of this study is to examine the mediating role of psychological well-being and the moderating role of cultural intelligence in the relationship between emotional intelligence and creative performance among individuals working in the information and communication technology (ICT) sector. Research data were obtained from 409 participants working in large-scale information and communication technology companies operating in Turkey, and the research hypotheses were tested using the bootstrap method with the Hayes PROCESS Macro. The findings show that emotional intelligence is positively related to psychological well-being and creative performance. Furthermore, psychological well-being significantly mediates the relationship between emotional intelligence and creative performance. The results show that cultural intelligence strengthens this indirect effect, and the contribution of emotional intelligence to creative performance is more pronounced at higher levels of cultural intelligence. The findings indicate that emotional intelligence is associated with creative performance both directly and indirectly through psychological well-being. However, the strength of this indirect association appears to vary across employees’ levels of cultural intelligence, which is considered an individual capability rather than a characteristic of the work environment. In this respect, the study contributes to the literature by examining psychological well-being as an explanatory pathway and cultural intelligence as an individual-level boundary condition in the association between emotional intelligence and creative performance.

Journal of Intelligence doi: 10.3390/jintelligence14090213

Authors: Yihui Liu Jinming Sun Weida Zhang

Generative artificial intelligence (GenAI) is increasingly embedded in higher education, but the association between students’ efficacy beliefs for AI-supported learning and their self-perceived digital creative functioning remains under-specified. This cross-sectional study tested a domain-specific efficacy account and examined a broad, study-specific AI learning risk-awareness measure as an exploratory boundary condition. Survey data from 920 Chinese higher education students were analyzed using confirmatory factor analysis and regression-based conditional process analysis with 5000 bootstrap resamples. The four focal measures showed a statistically distinguishable four-factor structure, although digital creative self-efficacy and self-perceived digital creative functioning remained conceptually close. AI learning self-efficacy was positively associated with self-perceived digital creative functioning, with a smaller statistical indirect association through digital creative self-efficacy. At mean risk awareness, the model-implied indirect point estimate was 0.177, compared with a direct association of 0.612. Conditional indirect point estimates decreased modestly from 0.196 to 0.157 as risk awareness increased. The conventional index of moderated mediation was negative and small and is interpreted here only as a statistical index of change in the conditional indirect association. These findings are consistent with a domain-specific efficacy account while indicating that the indirect component was meaningful but non-dominant. Single-wave self-reports preclude causal inference and do not constitute evidence of objectively rated creativity.

Journal of Intelligence doi: 10.3390/jintelligence14090212

Authors: Chulmin Kim Seongjin Ahn

Large language model (LLM)-based artificial intelligence is increasingly used in ethically consequential human decision-making, yet fully autonomous machine ethics remains unrealistic, motivating architectures that support rather than replace human ethical judgment. This study introduces the Cognitive–Reflective Equilibration Architecture (CREA), a cognitively grounded artificial moral advisor that operationalizes the Cognitive–Reflective Equilibration Model (CREM), in which reflective reasoning guides ethical judgment from intuitive cognition toward a more advanced equilibrium among competing values, drawing on Piaget and Rawls. CREA implements CREM’s 20-step process through four stage-aligned reasoning agents—Cognitive, Reflective, Equilibration, and Evaluation—coordinated via multi-LLM orchestration, in which auxiliary models independently explore principles, generate counterarguments, and score supporting and opposing considerations to externalize reflective deliberation. The architecture was empirically evaluated by comparing four configurations—single-agent, multi-agent, multi-LLM, and multi-LLM with knowledge- and reasoning-bank augmentation—across four indicators of advice quality using 500 matched execution units per configuration. All comparisons are system-internal: advice quality was scored by CREA’s own multi-LLM measurement pipeline rather than by human ethicists, so the findings reflect relative differences among architectures under LLM-based self-evaluation, not normative validity. Within that scope, distributing reflective reasoning across multiple models was associated with higher reason-giving (justifiability) and normative-alignment scores relative to simpler configurations. CREA therefore offers an empirically characterized, auditable advisor architecture whose potential to scaffold human ethical judgment remains a hypothesis for user-centered validation rather than a demonstrated outcome.

Journal of Intelligence doi: 10.3390/jintelligence14090211

Authors: Laura Macchi Manuela Turco Daniele Inglese Laura Caravona

Recent studies suggest forms of unconscious processing in classical insight problem solving. In particular, the unconscious analytic thought (UAT) approach speculates that the creative act of restructuring such very demanding problems implies a form of high-level unconscious thought, mainly active during incubation. The present study aims to investigate whether pupillary dilation, typically associated with conscious cognitive effort, can also represent an indicator of unconscious cognitive effort (UAT). In the experimental condition, participants attempted to solve the problem and, during the incubation phase, performed a series of single-digit arithmetic calculations. In the control condition, participants only performed the arithmetic calculations. Results showed a significant interaction between group (solvers, non-solvers and control) and temporal course on pupillary dilation. In particular, solvers, conversely to the other two groups, showed high and constant pupillary dilation during all the incubation phase, suggesting the maintenance of constant cognitive effort at an unconscious layer not attributable to anything other than the processing of the insight problem solution. In the other two groups, instead, the pupil size—starting at the same level as solvers—progressively decreased during the execution of the arithmetic task. Our results are discussed in the light of the debate on dual process theories.

Journal of Intelligence doi: 10.3390/jintelligence14090210

Authors: Ziyaeddin Halid İpek Ferudun Sezgin

This study aimed to explore how mindfulness is perceived and experienced by teachers and to develop and validate the Mindfulness in School Scale (MSS) for use in school settings. An exploratory sequential mixed methods design was employed. In the qualitative phase, phenomenological interviews were conducted with 23 teachers selected through purposive maximum variation sampling. Content analysis identified six themes: Self-Observation, Environmental Observation, Self-Awareness, Environmental Awareness, Acceptance, and Reactive Control. These findings informed the development of the MSS. In the quantitative phase, the scale was administered to 663 teachers. Exploratory and Confirmatory Factor Analyses supported a five-dimensional structure comprising Self-Observation, Environmental Observation, Self-Awareness, Environmental Awareness, and Acceptance, explaining 51.46% of the total variance. Standardized CFA factor loadings ranged from 0.50 to 0.90, and the measurement model demonstrated satisfactory fit (χ2/df = 1.20, RMSEA = 0.028, CFI = 0.984, TLI = 0.981). The five dimensions also demonstrated satisfactory internal consistency, with Cronbach’s alpha coefficients ranging from 0.80 to 0.93. Correlations among the dimensions were generally low to moderate, indicating that the MSS captures related but distinct aspects of mindfulness in school settings rather than a single underlying construct. The findings further suggest that mindfulness in schools is shaped by both intrapersonal and interpersonal processes, with environmental awareness emerging as a contextually salient dimension. Overall, the MSS provides initial evidence for the validity and reliability of its five dimensions as a multidimensional instrument for assessing mindfulness among teachers in school settings while offering a context-specific framework for future research examining the relationship between mindfulness and emotional intelligence.

Journal of Intelligence doi: 10.3390/jintelligence14090209

Authors: Sevinç Zeynep Kavruk Ahmet Bildiren

The Naglieri Nonverbal Ability Test–Individual Form (NNAT-I) is widely used to assess nonverbal reasoning while minimizing language demands, yet evidence regarding its psychometric properties in Turkish middle school students is limited. This study examined the reliability, convergent validity, internal structure, item functioning, and differential item functioning (DIF) of the NNAT-I in 1130 students aged 10–14 years attending schools across socioeconomic development strata in Aydın, Türkiye. Internal consistency, test–retest and alternate-form reliability, convergent validity, confirmatory factor analysis (CFA), Rasch analyses, multiple-group Rasch DIF analyses, and linear mixed-effects modeling were conducted. The NNAT-I demonstrated high internal consistency (Cronbach’s α = 0.93; McDonald’s ω = 0.94), strong test–retest and alternate-form reliability, and strong correlations with the Test of Nonverbal Intelligence–Third Edition (TONI-3) and Raven’s Standard Progressive Matrices (RSPM). CFA provided qualified rather than unequivocal evidence for a predominantly unidimensional structure, with model fit varying according to item response distributions and the treatment of structurally nonadministered responses. Rasch analyses of the calibrated items supported an ordered hierarchy of item difficulty, with all items showing Infit values within the prespecified descriptive reference range. After adjustment for multiple comparisons, DIF analyses identified significant gender-related DIF for only one item, while no items showed significant age-related DIF. Linear mixed-effects modeling showed that grade level and gender were significantly associated with NNAT-I scores, whereas chronological age was not independently associated with performance. These findings provide evidence regarding the reliability, convergent validity, internal structure, and item functioning of the NNAT-I in this Turkish middle-school sample, while also indicating that conclusions regarding unidimensionality should remain qualified. The findings underscore the importance of local psychometric evaluation before applying nonverbal measures in new educational and cultural contexts.

Journal of Intelligence doi: 10.3390/jintelligence14090208

Authors: Shuangshuang Li Jin Tian Mengting Chen Ziyue Wang

The rapid integration of generative artificial intelligence (GenAI) into educational settings has raised growing concerns regarding its potential influence on students’ higher-order thinking skills. However, existing findings remain inconclusive, and the role of individual factors in shaping these effects has received limited attention. Drawing on self-regulated learning theory, this study examined the relationships between GenAI usage and three higher-order thinking skills, critical thinking, creative thinking, and computational thinking, and investigated the moderating role of self-regulated learning. The results indicated that the associations between GenAI use and higher-order thinking skills varied across levels of self-regulated learning. Specifically, among students with lower levels of self-regulated learning, GenAI use was negatively associated with critical thinking, creative thinking, and computational thinking. In contrast, among students with higher levels of self-regulated learning, GenAI use was positively associated with critical thinking and computational thinking, while its association with creative thinking was not significant. Furthermore, self-regulated learning was positively associated with all three higher-order thinking skills and significantly moderated the relationships between GenAI use and critical thinking, creative thinking, and computational thinking. These findings highlight the complex cognitive consequences of GenAI usage and underscore the importance of self-regulated learning in shaping students’ cognitive outcomes in AI-assisted learning environments. The study contributes to a more nuanced understanding of how GenAI influences higher-order thinking and provides implications for promoting effective and responsible AI use in education.

Journal of Intelligence doi: 10.3390/jintelligence14090207

Authors: Faye Antoniou Mohammed Alghamdi

Assessed creative thinking is a cognitive individual-difference construct—related to, but not equivalent to, conventional intelligence—that PISA 2022 measured directly at scale. We asked how its associated resources are structured, and whether discrete student “types” or continuous dimensions better predict creative-thinking performance. Using 44 education systems (306,704 students; complete-indicator sample 235,240), we estimated senate-weighted Gaussian mixtures of nine within-system standardized creativity indicators and related them to performance (10 plausible values; Fay balanced repeated replication). A six-profile solution was reproducible given adequate optimization (mean adjusted Rand index 0.95 across 12 subsamples), yet no class count was selected—information criteria improved monotonically from 2 to 12 classes—and the profiles were predictively redundant: in leakage-safe cross-validation they explained less variance (R2 = 0.028) than the nine continuous indicators (0.044) or a flexible continuous model (0.068) across K = 2–12 and under hard or soft classification. After a validated classification-error correction, performance-proximal dispositions and curiosity, more than relational perceptions, predicted creative-thinking performance. For measuring and modeling creative ability, continuous resource dimensions were more informative than discrete types, which amounted to an enumeratively underdetermined, predictively redundant re-description of continuous variation.

Journal of Intelligence doi: 10.3390/jintelligence14090206

Authors: Helene M. von Gugelberg

Matrix reasoning performance is often treated as a static indicator of cognitive ability, although it reflects a dynamic process that unfolds across items. The present study examined how reasoning ability and visuospatial working memory (vsWM) interact with item difficulty and item position to shape problem-solving behavior. Eye movements were recorded from 300 participants during a figural matrices test. Three indices of problem-solving behavior were analyzed using Bayesian multilevel models: toggle rate (ToR), proportional time on matrix (PoM), and proportional time to first fixation on response alternatives (PoFA). Results showed that higher reasoning was consistently associated with behavior indicative of constructive matching, reflected in lower ToR, higher PoM, and higher PoFA. Further, individuals with higher reasoning ability showed stronger adaptations in problem-solving behavior in response to changing task demands. Evidence regarding vsWM was more nuanced, revealing selective interactions with reasoning ability and test characteristics rather than a consistent main effect across measures. Item position revealed substantial between-person variability, which exceeded the variability associated with item difficulty. This suggests that individuals differ more strongly in how their behavior evolves across the test than in their responses to increasing difficulty. Overall, the findings highlight that problem-solving behavior in reasoning tests reflects a dynamic interplay between individual characteristics and test characteristics, emphasizing the importance of modeling both simultaneously when investigating reasoning performance.

Journal of Intelligence doi: 10.3390/jintelligence14090205

Authors: Mauricio Barramuño-Medina Pablo Valdés-Badilla Pablo Aravena-Sagardia Jordan Hernandez-Martínez Edgar Vásquez-Carrasco Esteban Fariña-Hermosilla Wilson Pastén-Hidalgo Germán Gálvez-García

This study aimed to examine within-subject changes associated with a 12-week multicomponent training program combined with cognitive stimulation in cognitive–motor performance in older women, and to evaluate changes in cognitive status and health-related quality of life (HRQoL). Thirty-three participants aged 70.24 (4.32) years were assessed at baseline, 6 weeks, and 12 weeks. Cognitive–motor performance was evaluated using a Go/No-Go task, including reaction time (RT), omission errors, and commission errors. Cognitive status was assessed using the Memory, Fluency, and Orientation (MEFO) score, and HRQoL was evaluated with the 36-Item Short Form Health Survey (SF-36). Mixed-effects models showed a significant reduction in RT (Δ = 51.5 ms, p < 0.001; d = 0.50) and omission errors (OR = 0.04, p < 0.001), and an increase in commission errors (OR = 2.06, p = 0.008). Cognitive status improved significantly (Δ = 0.45, p = 0.021; d = 0.34), whereas HRQoL improved only in the general health dimension (Δ = 6.82, p = 0.004; d = 0.53). Overall, the intervention was associated with a reorganization of Go/No-Go performance, characterized by faster responding, but increased commission errors, indicating no improvement in motor inhibition. These findings suggest that this multidomain approach is feasible and warrants evaluation in randomized controlled trials.

Journal of Intelligence doi: 10.3390/jintelligence14090204

Authors: Difei Jia Xi Chen Yunsong Wang

Generative artificial intelligence (GenAI) has expanded opportunities for English learning beyond formal classroom settings. From an intelligence perspective, speaking in English as a foreign language (EFL) represents an applied domain in which learners must adaptively retrieve, integrate, monitor, and deploy linguistic knowledge under communicative demands. The present study examined the effects of a pedagogically guided form of AI-mediated informal digital learning of English (AI-IDLE) on Chinese university EFL learners’ speaking ability, speaking anxiety, and speaking enjoyment. To this end, 89 Chinese university EFL learners participated in a 12-week intervention and were assigned to an experimental group (EG) or a control group (CG). Pre- and post-intervention evaluations were administered to assess the results, including standardized speaking skills and valid questionnaires regarding enjoyment and speaking anxiety. The English-speaking skills test and the two questionnaires were administered at post-test and again 10 weeks after the intervention as the delayed post-test. Mixed-effects modeling (MEM) was used, and the findings showed that the EG demonstrated significantly greater improvements in speaking ability and enjoyment and a greater reduction in speaking anxiety than the CG, with these between-group advantages evident at the post-test. Overall, the findings suggest that structured AI-IDLE can provide a digitally mediated context for intelligence-relevant adaptive learning by supporting communicative performance while fostering affective conditions conducive to its development. These findings have implications for understanding how digitally mediated informal learning environments may support adaptive intelligence in applied language-learning contexts.

Journal of Intelligence doi: 10.3390/jintelligence14090203

Authors: Tikva Ovadiya

Teacher learning presents a unique cognitive challenge: educators must simultaneously develop mathematical expertise and pedagogical knowledge for teaching. This study examines how interactive worked examples function as cognitive tools to support this dual learning process. Using a qualitative, exploratory case-study design, I investigated how 30 in-service mathematics teachers engaged with interactive worked examples of non-routine problems in a professional development setting designed to reduce extraneous cognitive load while promoting germane processing. Drawing on Cognitive Load Theory, expertise development frameworks, and recent perspectives on cognitive offloading and affective engagement in multimedia learning, I analyzed the cognitive and emotional mechanisms underlying teachers’ acquisition of both mathematical content knowledge and pedagogical content knowledge. Results from teachers’ journals, reflective protocols, and self-assessed cognitive load measurements reveal three key findings: First, teachers experienced analyzing worked examples not as passive observation but as active problem-solving, engaging in self-explanation and schema construction. Second, the interactive format enabled them to discover new mathematical relationships and solution strategies that were previously unavailable to them. Third, and notably, teachers simultaneously developed pedagogical insights—recognizing instructional approaches and anticipating student difficulties—while working through the mathematical content. These findings extend worked example theory by demonstrating that carefully designed interactive examples can support dual-domain learning in professional contexts. By managing cognitive load through scaffolded interactivity, worked examples enable teachers to construct both mathematical and pedagogical schemas concurrently, suggesting a cognitively efficient pathway for teacher professional development. Given the study’s exploratory, single-setting design, these mechanisms are offered as a theoretically grounded account to be tested in future comparative research rather than as definitive causal claims.

Journal of Intelligence doi: 10.3390/jintelligence14090202

Authors: Zeltia Martínez-López Mª Emma Mayo Jose Eulogio Real Carolina Tinajero

Motivational self-regulation strategies (MRSs) play a key role in how students initiate, sustain, and direct effort during learning. However, little is known about their use among adolescents with different educational profiles. This study aimed to explore possible differences in the use of MRSs among adolescent students with varying learning potential (high-ability, learning-disability and regular students) and to examine whether the relationship between MRSs and academic adjustment differed across these educational profiles. We recruited a sample of 2191 students enrolled in compulsory secondary education, comprising regular students (81.8%, N = 1792), students with learning disabilities (15.8%, N = 347), and high-ability students (2.4%, N = 52). The participants completed self-report measures of MRSs and academic adjustment. Path models were used to examine cross-sectional direct and indirect associations among the study variables, and they revealed different patterns across the three groups. High-ability status was directly and positively associated with academic adjustment, while learning-difficulty status showed a direct negative relationship with adjustment. The three groups of students showed indirect associations with adjustment via various types of MRSs.

Journal of Intelligence doi: 10.3390/jintelligence14090201

Authors: Juan Wang Zichen Liu Jiaying Huang

Generative artificial intelligence (AI) is reshaping how students evaluate their abilities, occupational prospects, and the continuing value of professional education. This study examined how ability-based comparison with AI is associated with perceived professional value among journalism and communication students and why the same comparison may be linked to both threatening and adaptive responses. An explanatory sequential mixed-methods design was used. Survey data from 507 students in China were analyzed using partial least squares structural equation modeling, with PROCESS analyses as robustness checks, followed by semi-structured interviews with 20 students. Ability-based comparison with AI was positively associated with job replacement anxiety, which was negatively associated with perceived professional value, and with AI learning motivation, which was positively associated with perceived professional value. A positive direct association with perceived professional value also remained. AI hindrance appraisal strengthened the comparison–anxiety association, whereas AI challenge appraisal did not significantly moderate the comparison–motivation association. Interviews indicated that these relationships were task- and criterion-dependent and shaped by occupational interpretation and concrete learning demands. Human–AI comparison may therefore relate to professional value through simultaneous threat- and adaptation-oriented pathways rather than a uniformly positive or negative process.

Journal of Intelligence doi: 10.3390/jintelligence14090199

Authors: Mohammad Hamad Al-khresheh Suheyla Demirkol Orak Wael Alharbi Mayez Almayez

The growing integration of artificial intelligence (AI) into second-language reading instruction has intensified debate regarding whether AI-mediated support facilitates comprehension or reduces the cognitive effort required for independent meaning construction. Although previous research has reported generally positive effects of AI-assisted reading, limited attention has been directed toward the timing of support during the reading process itself. The present study examined whether immediate versus delayed AI support produced different effects on EFL learners’ reading comprehension performance. A pretest–posttest experimental design was employed, involving 90 undergraduate EFL learners, randomly assigned to three instructional conditions: Immediate AI Support, Delayed AI Support, and No-Support control. The intervention was conducted over a ten-week period using AI-assisted reading tasks administered under controlled instructional conditions. Reading comprehension performance was measured using validated pre- and post-test instruments that assessed literal, inferential, and global comprehension. Psychometric validation procedures incorporating Classical Test Theory and Item Response Theory supported the instrument’s reliability and measurement adequacy. One-way ANOVA revealed statistically significant differences across the three instructional conditions, F(2, 87) = 18.679, p < .001, η2 = 0.301. Post hoc comparisons demonstrated that learners receiving delayed AI support achieved significantly higher post-test reading comprehension scores than learners in both the immediate support and no support conditions. Learners receiving immediate AI support similarly outperformed the no-support group. The findings indicate that the effects of AI-assisted reading varied according to when explanatory support became available during the reading process. Delayed AI support produced stronger reading comprehension performance than immediate support and unsupported reading, suggesting that support timing constituted an important condition shaping comprehension performance in AI-mediated reading environments.

Journal of Intelligence doi: 10.3390/jintelligence14090200

Authors: Chien-Hong Yu Ching-Chih Kuo

Gifted students possess broad developmental potential; however, these characteristics may also create challenges in career decision-making. Previous research suggests that gifted individuals’ career development is influenced by multiple interacting factors across personal, familial, educational, and sociocultural domains. This retrospective study investigated the career development of individuals who participated in the preschool gifted program (PSMIGP: Problem-Solving and Multiple Intelligences for Gifted Preschoolers) launched by National Taiwan Normal University in 2003. Semi-structured interviews were conducted with 18 participants selected based on domain of strength, gender, cohort, and willingness to participate. Interview data were analyzed using NVivo 15 through qualitative coding procedures, with multiple data sources, collaborative review, and peer discussion employed to enhance trustworthiness and contextual understanding. Findings revealed that gifted individuals’ career decisions were shaped by interacting personal, family, learning, and sociocultural factors. Although participants across domains demonstrated strong self-awareness and agency, academic participants generally followed more stable career pathways, whereas artistic participants encountered greater complexity and practical constraints. Artistic participants also exhibited greater sensitivity to perceived sociocultural risks. Based on these findings, this study underscores the importance of parental support, strengthened career guidance, and diverse exploratory opportunities within educational settings and the promotion of diverse social values to support gifted students’ career development. Overall, gifted career development should be understood as a dynamic process shaped through continuous interaction between individuals and their broader developmental contexts.

Journal of Intelligence doi: 10.3390/jintelligence14090198

Authors: Jennifer Wiley Taylor Strickland Miller

There has been an increase in studies using Aha! ratings to explore insight in creative problem solving. Aha! moments are often assumed to serve as a marker for insight—meaning that solutions are achieved via restructuring. However, similar to metacognitive paradigms that collect FOKs (Feelings of Knowing) or JOLs (Judgments of Learning) and JOUs (Judgments of Understanding) to track monitoring of memory, learning, and comprehension processes, Aha! ratings may be better considered as “Judgments of Aha!” or JOAs derived from an inferential process based on a variety of cues. Only some of these may directly reflect underlying problem-solving processes. Using a set of spatial object-move puzzles, this study explored JOAs and how they may vary (or fail to vary) with initial fixation, correct solution, and individual differences in growth mindset. Instead of JOAs being highest for solutions where an initial representation needed to be revised, they were highest for solutions where solvers made a correct first move. This suggests that Aha! experiences are more closely tied to achieving correct solutions than to restructuring processes. Individuals with a stronger growth mindset were more likely to associate these initially correct solutions with Aha! experiences.

Journal of Intelligence doi: 10.3390/jintelligence14090197

Authors: Shaoyang Guo Yinmei Xu Xiangdong Yang Yanlei Chen Xiaoyu Li Xuefei Tian

Collaborative problem-solving (CPS) is an essential 21st-century competency, yet its valid and ecologically sound assessment remains challenging due to inconsistent frameworks, poorly aligned task design, and over-reliance on behavior frequency counting in scoring. This study developed and provided empirical support for a theory-driven game-based CPS assessment embedded in a real-time strategy game, using a customized Collaborative Siege Task with asymmetric roles and ill-structured problems, to elicit more realistic collaborative interactions. Based on a validated three-facet CPS framework from Chinese cultural context, the task systematically triggers social and cognitive processes of collaboration across five progressive stages. By sampling from the eastern provinces of China, a pilot study with 36 participants confirmed task feasibility and indicator effectiveness, while a formal study with 298 valid participants examined psychometric properties using item response theory and confirmatory factor analysis. Findings supported the hierarchical higher-order factor structure, satisfactory item difficulty, discrimination, reliability, and criterion-related validity against peer assessment, reasoning ability, and personality. The assessment showed no sex, major, or role bias. This study provides a reliable gamified assessment tool for CPS measurement with demonstrated potential for scalability pending further development of automated scoring procedures, offering implications for large-scale evaluation and competency-oriented instruction.

Journal of Intelligence doi: 10.3390/jintelligence14090196

Authors: Filiz Mizrak Gozde Mert Ahmet Erkasap Ali Ozcan Mehmet Mert

Generative artificial intelligence (GenAI) is increasingly embedded in consulting and professional-service work, yet employees may not always disclose its contribution to work outputs. This study examines hidden GenAI use as a social-cognitive, organizational, and human–AI interaction phenomenon and investigates how individual and organizational factors are associated with responsible AI disclosure and concealment. Survey data were collected from 564 consulting and professional-service employees in Türkiye. The analytical framework combined PLS-SEM with covariance-based SEM robustness checks, cross-validated predictive modeling with out-of-fold SHAP interpretation, and exploratory Choquet decision support. The structural results showed that AI disclosure anxiety and competence penalty fear were positively associated with hidden GenAI use, whereas social intelligence, official AI tool adequacy, and responsible AI disclosure were negatively associated with concealment. AI metacognition, human–AI trust, AI policy clarity, social intelligence, and official AI tool adequacy were positively associated with responsible AI disclosure. In the predictive analysis, Ridge regression achieved the strongest average cross-validated performance, while XGBoost was used for nonlinear interpretation through out-of-fold SHAP values; AI disclosure anxiety and competence penalty fear emerged as the strongest predictors. The exploratory Choquet analysis consistently prioritized a non-punitive AI-use disclosure system under alternative weighting and uncertainty scenarios. The study contributes by shifting attention from whether professionals adopt GenAI to whether, why, and under what organizational and social conditions they disclose or conceal its substantive contribution to professional work. The cross-sectional design supports association-based rather than causal interpretation.

Journal of Intelligence doi: 10.3390/jintelligence14090191

Authors: Mabel Urrutia Alonso Valderrama Susana Araya Luciano Silva Francisco Ortiz Cristian Medina Pedro Salcedo Pamela Guevara Esteban J. Pino Hipólito Marrero

Developing emotional competencies has been identified as a key component of personal well-being, school coexistence, and significant learning during childhood and adolescence. The aim of this study was to map and categorise the quantitative scientific evidence available between 2019 and 2024 regarding the use of artificial intelligence technologies and algorithms applied to emotional education in school-age populations. This scoping review followed the PRISMA Extension for Scoping Reviews (PRISMA-ScR) and searched seven databases (Web of Science, Scopus, PubMed, PsycNet, ERIC, IEEE Xplore and the ACM Digital Library). The reviewed studies were published between the years 2019 and 2024. Eighteen studies met the inclusion criteria required for the final analysis. The results show that the most trained competence was machine-based emotional recognition using convolutional neural networks (CNNs). The goal of using them was to provide teacher support, specifically during initial or diagnostic assessments of students’ competencies. However, there remains a gap in its integration into the teaching and training of complex human skills, such as empathy. More longitudinal research is required, as well as more robust experimental designs that go beyond basic emotion models.

Journal of Intelligence doi: 10.3390/jintelligence14090193

Authors: Carlos Saiz Miguel H. Guamanga Silvia F. Rivas Cristiano M. A. Gomes

This study reanalyzes data from a previous investigation on critical thinking, metacognition, empathy, and psychological well-being by incorporating academic performance into a graphical-causal model. The aim was to compare two explanatory structures: a performance-mediated model, in which metacognition relates to psychological well-being through academic performance, and a direct metacognitive model, in which metacognition maintains a direct relation with well-being. The sample included 155 undergraduate psychology students. Critical thinking, metacognition, empathy, academic performance, and psychological well-being were modeled using directed acyclic graphs, conditional independence tests, Gaussian Bayesian networks, and bootstrap-based arc stability analysis. The initial performance-mediated DAG was not supported because the implied independence between metacognition and well-being conditional on performance was violated. The refined model, which included a direct metacognition–well-being path, was more compatible with the data. Metacognition showed the strongest estimated path to psychological well-being, whereas academic performance did not operate as the main mediating mechanism. Bootstrap results supported the stability of the metacognition–well-being adjacency. Overall, the findings suggest that critical thinking relates to psychological well-being mainly through metacognitive processes, under the assumptions of the specified DAG.

Journal of Intelligence doi: 10.3390/jintelligence14090195

Authors: Yunhuan Pu Wei Zhao

This study involved 30 children with Autism Spectrum Disorder (ASD) aged 7–10 years and 31 typically developing (TD) children who were matched for age and IQ. The participants’ performance on visual and auditory statistical learning tasks within a segmentation paradigm was assessed. Generalised linear mixed models (GLMMs) and linear mixed models (LMMs) were used to analyse response accuracy and reaction times, respectively. The aim was to explore the characteristics of statistical learning in children with ASD. The results showed that the children with ASD failed to perform above chance level on the auditory statistical learning task (p = 0.156). Only 40% of ASD participants performed better than expected by chance. Their performance was significantly lower than that of the TD children. Conversely, children with ASD performed significantly above chance level in the visual statistical learning task (p < 0.001). GLMM analysis revealed a significant interaction between group and task modality (p = 0.023). Specifically, children with ASD demonstrated significantly lower accuracy in auditory statistical learning than TD children (p < 0.001). In contrast, no significant difference was observed between groups in the visual task (p = 0.468). LMM analysis showed a significant interaction between group and task modality regarding reaction times (p = 0.020). Children with ASD exhibited significantly longer reaction times than TD children in both the visual and auditory tasks (ps < 0.001). These findings suggest that children with ASD have difficulty with auditory statistical learning, but perform relatively well in visual tasks within the segmentation paradigm.

Journal of Intelligence doi: 10.3390/jintelligence14090194

Authors: Jingying Chen Tengfei Gao Rui Li Zhiyi Yang Yuhan Li

Learning potential assessment is important for understanding individual differences in learning processes and cognitive development. Traditional assessments mainly rely on questionnaires or behavioral data collected during tasks, emphasizing external performance while lacking sensitivity to learners’ internal cognitive activities. Therefore, they provide limited evidence about the neurocognitive mechanisms underlying individual learning potential. This study investigated whether task-evoked EEG responses acquired during a dynamic learning paradigm could characterize individual differences in task-specific learning performance under graduated instructional support. Participants were 50 undergraduate students who completed a dynamic cognitive task designed to evaluate learning performance under graduated instructional support. An electroencephalogram (EEG)-based Multi-scale Learning Potential Assessment (MS-LPA) framework was developed, including a Dynamic Learning Paradigm with progressively complex cognitive tasks and responsive cueing; multiscale Neural Complexity Representations quantifying signal complexity and inter-regional synchronization; and a ResNet-18 model with SHAP-based interpretability for classification and biomarker identification. Experiments showed that MS-LPA captured task-evoked neural responses and outperformed traditional outcome-driven evaluation schemes. Entropy-based features achieved an accuracy of 0.77, and functional connectivity matrices reached 0.83. These findings demonstrate the feasibility of using EEG-derived neural representations to differentiate individual differences in task-specific dynamic learning performance and provide insights into its underlying neural mechanisms.

Journal of Intelligence doi: 10.3390/jintelligence14090192

Authors: Taylor Stone Lesenia R. Fish Nadia Chernyak Gabriella Guadalupe Sara Cordes

Worked examples—step-by-step examples of how to solve a problem correctly or incorrectly—are known to promote algebraic thinking in middle and high school students. Whether they are useful in facilitating math learning in much younger students is unclear, however. The present study investigated how highlighting correct versus incorrect division strategies in a sharing context supports preschool children’s learning and transfer across division tasks. Two hundred and one children (ages 4.01–6.96 years, M = 5.57) viewed videos of agents dividing toys using either a correct (one-to-one turn-taking) or incorrect (i.e., error prone) strategy. The children were asked to explain what the agent had done correctly, what the agent had done incorrectly, or to simply identify the color of the agents’ shirts (Baseline condition). Afterwards, they completed two division tasks which required them to divide resources between two recipients (similar to training) and then between three recipients (novel division task). The results revealed that highlighting the correct or incorrect procedures led to improved performance on the division task relative to baseline, particularly for the youngest children completing the most challenging trials. Moreover, analyses of children’s explanations revealed that those children who spontaneously referenced the division process outperformed those who referenced the outcome or another aspect. These findings suggest that drawing attention to procedural strategies can facilitate division learning in early childhood, highlighting the importance of process-focused instruction as a mechanism for supporting early mathematical development.

Journal of Intelligence doi: 10.3390/jintelligence14080190

Authors: Liming Jiang Fang Luo Xuetao Tian

In an era of information explosion and overload, individuals face a constant influx of mixed-quality, contradictory data. Navigating this complex landscape requires critical thinking, which enables individuals to dynamically adjust their reasoning and refine judgments. Consequently, accurately measuring this dynamic ability is essential. However, existing assessments predominantly focus on single-round argumentation and fail to capture the iterative process, potentially yielding unrepresentative evaluations of real-world performance. To address this gap, this study develops the Iterative Argumentation Task (IAT), a novel assessment tool that presents participants with a sequence of interrelated passages simulating real-world information flow. By capturing the process of evaluation, position changes, and argument revisions, the IAT evaluates three ability dimensions: single-round argument analysis, inter-argument relationship analysis, and argument integration. We recruited 355 undergraduates to validate the IAT. Confirmatory factor analysis provided evidence for the proposed measurement structure. Furthermore, correlations with established critical thinking and reasoning tests demonstrated convergent and discriminant validity, respectively. Moreover, performance differences between the undergraduates and 33 debaters established known-groups validity. Finally, latent profile analysis identified four distinct profiles across the IAT dimensions. Together, these findings establish the IAT as a theoretically sound and practically effective tool for assessing critical thinking.

Journal of Intelligence doi: 10.3390/jintelligence14080189

Authors: Lihang Guan Qingbo Yang

Artificial intelligence (AI) creates personalized, interactive, and low-risk learning opportunities for informal digital learning of English (IDLE). However, effective AI tools do not necessarily lead learners to engage autonomously or productively. Guided by the person–behavior–environment framework of social cognitive theory, this systematic review synthesized 58 empirical studies published in SSCI-indexed journals between 2020 and 2025 to examine how AI affordances are translated into learning behaviors and intelligence-related processes in AI-mediated IDLE. Speech recognition, natural language processing, and personalized feedback were the most frequently investigated affordances, providing opportunities to observe, imitate, and adapt authentic language use. Engagement was mediated primarily by emotional, motivational, and self-belief mechanisms. These mechanisms could facilitate or hinder engagement depending on technological accuracy, perceived human responsiveness, learner goals, and proficiency-task alignment. From an intelligence perspective, the findings illustrate how learners evaluate AI-generated information, regulate affect and motivation, calibrate self-perceptions, and adapt learning behaviors. Thus, AI-mediated IDLE may serve as a site for activating adaptive intelligence and learner self-regulation through context-sensitive evaluation, regulation, and behavioral adaptation.

Journal of Intelligence doi: 10.3390/jintelligence14080188

Authors: Şule Betül Tosuntaş

Academic procrastination is a widespread self-regulatory failure among university students and a costly one for both academic achievement and psychological well-being. In this study, procrastination was examined within the framework of non-cognitive individual differences. The model tested here posits that perfectionism and locus of control affect procrastination through negative automatic thoughts and general self-efficacy. These variable-centered analyses were then complemented with a person-centered approach in order to capture distinct risk profiles. Cross-sectional data were collected from 1003 Turkish university students (71.5% women; Mage = 21.93 years) and analyzed with structural equation modeling, latent profile analysis (LPA), and random forest regression. The structural model accounted for 62% of variance in procrastination (CFI = 0.958, RMSEA = 0.074). Based on these results, general self-efficacy was identified as a common pathway through which distal non-cognitive influences converge on procrastination. The direct effect of automatic thoughts was non-significant, whereas the indirect effect through self-efficacy was strong (indirect effect = 0.51, 95% CI [0.44, 0.58]). LPA identified three profiles: adaptive, average, and vulnerable; the vulnerable profile showed markedly higher procrastination than the adaptive profile (d = 1.21). Random forest analysis indicated no meaningful nonlinearity, with self-efficacy as the dominant predictor. These findings suggest that interventions may benefit from prioritizing general self-efficacy beliefs.

Journal of Intelligence doi: 10.3390/jintelligence14080187

Authors: Yang Liu Heying Liu Pengyang Kang

Background: Developmental topographical disorientation (DTD)-like navigational difficulties can impair everyday orientation and spatial information processing. Objective: We examined whether 8-week virtual reality (VR) orienteering training improves performance on a laboratory spatial memory task in college students with operationally defined DTD-like navigational difficulties and examined task-related cortical changes. Methods: In a randomized controlled trial, 96 students were assigned to VR orienteering, VR exercise, or control groups (n = 32/group). The exercise groups trained twice weekly for 45 min for 8 weeks at moderate intensity (64–76% HRmax). Before and after the intervention, all participants completed a delayed visuospatial recognition task designed to assess spatial-memory-related processing, while functional near-infrared spectroscopy measured hemodynamic changes in prefrontal and sensorimotor cortices. The post-test session was conducted 48 h after the final scheduled session, with no immediate post-exercise assessment. Results: VR orienteering produced greater improvement in spatial-memory-task accuracy and reaction time than the other groups. After BH-FDR correction across eight ROIs, training-specific reductions in task-related activation were observed in the right primary motor cortex and left frontopolar area. Within the VR orienteering group, pre-to-post change in left orbitofrontal activation was negatively associated with change in spatial-memory-task accuracy, Pearson r(30) = −0.819, 95% CI [−0.908, −0.658], p < 0.001, R2 = 0.670; the association remained significant after eight-ROI BH-FDR correction (adjusted p < 0.001). Conclusions: VR orienteering improved performance on the laboratory spatial memory task and was accompanied by reduced task-related cortical recruitment. Greater reductions in L-OFC activation were linked to greater improvements in accuracy, identifying L-OFC activation change as a neural correlate of training-related spatial-memory improvement.

Journal of Intelligence doi: 10.3390/jintelligence14080186

Authors: Yuan Liu Chenyang Zhao Zheping Gu Hong Jun Yanyan Hu He Liu Xiaojian Yin

Executive function (EF) is a core component of human cognitive performance. This study investigated whether increasing the duration of moderate-to-vigorous physical activity (MVPA) could enhance this critical cognitive ability in Chinese adolescents. Adolescents (13.4 ± 0.5 years) from two secondary schools in Jiangsu Province were recruited, screened, and individually randomized into an intervention group (n = 101) and a control group (n = 105); after three withdrawals, the final analytical sample comprised 98 and 105 participants, respectively. EF was compared using independent and paired sample t-tests, analysis of covariance (ANCOVA), and effect size tests. Post-intervention, adolescents in the intervention group showed a significantly longer daily MVPA duration than the control group (mean difference 25.16 min, p < 0.05). As an exploratory analysis, sex-stratified ANCOVAs were conducted. After adjusting for pre-intervention inhibition control reaction times (RTs), ANCOVA revealed that boys in the intervention group had a lower mean inhibition control RT (1.03 ms lower than controls, p < 0.05). With pre-intervention 2-back RTs as a covariate, ANCOVA indicated significant reductions in mean 2-back RTs for intervention group boys (191.09 ms lower, p < 0.05) and girls (175.62 ms lower, p < 0.05). After adjusting for pre-intervention cognitive flexibility RTs, boys in the intervention group also exhibited a significantly lower mean cognitive flexibility RT (102.7 ms lower, p < 0.05). In conclusion, a combined aerobic and resistance (HIIT) exercise intervention effectively increased adolescents’ physical activity time, and concurrent improvements were observed in key components of human cognitive performance—specifically inhibition control, working memory, and cognitive flexibility in boys and working memory in girls.

Journal of Intelligence doi: 10.3390/jintelligence14080185

Authors: Wenzhang Li Weida Zhang Mengling Li Yi Zhang

Generative artificial intelligence (GenAI) provides a new context for examining how external intelligent systems are associated with human cognitive agency and perceived creative intellectual functioning. Drawing on cognitive offloading theory and AI literacy research, this cross-sectional study tested a moderated mediation model linking GenAI use intensity, GenAI cognitive dependence, AI literacy, and perceived academic creativity. Survey data were collected from 936 undergraduates at six universities in China. Confirmatory factor analysis supported the empirical distinctiveness of the four focal constructs, and latent structural equation modeling was used to examine the hypothesized associations. GenAI use intensity showed a positive direct association with perceived academic creativity and a positive association with GenAI cognitive dependence; cognitive dependence was negatively associated with perceived academic creativity. The standardized direct and indirect associations operated in opposite directions, leaving a small model-implied net association (β ≈ 0.040). AI literacy was negatively associated with cognitive dependence and positively associated with perceived academic creativity. The use-dependence association was weaker at higher levels of AI literacy, and the negative indirect association through dependence was correspondingly smaller. These cross-sectional results are interpreted as associations and do not establish temporal or causal ordering.

Journal of Intelligence doi: 10.3390/jintelligence14080184

Authors: László Balázs

Research exploring the relationship between emotional intelligence and workplace outcomes is predominantly variable-centered and treats the dimensions of emotional–social competencies as separate from one another, thereby losing sight of the intra-individual interplay of these competencies and the potential heterogeneity of their effects. This study examines teachers’ emotional–social intelligence using a person-centered approach. Using a sample of 587 Hungarian teachers from 26 public education institutions, we conducted latent profile analysis based on Bar-On’s five composites and then used unbiased three-step procedures (BCH; categorical BCH) to examine the antecedents of profile membership, as well as perceptions of school organizational climate and dominant conflict management style. Four distinct profiles emerged, primarily differing at the global level—a uniformly low profile (12%), a medium-low profile (25%), a dominant, balanced medium-high profile (53%), and a uniformly high profile (9%). Participants with higher profiles perceived the school’s organizational climate as more supportive (η2 = 0.07) and exhibited more adaptive conflict management patterns (less avoidance, more problem-solving; χ2(12) = 37.24, p < .001). A compositional (log-ratio) reanalysis of the complete conflict-mode profiles, reported as the primary conflict analysis, showed that the relative weight of problem-solving rises across profiles and plateaus between the modal and the highest profile, that avoidance declines with profile elevation, and that an apparent surplus of accommodating in the highest profile was an artifact of dominant-style coding rather than an indication of a broader repertoire. At the same time, profile membership was essentially independent of job position, career stage, and gender, and the relationships with outcomes were of modest to moderate strength. The results extend person-centered, profile-based intelligence research to the emotional–social domain and identify promising directions for research on teacher professional development.

Journal of Intelligence doi: 10.3390/jintelligence14080183

Authors: Xiaojun Guo Xiaohua He Xiaoyun Bai Juan Yan

In computerized assessments, response accuracy and response time together reflect how examinees engage with test items. Existing joint models typically assume a linear relationship between these two outcomes, yet empirical observations often reveal more complex patterns. This study proposes a Quadratic Bifactor Hierarchical Model (QBi-HM) that captures nonlinear speed–accuracy dependencies through a shared general factor with quadratic terms, while preserving separate ability and speed components. Simulation results across sample sizes (N = 1000, 1500, 3000) and test lengths (m = 30, 60) demonstrated that QBi-HM recovered parameters accurately under both linear and nonlinear conditions, whereas the conventional linear bifactor model produced biased estimates when nonlinearity was present. An empirical application to PISA 2012 computer-based mathematics data (N = 1527 examinees from four economies, m = 10 items) showed that QBi-HM achieved superior model fit (AIC = 41,825.190, BIC = 42,251.675, SABIC = 41,997.535) compared with Bi-HM (AIC = 41,969.750, BIC = 42,289.613, SABIC = 42,099.099) and revealed an inverted U-shaped speed-accuracy pattern. These findings may suggest that modeling nonlinear dependencies can improve the precision of ability estimation in large-scale assessments and inform item design by identifying task features that elicit distinct response processes. The QBi-HM framework offers a flexible tool for researchers and practitioners seeking to leverage response time data to better understand test-taking behavior.

Journal of Intelligence doi: 10.3390/jintelligence14080182

Authors: Jing Su Zhuo Wang Zhen Qiang Zhonghou Wang

Cognitive offloading—delegating cognitive work to external tools—is basic to human cognition, but generative AI (GenAI) amplifies it radically: entire cognitive products can now be produced on request. This sharpens a question about agency over one’s own thinking: when cognition is habitually offloaded, does its product still feel like one’s own? We surveyed 239 pre-service teachers, operationalizing the felt loss of cognitive ownership as concern over the erosion of teaching-design subjectivity (TSC)—the metacognitive appraisal that AI-assisted work is not genuinely one’s own and that independent capacity is declining. Guided by the Interaction of Person-Affect-Cognition-Execution (I-PACE) model, we tested whether AI anxiety/fear of missing out (affective) and impulsivity (self-regulatory) relate to this concern through behavioral GenAI dependency—habitual offloading. In a structural equation model with bias-corrected bootstrapping, dependency strongly predicted TSC and partially mediated the effect of AI anxiety; impulsivity raised dependency but showed a suppression pattern (a positive indirect effect offset by a null total effect), and the model explained 35% of the variance in the concern. Habitual offloading—that is, GenAI dependency—rather than generic AI use, accompanies the metacognitive loss of cognitive ownership.

Journal of Intelligence doi: 10.3390/jintelligence14080181

Authors: Funda Uğurlu Filiz Evran Acar

Digital game-based learning is increasingly used to support children’s engagement, interaction, and social–emotional development. However, its contribution to emotional intelligence, particularly in inclusive classroom contexts, remains theoretically underdeveloped and methodologically fragmented. This systematic review and critical narrative synthesis examines how teacher- or facilitator-mediated digital games address children’s emotional intelligence and related social–emotional outcomes in inclusive educational contexts. Following PRISMA 2020 reporting guidance, the review examined peer-reviewed articles indexed in Web of Science and Scopus between 2015 and 2026. After screening 347 unique records and assessing 31 full-text reports, 14 studies were included in the final synthesis. The review focused on four interrelated areas: the emotional and social–emotional outcomes addressed in digital game-based studies; the roles of teachers or facilitators in structuring, mediating, and debriefing game-based activities; the inclusive classroom conditions that support equitable participation and emotional safety; and the risks, limitations, or implementation challenges associated with classroom use. The findings suggest that digital games may become educationally meaningful when they are linked to explicit emotional learning goals, embedded in structured classroom or school-based activities, supported by teacher or facilitator mediation, and followed by reflection or debriefing. Based on the synthesis, the review proposes a teacher-mediated inclusive digital game-based emotional learning framework consisting of seven dimensions: emotional learning target, game selection and pedagogical fit, teacher mediation, inclusive participation, classroom interaction, reflective debriefing, and risk monitoring.

Journal of Intelligence doi: 10.3390/jintelligence14080180

Authors: Jingying Chen Yijia Zheng Yuetong Li Xiaodi Liu Rui Li

Foreign language reading comprehension is a critical indicator of university students’ academic cognitive ability and global competence, yet effectively assessing their latent learning potential remains a major challenge. Therefore, from a neurocognitive perspective, this study combined functional near-infrared spectroscopy (fNIRS) with dynamic assessment (DA) to investigate the activation characteristics of the left frontal and temporal regions in non-English majors during reading tasks across two difficulty levels (CET-4/CET-6) and two assessment conditions (static/dynamic testing). The results showed that reading tasks activated the dorsolateral prefrontal cortex and middle temporal gyrus. Compared with CET-4, the more difficult CET-6 task showed a broader descriptive cortical activation pattern. Notably, mediated prompts were associated with different behavioral and activation-frequency patterns across task difficulty levels: in the CET-4 task, dynamic assessment showed higher normalized activation-frequency values than static assessment in areas corresponding to the middle and inferior frontal gyrus, accompanied by rising prompt scores; however, the CET-6 task exhibited the opposite pattern. These findings suggest that mediated prompts may be more effective in stimulating learning potential when task difficulty aligns with the learner’s zone of proximal development. Overall, by considering behavioral dynamic-assessment indicators in foreign-language reading together with fNIRS-based cortical activation patterns, this exploratory study offers preliminary insights into learning potential, personalized instructional interventions, and academic cognitive assessment.

Journal of Intelligence doi: 10.3390/jintelligence14080179

Authors: Yuqiu Fu Yuxi Wang Hongzhao Xie Shiyun Zhao Mingyuan Liu Yujie Lu Xinyi He Zhenku Cheng Yujia Peng Zhenliang Zhang

Current artificial intelligence evaluation often relies on static benchmarks and task-specific performance metrics, which are useful for system comparison but limited in assessing developmental progression, cognitive architecture, adaptive transfer, and real-world applicability. Building on developmental approaches to AGI testing, including our earlier conceptual proposal, this article develops an operational formulation of the General–Specialized–Applicable (GSA) framework for artificial general intelligence (AGI) evaluation. The framework organizes evidence into three non-interchangeable stages: the General stage, which assesses foundational capacities for open-ended generalization, value-oriented regulation, and autonomy; the Specialized stage, which evaluates stable domain-specific competence; and the Applicable stage, which examines whether such competence can be deployed safely and robustly in realistic environments. The article further introduces a dynamic task-generation pipeline and a provisional operational rubric for interpreting stage-specific evidence. An illustrative case study using household tasks in a simulated embodied environment illustrates how GSA can provide diagnostic information beyond aggregate benchmark scores by identifying capability gaps in current multimodal large language model (MLLM) agents. Rather than offering a final universal standard, the GSA framework provides an evaluation-oriented structure for connecting benchmark performance with architectural readiness, specialization, and deployment-level applicability.

Journal of Intelligence doi: 10.3390/jintelligence14080178

Authors: Jinjushang Chen Qian Zhang Yue Yin Jeannine E. Turner Liang Luo Chunliang Yang

Metacognitive monitoring, students’ evaluations of their own performance, is central to self-regulated learning but often misaligns with actual performance. While situational factors have been studied extensively, less is known about how learning-related beliefs and motivational dispositions—hereafter termed learning-related attributes—relate to individual differences in metacognitive monitoring. This study examined how these learning attributes, alongside cognitive and demographic variables, were associated with students’ general confidence and resolution. Data were drawn from 3946 Chinese sixth-grade students who completed academic assessments in the verbal and mathematical domains, provided post-test performance estimates, and completed questionnaires measuring their learning attributes and demographic characteristics. We conducted multilevel analyses with a random-intercept–random-slope specification and identified several learning attributes as significant predictors of metacognitive monitoring, primarily in terms of general confidence. Adding learning-related attributes after the cognitive and demographic blocks increased marginal R2, primarily for general confidence; this order-dependent increment was not interpreted as evidence that the block is inherently more important. The findings extend existing accounts of metacognitive monitoring by suggesting that students’ general confidence is associated with non-cognitive learning attributes—and may thus be partly separable from cognitive ability—whereas their resolution is comparatively less strongly associated with these attributes. Theoretical as well as practical implications are discussed.

Journal of Intelligence doi: 10.3390/jintelligence14080177

Authors: Jingzhan Ren Xiaotong Du Wei Wang Sijing Fan Han Wang Gongxing Fang Yuan Li Ruilong Wang Hegui Bao Chenyu Zhang Ruohan Zhu Xinming Ye Wen Fang

Background: As population aging accelerates, pharmacological treatments provide limited benefits for cognitive decline. Exercise and technology-assisted exercise have thus emerged as important non-pharmacological approaches for supporting cognitive health in older adults. However, comparative evidence on the relative effectiveness of different intervention modalities across cognitive outcomes remains limited. Objective: This study used a network meta-analysis to systematically compare and rank the effects of traditional and non-traditional exercise interventions on cognitive function in older adults. Methods: This systematic review and network meta-analysis was registered on PROSPERO (CRD420261278685). PubMed, Embase, the Cochrane Library, Web of Science, and Scopus were searched from inception to May 2026. Randomized controlled trials enrolling participants aged 60 years or older were included. Interventions comprised aerobic, resistance, mind–body, finger, and multicomponent exercise, as well as virtual reality-based interventions, artificial intelligence-assisted exercise, and wearable exoskeleton training. Primary outcomes included global cognition, executive function, memory, attention, and activities of daily living. A random effects network meta-analysis was applied to estimate standardized mean differences with 95 percent confidence intervals, and intervention rankings were derived using SUCRA values. Results: A total of 70 randomized controlled trials involving 5573 older adults were included. All active interventions demonstrated overall cognitive benefits compared with usual care or non-active control conditions. Mind–body exercise showed the highest probability of improving global cognitive performance. Virtual reality-based interventions were particularly effective for executive function, memory, and activities of daily living, while artificial intelligence-assisted exercise showed favorable rankings for attention outcomes. Mind–body exercise showed favorable effects on global cognition, whereas several technology-assisted interventions ranked highly for selected cognitive and functional outcomes. Conclusions: Both traditional and non non-traditional exercise interventions improve cognitive function in older adults, although their effects differ across cognitive domains. Technology-assisted approaches appear more effective for enhancing specific cognitive functions, while traditional interventions such as mind–body exercise are more advantageous for preserving overall cognitive performance. These findings highlight the potential value of stage specific and combined intervention strategies for promoting cognitive health in aging populations.

Journal of Intelligence doi: 10.3390/jintelligence14080176

Authors: Natalie A. Ceballos Ruby N. Sweet Carmen E. Westerberg

Creativity has a dark side—malevolent creativity (MC), which is the application of original ideas to purposely harm others; i.e., creative problem-solving with malicious intent. Here, the influence of past MC behavior on MC ideation was examined while controlling for other individual-level factors relevant to the AMORAL model of MC. College undergraduates (N = 276 initially recruited; N = 248 final sample analyzed; 82% female) were surveyed on demographics, self-perception of problem-solving potential (SP-PSP), dark triad traits, and past MC behavior, and completed an adapted Malevolent Creativity Task (MCT), generating revenge ideas to an unfair scenario. Sequential regression models were used to predict MCT outcomes (fluency index, malevolence index, and composite score), incrementally controlling for age, sex, SP-PSP, dark triad traits, past MC behavior, and provocation level. In the final model, SP-PSP significantly predicted the MCT fluency index (R2 = 0.065). Past MC behavior significantly increased the R2 of the MCT malevolence index; the final model was significant (R2 = 0.084). Past MC behavior and narcissism significantly predicted the MCT composite score (R2 = 0.091). These findings align with and extend previous research linking past MC behavior to MC ideation by highlighting unique and incremental impacts of SP-PSP, narcissism, past MC behavior, and provocation on MC ideation.

Journal of Intelligence doi: 10.3390/jintelligence14080175

Authors: Tindara Caprì

The present study addressed a timely and important issue in cognitive and media psychology about the relationship between media multitasking and cognition. This study examined attention control, working memory, convergent associative thinking, and fluid reasoning, examining multiple functions within the same analytical framework and moving beyond the narrow focus on single cognitive domains that characterizes much prior research. The main aim was to deepen the understanding of the associations of media multitasking with these core domains of human cognition. A total of 214 healthy young adults (18–30 years) completed the Short Media Multitasking Questionnaire, Raven’s Progressive Matrices, the Digit Span Test, the Trail-Making Test, and the Remote Associates Task. Multivariate general linear models and subsequent regression analyses were performed while controlling for age and gender. The results did not reveal a significant overall association between media multitasking and the combined cognitive outcomes. At the univariate level, higher media multitasking scores were associated with lower performance on the Remote Associates Task, although the observed effect size was small. Given the cross-sectional design and reliance on self-reported media multitasking, longitudinal studies employing objective measures of media multitasking use are needed to clarify the robustness and directionality of these associations.

Journal of Intelligence doi: 10.3390/jintelligence14080174

Authors: Athanasios Kolovelonis Ioannis Syrmpas

This exploratory study examined students’ acute responses in soccer performance, executive functions, and calibration accuracy to a single soccer session delivered using three different teaching styles (i.e., command, practice, and self-check) in physical education. One hundred and forty-seven students of four elementary schools attending four fifth-grade (80 students) and four sixth-grade (67 students) classes participated in the study. Two classes (one fifth- and one sixth-grade) from each elementary school were randomly assigned to one of the four groups of the study (i.e., self-check style, command style, practice style, and control group). Students in all experimental groups were taught the soccer-passing skill in a single 45 min session using identical activities and drills but different teaching styles. Control-group students attended a classroom-based session on the ancient Olympic Games. Students were pre- and post-tested in soccer-pass performance, estimation of soccer-pass performance, and design fluency test. Students in the self-check group demonstrated more favorable changes in inhibition and cognitive flexibility than students in the command- and practice-style groups, whereas students in the practice group also showed improvements in cognitive flexibility over time. No differences between groups were found in calibration accuracy. For soccer performance, although students in the self-check group improved from pre- to post-test, nonsignificant differences between groups were observed. The findings reflected students’ immediate responses under different instructional conditions during a single soccer session, rather than outcomes of the longitudinal learning process. These results are discussed in relation to the cognitive and metacognitive aspects of learning and performing sports skills in physical education, with particular emphasis on how students responded to different teaching styles.

Journal of Intelligence doi: 10.3390/jintelligence14080173

Authors: Ioseba Iraurgi Miryam Martínez-Izaguirre Alexander Álvarez-González

Guaranteeing high-quality, equitable, and inclusive basic education requires upholding the right of all students to reach their full potential and ensuring their well-being. Students with High Intellectual Ability (HIA) represent a vulnerable group that is often under-identified in schools, a fact that hinders the personalization of their education and the ability to meet their specific needs. In 2019, as part of its plan to address the educational needs of this group, the Department of Education of the Basque Government prioritized the design and validation of a screening protocol to support the identification of students with HIA. This article presents a study on the validity and reliability of the High Intellectual Ability Screening Protocol (SACI, by its Spanish acronym), applicable in Primary Education. A two-phase controlled screening design was employed, involving the random assignment of participants and the use of specific statistical techniques to evaluate the psychometric soundness of the measure. The results provide highly promising evidence, supporting the conclusion that the SACI protocol is a valid and reliable screening instrument for use with Primary Education students in the Basque Autonomous Community.

Journal of Intelligence doi: 10.3390/jintelligence14080172

Authors: Meng Sun Jiutong Luo

With the widespread use of digital technologies, it has become increasingly important to understand how children adapt to digital environments. However, limited research has examined digital maturity among children beyond digital access and familiarity with technologies, particularly in non-Western contexts. Building on recent work on digital maturity, the present study aimed to validate the Digital Maturity Inventory (DIMI) among Chinese primary school students and to identify their latent profiles. A sample of 434 students (221 girls; Meanage = 10.68, SD = 0.86) completed the DIMI for validation analyses, and an additional sample was simultaneously recruited, resulting in a total sample of 1026 students (500 girls; Meanage = 10.38, SD = 1.09), for latent profile analysis. The findings demonstrated that the DIMI exhibited satisfactory psychometric properties in the Chinese context. Moreover, digital maturity, in addition to digital native, emerged as a unique and significant predictor of artificial intelligence (AI) literacy. Three distinct subgroups of students with heterogeneous patterns of digital maturity were also revealed: an unbalanced low group, an inconsistent medium group, and a relatively high group. These findings not only support the cross-cultural applicability of DIMI, but also demonstrate its relevance to children’s AI-related competencies. The implications of this study were also discussed.

Journal of Intelligence doi: 10.3390/jintelligence14080162

Authors: David Tzuriel Ravit Hozez Sagi

The present study examined the effects of a brief motoric mediation intervention, delivered within a dynamic assessment (DA) framework, on mental rotation (MR) performance among third-grade children. The study also investigated the roles of sex and spatial working memory (SWM). Participants completed two MR tasks, a Windows Mental Rotation (WMR) test and the Seria-Think–MR (STI-MR) test, assessing generalization. Each task was administered before and after a single mediation session. Four major findings emerged. (a) Children who received individualized motoric mediation demonstrated significantly greater MR gains than control children on both tasks. (b) The intervention produced a generalization effect, with improvements observed even on the structurally distinct STI-MR task. (c) No significant sex differences or sex-related interactions were found. (d) The positive pre-intervention correlation between SWM and WMR weakened after mediation, suggesting reduced reliance on WM resources following the acquisition of more efficient MR strategies. Across both tasks, a consistent Degree of Rotation × Time interaction indicated disproportionately larger gains at higher rotation angles, and children showed evidence of strategic reframing, particularly in interpreting 270° rotations as equivalent to 90° in the opposite direction. These findings support the theoretical framework of Structural Cognitive Modifiability and highlight the potential of motoric mediation to foster flexible spatial strategies. The results also underscore the educational value of integrating spatially rich, hands-on activities into early curricula and demonstrate the diagnostic utility of DA for identifying children’s learning potential. Implications, limitations, and directions for future research are discussed.

Journal of Intelligence doi: 10.3390/jintelligence14080171

Authors: Shuting Liu Yong Mei

Surprisal, the negative log probability of a word given its context, is the dominant computational metric for quantifying reading difficulty and a common item difficulty estimator in reading research. Yet how the language model family, surprisal granularity, and corpus type jointly shape the surprisal–reading time link remains unclear. We conducted a secondary analysis of two public English reading corpora: the Natural Stories Corpus (self-paced reading, 181 readers) and the Provo Corpus (eye tracking with cloze norms). Surprisal was computed from GPT 2 (Small, XL), Llama 2 (7B, 13B), and Llama 3 (8B), together with a 5-gram baseline and human cloze norms. Linear mixed-effects models tested the baseline contribution of neural surprisal, inverse scaling within model families, word-level versus sub-word aggregation, and the linking function shape. Neural surprisal contributed reliable variance above a strong baseline in both corpora. A clear inverse scaling pattern emerged: GPT 2 Small produced the largest fit improvements, exceeding GPT 2 XL, Llama 2 13B, and Llama 3 8B. Word-level aggregation outperformed sub-word aggregation, especially for measures of early lexical access. Non-parametric analyses supported an approximately linear linking function, and cloze norms carried information not fully captured by neural surprisal. These findings show that larger models are not automatically better cognitive models of reading and that surprisal granularity is not a neutral analytic choice.

Journal of Intelligence doi: 10.3390/jintelligence14080170

Authors: Ahmed Sadek Abdelmagid Naif Mohammed Yahya Jabli Adel Ibrahim Qahmash

Higher education is undergoing a rapid transformation driven by advancements in artificial intelligence (AI), creating new opportunities to support the development of higher-order cognitive skills through carefully designed learning environments. This study aimed to examine the effectiveness of an educational program that integrates AI-powered intelligent assistants into a structured e-learning environment to support the development of digital entrepreneurship and future-oriented problem-solving skills among graduate students. A quasi-experimental design was used with 82 graduate students, randomly assigned to two groups: an experimental group (n = 41) and a control group (n = 41). Both groups studied the same educational content and completed identical learning activities; however, the experimental group participated in a structured learning framework that included AI-powered intelligent assistants, guided learning activities, and instructor facilitation, while the control group completed the same activities without AI support. The educational program comprised five modules focusing on chatbot design, intelligent platform development, digital content production, educational data analysis, and future-oriented problem-solving. Learning outcomes were assessed using a Digital Entrepreneurship Product Rating Scale and a Future-Oriented Problem-Solving Scale, which measures skills in visualization, prediction, foresight, and planning. The results revealed statistically significant differences favoring the experimental group on both outcome scales, with large effect sizes. These findings suggest that integrating AI-powered intelligent assistants within a structured learning framework may support the development of digital entrepreneurship and future-oriented problem-solving skills among graduate students. This study contributes empirical evidence to the potential educational value of integrating AI-powered intelligent assistants with pedagogically designed learning activities to foster higher-level cognitive development in higher education.

Journal of Intelligence doi: 10.3390/jintelligence14080169

Authors: Xiaoming Song Jinmei Leng Xia Wu Sheng Yang Fang Luo

Emotion recognition is essential for social adaptation and mental health. However, existing measures still rely on limited stimulus materials or simplified task formats, which restricts ecological validity. The present study aimed to construct a localized multimodal emotion expression database in the Chinese context and to develop a multimodal emotion recognition ability test integrating emotion category and intensity recognition. In Study 1, a multimodal emotion expression database was established through actor performances, expert evaluation, and participant validation. The final database contained 2160 videos representing six basic emotions, five levels of emotional intensity, and two expression modes (vocal and non-vocal). In Study 2, 90 videos were selected under three recognition conditions (vocal-expression, muted vocal-expression, and muted non-vocal-expression). An integrated scoring approach combining emotion category and intensity was used to develop the test. A total of 236 valid participants completed the test. The results indicated that, after 61 video clips were retained, the test demonstrated good reliability and item discrimination. Evidence for criterion-related validity and construct validity was also obtained. Overall, the findings suggest that the present measure is a reliable and valid tool for assessing multimodal emotion recognition ability in the Chinese cultural context and offers the first empirical evidence that integrating category and intensity recognition can better capture the multidimensional nature of this ability.

Journal of Intelligence doi: 10.3390/jintelligence14080161

Authors: Visheeta J. Chandolia Matthew A. Kidd Morgan S. Paladino Steven M. Smith

Responses that come to mind as aha! moments occur suddenly, rapidly, and without deliberate intention. Aha! moments are typically associated with insight, the sudden realization of a problem’s solution, and have recently been reported in generative tasks, including divergent thinking and category generation. The present experiment compared the frequency and strength of aha! moments for responses in verbal problem solving (Compound Remote Associates and Polysemous Associates Test problems) with responses in ad hoc and taxonomic category generation. Comparisons were made using a two-step procedure in which each response was first reported as an aha! or non-aha! response, and, for aha! responses, rated in subjective strength. No difference in aha! strength was found between problem solutions and category responses, regardless of which task was given first, nor did aha! strength differ between ad hoc and taxonomic category members. In generative tasks, aha! strength, pre-response pause time, and response novelty increased over time during generation. Although fewer responses occurred over time, later responses were increasingly likely to be aha! responses. These results support the use of aha! reports during idea generation and link such reports to retrieval dynamics seen in many generative tasks, including divergent thinking.

Journal of Intelligence doi: 10.3390/jintelligence14080168

Authors: Fang Yuan Chunlei Liu Jiaqin Yang

The rapid proliferation of generative artificial intelligence (AI) has raised concerns about its impact on creativity, yet mechanisms and boundary conditions remain underexplored. Using one survey and two experiments, we investigated how AI dependence affects university students’ creativity. Study 1 (cross-sectional) found AI dependence was negatively related to creativity, fully mediated by critical thinking. Study 2 (between-subjects experiment) experimentally demonstrated that high AI dependence was associated with lower creativity than low dependence, with critical thinking again mediating. Study 3 employed a 2 × 2 design with thinking orientation as a moderator in a first-stage moderated mediation model. Thinking orientation moderated the “AI dependence → critical thinking → creativity” pathway: the indirect effect was significant under an outcome-oriented condition but non-significant under a process-oriented condition; the index of moderated mediation was 0.22. These findings indicate that critical thinking is a central pathway through which AI dependence undermines creativity, and process-oriented thinking buffers this negative impact by safeguarding critical thinking. The study provides theoretical foundations and intervention directions for the judicious use of AI in education.

Journal of Intelligence doi: 10.3390/jintelligence14080160

Authors: Jacqueline E. McGuinness Jonathan W. Schooler Shelly L. Gable Madeleine E. Gross

One of the defining characteristics of the Aha! experience is the profound sense of certainty with which it arrives—you don’t just know, you know you know. A large body of research suggests this confidence is generally warranted; ideas that arrive with a feeling of insight are more likely to be accurate than those that do not. However, the insight–accuracy relationship has been examined almost exclusively in convergent thinking tasks, leaving the diagnostic value of insights in divergent contexts uncertain. The present research tests whether this relationship differs between problem-solving contexts. Study 1 re-analyses daily diary data from professional writers and physicists, used as proxies for divergent and convergent domains. Writers initially rated their Aha! ideas as highly creative but downgraded them at 6-month follow-up; physicists’ Aha! ideas, by contrast, retained or gained value over time. Study 2 used a within-subjects design with laboratory measures of convergent and divergent thinking. Insight intensity predicted accuracy on both tasks but predicted subjective creativity ratings far more strongly than objective ones on the divergent task, producing a systematic bias we term the Aha! glow. Implications are discussed, including a proposed adaptive, motivational function for this phenomenon.

Journal of Intelligence doi: 10.3390/jintelligence14080167

Authors: Alessio Tesi

The present study examined how different combinations of promotion and prevention focus are associated with the experience of flow during fluid reasoning tasks of different difficulty, and whether flow is in turn associated with fluid intelligence assessed through a fluid reasoning task. Promotion focus reflects a self-regulatory orientation toward ideals, advancement, and gains, whereas prevention focus reflects an orientation toward duties, security, and non-losses. A total of 242 participants completed self-report measures of regulatory focus and task-related flow, along with a fluid reasoning task. One group completed a lower-demanding version of the task (N = 136), whereas the other completed a higher-demanding version (N = 106). A three-way moderated mediation analysis was conducted to test the indirect association between a promotion focus and fluid reasoning, mediated by flow, at the conditional level of high versus low levels of prevention for each difficulty condition. The results showed that, in the lower-demanding condition, promotion focus was more strongly positively associated with flow when prevention focus was low rather than high. In the higher-demanding condition, promotion focus was more strongly positively associated with flow when prevention focus was high rather than low. Across both conditions, higher flow was associated with better performance in the fluid reasoning task. Overall, the findings suggest that different motivational configurations may be differentially compatible with fluid reasoning task demands, and that flow may represent one process through which motivational orientations are associated with fluid intelligence.

Journal of Intelligence doi: 10.3390/jintelligence14080166

Authors: Norah Almulhim Abdullah M. Aljughaiman Fahad S. Alfaiz

This study examined whether a single total-score cutoff in the Mawhiba Multiple Cognitive Aptitude Test (MMCAT) may conceal meaningful cognitive profile differences between students across four domains: mental flexibility, verbal reasoning and reading comprehension, mathematical–spatial reasoning, and scientific–mechanical reasoning. Archival score data from 5146 school-nominated students across three assessment levels in Al-Ahsa, Saudi Arabia, were analyzed. Latent profile models were estimated separately for each assessment level. A three-profile solution was retained for Levels 1 and 2, and a four-profile solution for Level 3. The retained profiles differed mainly in the overall performance level rather than the profile shape. The operational cutoff of 1540 captured globally high-performing profiles at considerably higher rates than adjacent middle-range profiles, and the identified groups remained internally heterogeneous across all levels. The findings indicate that the current identification rule effectively captures broad, high-level performance, yet it does not represent all adjacent cognitive profiles equally. This has implications for identification policy and educational programming for gifted students.

Journal of Intelligence doi: 10.3390/jintelligence14080165

Authors: Jin-Long Liang Chin-Hsuan Hsieh Lung-An Shen Hsin-Chen Lu Ho-Tang Wu Xin-Ling Cui

This study examined longitudinal between-child and within-child associations between executive function (EF) and language expression (LE) in preschool children. Data were drawn from four waves of the Kids in Taiwan: National Longitudinal Study of Child Development and Care at 36, 48, 60, and 72 months (N = 1694). Longitudinal measurement invariance was examined before estimating a random intercept cross-lagged panel model (RI-CLPM), which separated stable between-child differences from within-child temporal fluctuations. EF and LE showed temporal continuity and positive concurrent associations. At the within-child level, occasions on which children displayed stronger LE than their own typical level consistently preceded relatively stronger EF one year later, whereas the reverse pathway was not consistently supported. These findings refine language-based accounts of self-regulation by suggesting that children’s everyday use of language to answer, explain, clarify, and organize information may be relevant to subsequent regulatory functioning. Practically, language-supported activities involving rule explanation, self-guidance, and reflective dialogue may be useful components of early EF support, although causal conclusions are not warranted from this observational study.

Journal of Intelligence doi: 10.3390/jintelligence14080159

Authors: Javier López Ana Belén Navarro Juan Carlos Meléndez Miguel Angel Millan

Depressive symptoms and cognitive functioning are intertwined in later life, but it is unclear whether their time-ordered associations differ by grandparenthood status. Using SHARE (2006–2022) data from adults aged ≥50 with consecutive observations, we fit lagged linear mixed-effects models with person-specific random intercepts and autoregressive adjustment, controlling for sociodemographic and health covariates and the time between waves. Higher prior EURO-D scores predicted slightly lower subsequent cognitive performance, and higher prior cognition predicted fewer subsequent depressive symptoms; in both directions, cross-domain effects were small relative to within-domain stability. Domain-specific analyses indicated the most consistent prospective links for memory and orientation, with weaker evidence for numeracy. Grandparenthood (having grandchildren) was not associated with the cognitive composite but was associated with slightly lower depressive symptoms. As a secondary aim, interaction models tested whether grandparenthood moderated the lagged associations; interactions were small and sensitive to specification, suggesting modest heterogeneity (including a crossover pattern for cognition predicting depression) rather than a uniform “enrichment” effect. Overall, depressive symptoms and cognition show modest bidirectional prospective links, and grandparenthood status marks limited, non-causal differences in affect–cognition coupling within an observational design.

Journal of Intelligence doi: 10.3390/jintelligence14080158

Authors: Renato Orti Michela Possenti Luigi Lorenzo Luca Napolitano Michela Romano Sabrina Iuliano Scila Nunziata Angelo Lucio Silvino Francesco Ruotolo

Understanding how environmental contexts influence cognitive functioning is a key issue in cognitive science and applied ergonomics. This study examined whether everyday environmental settings modulate task performance and perceived cognitive workload during a card-matching memory task, using high-fidelity immersive Virtual Reality (VR) environments. Participants performed the task by sequentially uncovering cards and remembering their spatial locations to identify matching pairs in two virtual scenarios: a naturalistic environment (urban park) and an artificial setting (office). The number of moves and time of execution were collected as behavioral measures. Perceived cognitive workload was assessed after task completion in each condition using the Simulation Task Load Index (SIM-TLX). Behavioral performances did not differ between environments. However, participants reported lower perceived cognitive workload when performing the task in the naturalistic environment. This effect was not uniform across individuals: the reduction in perceived cognitive workload was primarily observed among participants exhibiting high levels of task efficiency. These findings suggest that environmental context interacts with individual cognitive efficiency by modulating the perceived cognitive workload without necessarily affecting objective performance outcomes. Overall, the results suggest that naturalistic immersive virtual contexts may mitigate perceived cognitive workload during memory-based tasks, highlighting the role of contextual factors in influencing cognitive efforts under ecologically valid conditions.

Journal of Intelligence doi: 10.3390/jintelligence14080164

Authors: Ángel Romero-Martínez Carolina Sarrate-Costa Marisol Lila Luis Moya-Albiol

Background: The interplay between intervention dropout and increased recidivism among men convicted of intimate partner violence (IPV) underscores the need for a model that investigates the underlying variables associated with IPV perpetrator dropout. While sociodemographic factors are traditional predictors of dropout, the potential role of cognitive functioning remains insufficiently researched. This study examined whether certain neuropsychological domains, previously signaled by the scientific literature, including general intelligence, cognitive flexibility, planning abilities, verbal fluency, attention, and working memory, were associated with treatment attrition among IPV perpetrators. Methods: Participants (n = 309) underwent a comprehensive neuropsychological battery (K-BIT, WCST, BADS, CPT-III, and WAIS-III) prior to a court-mandated program. Results: The final regression model indicated that higher perseverative errors (WCST), lower nonverbal IQ, impaired planning abilities (Key Search Test), and a higher number of omissions in the CPT-III were significantly associated with dropout while controlling for alcohol misuse and sociodemographic variables. The overall predictive power of the neuropsychological model was moderate, offering a valuable, although complementary, framework to traditional models that only focus on sociodemographic variables. Conclusions: Early screening of neuropsychological variables could potentially help predict which individuals are at high risk of dropout. While these observational findings require external validation, integrating cognitive considerations into intervention planning warrants further investigation to understand their association with treatment adherence.

Journal of Intelligence doi: 10.3390/jintelligence14080163

Authors: Claudia Flores Débora Areces Celestino Rodríguez Inés López-Manrique

Divergent thinking is a key educational skill associated with cognitive and pre-literacy abilities. However, the limited research conducted with preschool populations makes it difficult to draw firm conclusions and to inform curriculum changes at early ages. This study aimed to (1) examine sex and school grade differences in divergent thinking at this stage, and (2) analyze the explanatory power of creative variables for preliteracy skills in second-cycle preschool students. A total of 80 children aged 3 to 5 years (M = 4.44; SD = 0.57) participated in this study. They completed an ad hoc figurative drawing test based on widely used instruments created for older age groups. More specifically, this test assessed fluency, flexibility, originality, and elaboration (Cronbach’s alpha = 0.89). Additionally, the AEI-R battery was administered to evaluate cognitive and pre-literacy skills. Multivariate analyses revealed only limited sex-related differences in divergent thinking variables, while no statistically significant differences emerged by grade level. Furthermore, hierarchical regression analysis indicated that creative elaboration offered incremental explanatory power, contributing uniquely and significantly to the variance in pre-literacy skills. Overall, these findings suggest associations between specific dimensions of divergent thinking and cognitive and pre-literacy variables in early childhood.

Journal of Intelligence doi: 10.3390/jintelligence14080157

Authors: Karen L. Siedlecki Veronika Kobrinsky

This study systematically examined whether long COVID negatively impacts subjective and objective measures of cognition in a community-based sample of 251 participants. Participants completed a battery of neurocognitive tasks (including assessments of short-term memory, working memory, episodic memory, processing speed, executive functioning, and reasoning) and provided self-ratings of their cognition (mental fatigue and subjective cognitive difficulties), depressive symptoms, the impact of symptoms on their life, and completed surveys assessing other psychosocial outcomes. Subgroups included individuals with no known lifetime history of COVID-19 illness (n = 66), those who had recovered from a short-term COVID infection (n = 119), and those who had been diagnosed with long COVID (n = 62). Analyses indicated that long COVID was associated with worse performance on measures of processing speed and executive functioning. In addition, those with a history of long COVID reported higher levels of mental fatigue and subjective cognitive difficulties compared to the other subgroups. Analyses within the long COVID subsample indicated that after controlling for covariates, depressive symptoms were significantly associated with mental fatigue and self-reported cognitive difficulties, and ratings of symptom impact also significantly predicted mental fatigue. These findings reinforce the value of incorporating both subjective and objective assessments when evaluating cognitive outcomes in long COVID, as each captures distinct aspects of functioning. Furthermore, these findings underscore the importance of assessing subjective cognitive concerns and functional symptom burden in clinical encounters, even when objective performance is generally unimpaired.

Journal of Intelligence doi: 10.3390/jintelligence14080156

Authors: Francisco D. Fernández-Martín Patricia Ayllón-Salas José L. Arco-Tirado

Service-learning interventions have been proposed as a context for examining trait emotional intelligence, a core socioemotional competence associated with cognitive and behavioral outcomes. This pilot quasi-experimental study examined whether participation in a PLACE model (Prepare, Link, Action, Celebrate, Effect) service-learning program was associated with differences in trait emotional intelligence; civic engagement attitudes and behaviors; academic motivation and perseverance; learning mindset; school belonging; dropout intentions; and grade point average among secondary school students. Participants were assigned to an intervention group (n = 30) or a control group (n = 27). Data were collected using validated instruments, including the Trait Emotional Intelligence Questionnaire, the Civic Engagement Scale, and the Students’ Engagement and Motivation Scale 3.0, along with a single-item measure of dropout intention and official grade point average records. Analyses of covariance, controlling for baseline scores, revealed no statistically significant group differences. Effect sizes showed a mixed pattern, with small effects favoring both groups depending on the outcome. These patterns should be interpreted cautiously, as the study was underpowered to detect small effects. Overall, the findings do not support intervention efficacy but may inform future research with larger samples.

Journal of Intelligence doi: 10.3390/jintelligence14080155

Authors: Xuping Xiong Fuhai An

The development of Artificial Intelligence (AI) poses significant challenges to creativity research and practice. Managing the degree of AI dependence during the creative process is therefore critical to fostering the healthy development of human creativity. To address this concern, this study proposes the CC-PACE model by integrating the 5A’s framework with the I-PACE model, and examines factors associated with AI dependence during the creative process. A cross-sectional survey was conducted with 813 college students in China, and the data were analyzed using partial least squares structural equation modeling (PLS-SEM). The results showed that (1) both creative personal identity and creative self-efficacy were positively associated with closeness with the audience and the perception of AI as a creative agent. (2) Both closeness with the audience and the perception of AI as a creative agent were positively associated with time spent using AI, usage preference, and perceived difficulty in avoiding AI use. (3) Time spent and usage preference were positively associated with students’ creative self-assessment, whereas perceived difficulty in avoiding AI use was not significantly associated with creative self-assessment. This study provides a viable analytical framework for understanding AI dependence in creative contexts. The findings further suggest that teachers should help students develop an appropriate understanding of AI’s role, strengthen students’ creative agency, and emphasize process-oriented assessment in creative activities.

Journal of Intelligence doi: 10.3390/jintelligence14080154

Authors: Şahin Danişman Filiz Evran Acar

This study investigated how pre-service teachers’ frequency of artificial intelligence (AI) tool use is associated with their perceived workload and perceived task performance in AI-supported pedagogical design tasks, grounded in mental workload theory. Using a quantitative cross-sectional survey design, data were collected online from 464 pre-service teachers through a structured questionnaire covering AI use profiles, purposes of AI use, perceived usefulness of AI functionalities, and NASA-TLX-based workload ratings. Perceived workload was assessed using the raw NASA-TLX after participants completed an AI-supported lesson planning task involving the design of an instructional activity and teaching materials. Descriptive findings showed that AI use was widespread, with most participants reporting weekly or daily engagement. One-way ANOVA results indicated significant differences across AI use-frequency groups on all NASA-TLX dimensions. Infrequent users reported higher mental demand, temporal demand, effort, and frustration, as well as lower perceived task performance. However, polynomial trend analyses showed that the association was not strictly linear across all dimensions. Some workload dimensions were lowest among occasional users, whereas frustration decreased and perceived task performance increased as AI use frequency increased. These findings suggest that the workload implications of AI engagement may vary depending on the specific dimension of task demand considered.

Journal of Intelligence doi: 10.3390/jintelligence14080153

Authors: Jiaqi Tang Huichao Bi Jun Wei

In the era of digital transformation and artificial intelligence, strengthening STEM education is indispensable to technological innovation and social progress. However, the scarcity of effective strategies and empirical evidence for the early identification of STEM-gifted learners constitutes a critical research gap. This study employed a systematic review approach to synthesize existing identification strategies for K-12 STEM-talent learners. Following the PRISMA framework, 26 peer-reviewed empirical and review articles were retrieved and analyzed from nine databases. Findings revealed that the majority of studies originated in the United States, peaked in 2018, and predominantly comprised empirical designs targeting high school students. Identification methods demonstrated an observed shift from holistic to more individualized approaches, encompassing intelligence and spatial ability testing, academic achievement measures, creativity and problem-solving assessments, portfolio evaluations, subject-specific competitions, and teacher or parent recommendations. Cognitive reasoning, creativity, spatial ability, and problem-solving emerged as core indicators, whereas socioeconomic status, cultural background, teacher and parental support, and gender expectations significantly influenced identification processes and talent development. Future research should empirically validate existing identification framework, account for cross-cultural variability, and examine the longitudinal efficacy of emerging artificial intelligence applications. This review synthesizes existing scholarship on K-12 STEM talent identification and delivers evidence-based recommendations to refine identification and assessment frameworks across diverse educational contexts.

Journal of Intelligence doi: 10.3390/jintelligence14080152

Authors: Mehmet Ertürk Geçici Fidan Geçici Metin Kaya

With the aim of determining the relationship between language and mathematics performance using a more general approach, this study examines the findings of primary meta-analyses using the meta-meta-analysis method. To achieve this, the study synthesized 12 independent primary meta-analyses representing an aggregated sample of more than 3.6 million participants. Because some primary studies overlapped across reviews, this figure reflects the cumulative rather than the unique number of participants. The results suggest a positive, moderate, and statistically significant relationship between language and mathematics performance (ES = 0.51, z = 19.20, p < .001). To explain the heterogeneity identified among the studies, the study further conducted moderator analysis on various categorical and continuous variables. The findings show that meta-analyses that focus on a specific language skill have larger effect sizes than studies that deal with language in general. The study also determined that the effect sizes differed statistically significantly across the language subgroups. A striking finding is that the relationship of reading, listening, writing, and vocabulary skills with mathematics performance is stronger than that of phonological processing skills. In conclusion, the study highlights the need to redesign mathematics instruction holistically, incorporating practices such as reading/listening comprehension, vocabulary development, and the effective use of mathematical language.

Journal of Intelligence doi: 10.3390/jintelligence14070151

Authors: Yawen Cheng Xiaoting Huang

Creative thinking has become an increasingly important competence in education, yet less is known about how it may be supported in human–AI collaboration as large language models (LLMs) are increasingly employed as partners in learning and problem solving. This study examined whether and how cognitive support and social interaction in AI-assisted collaboration were associated with divergent and convergent thinking. A total of 92 undergraduate students completed creative problem-solving tasks with GPT-4.0. Data were collected from task outputs and self-report measures, and the hypothesized associations were tested using structural equation modeling, with indirect effects examined as supplementary analyses. Results showed that in human–AI collaboration, social interaction rather than cognitive support was associated with more positive affective experiences, and this affective pathway was linked specifically to convergent thinking rather than divergent thinking. These findings suggest that associations between AI-assisted collaboration and creative thinking may depend less on perceived cognitive support alone than on the extent to which collaboration is experienced as socially engaging and affectively positive, particularly for convergent thinking.

Journal of Intelligence doi: 10.3390/jintelligence14070150

Authors: Deao Song Jien Guo Xuaner Rao Xinyu Hu Xinyuan Gu Junming Chen

Gamification increasingly influences learning experiences, cognitive engagement, and behavioral performance in digital learning, cognitive training, and health intervention contexts. However, the mechanisms underlying cognitive factors, along with related research hotspots and evolutionary trends, have not been adequately synthesized. Using the Web of Science Core Collection, this study analyzes 813 publications on gamification and cognitive factors published between 2012 and 2024. Using a bibliometric method, CiteSpace was utilized to analyze publication trends, collaborations, keyword co-occurrence, cluster structures, burst terms, cited references, and knowledge-map visualizations. The cluster analysis produced 10 interrelated themes: “flipped classroom,” “active learning,” “continuance intention,” “dementia,” “executive function,” “cognitive control training,” “computational thinking,” “cognitive training,” “cognitive load” and “user experience”. Potential future directions suggested by the bibliometric patterns include: (1) expanding gamification across educational contexts; (2) refining gamification theory models that focus on cognitive processes by examining user experience and cognitive load as potential mechanisms that link gamification design features to outcomes such as motivation, self-efficacy, task performance, and continuance intention; (3) promoting applications in cognitive training, cognitive impairment intervention, and digital health; (4) optimizing experimental design, data collection, interdisciplinary collaboration, and personalized design; and (5) clarifying how gamification shapes cognitive processes such as attention allocation, cognitive load regulation, problem solving, executive function, and computational thinking. This study does not aim to establish causal effects; rather, it uses bibliometric evidence to reveal the developmental trajectory, thematic structure, and emerging directions of research on gamification and cognitive factors.

Journal of Intelligence doi: 10.3390/jintelligence14070149

Authors: Nejla Gürefe Nursena Evrensel Yumrutepe

The examination of thinking skills has become a contemporary and imperative focus within educational research, capturing considerable attention from scholars. The present study endeavors to explore the mediating role of critical thinking tendency in the correlation between affective entry characteristics toward mathematics and reflective thinking skills related to problem-solving. This study was designed as quantitative research under the cross-sectional survey research design. The sample comprises 649 middle school students, with data collected through the Affective Entry Characteristics Towards Mathematics Scale, Critical Thinking Tendency Scale, and Reflective Thinking Skills Towards Problem Solving Scale, administered via an online survey platform. Mediation analyses were conducted. Results suggested statistically significant indirect associations among these constructs, with critical thinking tendency identified as a partial mediator between affective entry characteristics toward mathematics and reflective thinking skills for problem-solving. These findings suggest a complex interplay among these constructs, warranting further investigation in future studies.

Journal of Intelligence doi: 10.3390/jintelligence14070148

Authors: Aser Altalib

Trait emotional intelligence (EI) has been consistently associated with academic achievement, yet most empirical work has used global EI scores rather than examining how its four dimensions (well-being, self-control, emotionality, and sociability) contribute differentially. The present study addresses this gap through a dimensional analysis of trait EI as a correlate of English language achievement in a sample of 666 Saudi undergraduate students. Participants completed the Short Form of the Trait Emotional Intelligence Questionnaire (TEIQue-SF), and their final English course grades served as the measure of language achievement. Global trait EI correlated positively with achievement (r = 0.32), and the four dimensions jointly accounted for 16.6% of the variance, more than the 10.5% accounted for by the global score alone. Emotionality showed the strongest unique association with achievement (β = 0.35), followed by self-control (β = 0.13), while the well-being coefficient was negative in the unadjusted dimensional model (β = −0.14). After adjustment for gender, age, and academic major, emotionality and self-control remained the only significant EI dimension-level correlates of grade. Overall, the findings support reporting trait EI at both global and dimensional levels, with the strongest interpretation placed on emotionality and self-control.

Journal of Intelligence doi: 10.3390/jintelligence14070147

Authors: Kağan Kırcaburun

As generative artificial intelligence (GAI) technologies become increasingly integrated into university students’ academic routines, concerns have emerged regarding their associations with critical thinking dispositions during GAI use. The main objective of the present study was to examine whether metacognitive weakness and epistemic laziness individually and sequentially accounted for the association between GAI tool usage and critical thinking dispositions during GAI use across two distinct student samples. The Turkish-speaking sample consisted of 441 students, whereas the English-speaking sample included 218 students. Participants completed measures assessing GAI tool usage, critical thinking dispositions in GAI use, metacognitive weakness in GAI use, and epistemic laziness. Measurement invariance analyses indicated configural and metric invariance across groups. Path analyses revealed that GAI tool usage had significant indirect associations with lower critical thinking dispositions during GAI use via metacognitive weakness, epistemic laziness, and a sequential indirect pathway involving both constructs. The sequential indirect associations were significant in both the Turkish-speaking sample and the English-speaking sample. Overall, the findings suggest that greater GAI tool usage among university students is associated with lower critical thinking dispositions, with this association statistically carried through metacognitive weakness and epistemic laziness. The consistency of this pattern across two distinct student samples provides preliminary support for a process-oriented interpretation of how GAI use may relate to critical thinking dispositions.

Journal of Intelligence doi: 10.3390/jintelligence14070146

Authors: Şehriban Koca Çağlanur Sakarya Atakan Kutlu Alican Gülle Kıvılcım Pekpak Aslıhan Ergin

Instrumental learning in pre-professional music education involves complex associations among motivational orientations, cognitive strategies, and practice habits; however, their individual-level interrelations have received limited empirical attention in domain-specific research. This study examines these associations among pre-professional music students enrolled in Fine Arts High Schools (FAHS) in Türkiye (N = 332). A cross-sectional correlational design was employed, incorporating descriptive statistics, independent-samples t-tests, one-way ANOVAs with Scheffé post hoc tests, Pearson correlation analyses, multiple linear regression, and confirmatory factor analysis (CFA). The Instrument Learning Approaches Scale (ILAS) was used to assess four subscales: deep motive (DM), deep strategy (DS), surface motive (SM), and surface strategy (SS). Students predominantly reported deep motivational orientations (M = 3.57, SD = 1.19) and deep learning strategies (M = 3.47, SD = 1.21), with surface-oriented patterns reported at substantially lower levels (SM: M = 1.84, SD = 1.17; SS: M = 2.09, SD = 1.20). Small but statistically significant group differences were observed by gender (d = 0.25–0.26 for DM and DS) and grade level (ηp2 = 0.03 for DM). Daily practice time was positively associated with deep approach scores (r = 0.409 and r = 0.386, both p < .001) and negatively associated with surface strategy (r = −0.149, p = .007), disconfirming H4 for the surface strategy component. Regression modeling showed that DM and DS shared 70.2% of their variance (R2 = 0.702), while SM and SS shared 63.1% (R2 = 0.631). CFA revealed acceptable global fit (χ2/df = 1.886, RMSEA = 0.054, CFI = 0.930); however, the Fornell–Larcker discriminant validity criterion was violated for both within-approach factor pairs (DM–DS: r = 0.836 > √AVE; SM–SS: r = 0.789 > √AVE), and a competing two-factor model (deep vs. surface) fit the data equivalently [Δχ2(5) = 3.12, p = .681]. McDonald’s omega (ω = 0.793–0.895) and corrected item-total correlations (all r > 0.30) supported subscale-level reliability. These findings indicate that the ILAS functions as a two-dimensional rather than four-dimensional instrument in this pre-professional sample—a novel psychometric contribution. The most defensible empirical contributions are: the demonstration of two-dimensional ILAS collapse in this population; the independent association of practice duration with lower surface strategy endorsement; and small but replicable group differences by gender and grade level. Implications for ILAS validation, music pedagogy, and future longitudinal research are discussed.

Journal of Intelligence doi: 10.3390/jintelligence14070145

Authors: Meng Liu Yanlu Qi Mingliang Hu Ningxin Su Baike Li

An emerging body of research has shown that requiring participants to make concurrent JOLs during learning can reactively influence memory performance, a phenomenon known as the reactivity effect. However, it remains unclear whether making JOLs can also reactively alter metacognitive control. The present study addressed this question across three types of materials: unrelated word pairs, general knowledge questions, and related word pairs. After studying each item, participants decided whether they would choose to restudy it. In JOL conditions, participants made an item-by-item JOL during learning, while in no-JOL conditions, no such judgment was required. Across all materials, making JOLs reliably increased the proportion of items selected for restudy, indicating a robust reactivity effect on metacognitive control. These findings suggest that JOL reactivity should be considered not only in relation to memory performance but also in relation to metacognitive control.

Journal of Intelligence doi: 10.3390/jintelligence14070144

Authors: Maha Alfaleh

The rapid integration of generative artificial intelligence (GenAI) into education offers significant opportunities for talent development; however, current applications remain predominantly tool-centric and lack grounding in cognitive and intelligence theory. To address this limitation, the present study introduces the Human-AI Co-Competency Framework for Education (HACC-E), a human-centered, mastery-driven model intended to promote the synergistic development of creativity and cognitive intelligence. Drawing on contemporary intelligence research, including executive function theory, creative cognition models, and socio-emotional competence frameworks, HACC-E positions AI as a cognitive amplifier that supports higher-order reasoning, reflective thinking, and adaptive problem-solving, rather than as a replacement for human judgment. Employing a design science research methodology, the framework integrates competency-based progression, AI-augmented formative feedback, and simulation-based experiential learning within a governance structure that ensures human oversight and ethical accountability. The model aligns fluid reasoning, systems thinking, creative ideation, emotional regulation, and collaborative intelligence with structured AI-mediated feedback loops and authentic performance evidence. Analytical validation confirms pedagogical coherence, measurable competency mapping, and mitigation of critical AI-related risks, including automation bias and overreliance. By operationalizing distributed human–AI agency within a mastery-based progression architecture, HACC-E advances a theoretically grounded approach to AI-enabled talent development. This framework contributes to the ongoing discourse on human-centered artificial intelligence by providing a scalable and ethically robust model for cultivating cognitive, creative, and socio-emotional intelligence in educational systems designed for future readiness.

Journal of Intelligence doi: 10.3390/jintelligence14070143

Authors: Teófilo Félix Valentín Melgarejo Gastón Jeremías Oscátegui Nájera Dora Marina Hachoque Aguirre Ulises Espinoza Apolinario Isela Silvia Cruz Quinto Fidel Alberto García Yale Liz Ketty Bernaldo Faustino Clodoaldo Ramos Pando Josué Chacón Leandro Alexandra Rivas Meza Pablo Lenin La Madrid Vivar José Rovino Alvarez Lopez Pablo Lolo Valentín Melgarejo Flaviano Armando Zenteno Ruiz

Reading comprehension is a critical cognitive competency in higher education, although learners demonstrate substantial variability in responsiveness to metacognitive instructional interventions. The study focused on individual cognitive-response processes within the framework of the adaptive metacognitive reading system, which was realized through precision-learning architecture, which integrates latent learner-response phenotyping, explainable machine learning, Markov transition analysis, Bayesian adaptive inference, and reinforcement-learning optimization. The study employed a quasi-experimental longitudinal design involving an eight-week structured metacognitive reading intervention delivered through planning, monitoring, evaluation, strategic flexibility, and reading self-regulation activities. The psychometric analyses demonstrated satisfactory reliability of the adapted Metacognitive Awareness Inventory (MAI), with Cronbach’s α ranging from 0.83 to 0.89. A latent-profile model revealed significant heterogeneity of learner-response patterns among learners, with four learner-response phenotypes: High Responders, Strategic Improvers, Monitoring-Dependent Learners, and Low Responders. Explainable machine-learning models performed well in predicting individualized comprehension gains, with the model with the highest predictive accuracy being XGBoost (R2 = 0.61). Markov transition modeling identified exploratory learner-state redistribution patterns following the intervention. Bayesian adaptive inference and reinforcement-learning optimization were subsequently conducted as post hoc simulation procedures to estimate hypothetical adaptive instructional calibration scenarios rather than as real-time instructional decision systems. Overall, the proposed Adaptive Cognitive Intervention Architecture (ACIA) should be interpreted as an exploratory computational framework for modeling learner heterogeneity, predicting comprehension gains, and simulating post hoc computational optimization in higher-education learning environments.

Journal of Intelligence doi: 10.3390/jintelligence14070142

Authors: Christopher R. Niileksela Daniel B. Hajovsky Francisco Philippe Cocar-Montenegro William Joel Schneider Dawn P. Flanagan Vincent C. Alfonso

The relationships between cognitive abilities and mathematics skills are important to examine to help clarify how cognitive abilities promote mathematics development and inform why some individuals have difficulty acquiring mathematics skills. Meta-analyses of cognitive–mathematics relations can help summarize this research across a wide variety of test batteries and samples. This study compiled data from technical manuals of 122 norm-referenced standardized test batteries and included over 47,000 correlations from over 550 correlation matrices to summarize cognitive–mathematics relations using meta-analysis. The meta-analytic correlations were used to estimate cognitive–mathematics relations in an integrated model of cognitive abilities and mathematics skills. Fluid reasoning and comprehension-knowledge were two of the most consistent cognitive predictors of mathematics skills, and foundational mathematics skills (e.g., number sense and math fluency) were consistent predictors of advanced mathematics skills (e.g., math problem solving). A supplemental analysis examined how several narrow cognitive abilities (e.g., lexical knowledge, induction) predict mathematics skills. Results suggested that some narrow cognitive abilities, like general knowledge and general sequential reasoning, were consistent predictors of mathematics skills. This study summarizes a large amount of data from norm-referenced standardized test batteries to clarify how cognitive abilities predict mathematics skills. These results can inform both theoretical models of mathematics development and practical strategies when evaluating individuals who may have difficulty acquiring mathematics skills.

Journal of Intelligence doi: 10.3390/jintelligence14070141

Authors: Rossella Suriano Alessio Plebe

The increasing use of Neural Language Models (NLMs) calls for strategic and metacognitive skills in prompt formulation, monitoring, and adaptation. This study presents the Strategic Prompting Scale (SPS), designed to measure the extent to which users apply planning, adaptation, and evaluation strategies during interactions with generative language models. After expert review and pilot testing, an exploratory factor analysis with 187 participants revealed a three-factor structure—Planning, Adaptation, and Evaluation—explaining 60.4% of the total variance. The final 15-item scale demonstrated good internal consistency. Confirmatory factor analysis with an independent sample of 406 participants supported the proposed factor structure. The SPS also showed theoretically expected associations with related constructs, including positive correlations with metacognition and critical thinking, and negative correlations with disengagement in AI interaction. Overall, the findings provide preliminary psychometric evidence for the SPS as a theory-driven instrument to assess metacognitive regulation and self-regulated behaviors in generative AI use, with potential applications in educational, professional, and research contexts.

Journal of Intelligence doi: 10.3390/jintelligence14070140

Authors: Yuan Tao Hao Li Haisheng Fu Xiaotong Zhang Xiao Yu Andy Yu

The efficacy of working memory training (WMT) in facilitating academic achievement remains heavily debated. To resolve this efficacy paradox, the current study implemented an integrated adaptive WMT (IA-WMT) paradigm to investigate near transfer to inhibitory control (IC) and far transfer to mathematical abilities in school-age children. We assigned 89 typically developing children (Mage = 109.08 ± 3.90 months; 43 girls) to an IA-WMT group (29 students), an active control group (30 students), or a passive control group (30 students). Pretest and posttest assessments measured core working memory capacity, IC, exact calculation (EC), and approximate calculation (AC). Results revealed that six sessions of IA-WMT significantly enhanced visual–spatial and verbal working memory. Furthermore, the training induced potential near transfer to IC and asymmetric far transfer to EC, while yielding no specific gains for AC. This functional dissociation provides valuable alignment with the dual-system model of mathematical cognition. The findings suggest that far transfer may not operate as a generalized spillover effect but appears to be governed by the degree of functional overlap between trained executive processes and inherent task demands. Ultimately, organically integrated cognitive training represents a promising targeted intervention tool to facilitate cognitive and academic development during early school years.

Journal of Intelligence doi: 10.3390/jintelligence14070139

Authors: Mohammad Hamad Al-khresheh Suheyla Demirkol Orak Tariq Elyas Amal Abdullah Aboras

Research on AI-supported EFL learning has largely focused on learning outcomes, with limited attention to the cognitive mechanisms underlying creativity and the role of AI experience in shaping this process. This study examined how AI cognitive stimulation relates to EFL learners’ cognitive creativity, as measured by divergent thinking, with attention to the conditioning role of AI experience. Using a quantitative, correlational, cross-sectional design, data were collected from 878 undergraduate EFL learners recruited through stratified cluster sampling across four public universities in Saudi Arabia. Following instrument validation through exploratory and confirmatory factor analysis, structural equation modelling was employed to test direct, indirect, and conditional relationships among the study variables. The findings show that AI cognitive stimulation was strongly associated with divergent thinking (β = 0.52), and divergent thinking was strongly associated with cognitive creativity (β = 0.46). Although AI cognitive stimulation was directly associated with cognitive creativity, this effect was relatively limited. Mediation analysis showed that divergent thinking accounted for the primary pathway linking AI cognitive stimulation to cognitive creativity. Moderated mediation analysis further showed that learners’ duration of AI experience modestly but meaningfully conditioned this indirect pathway without altering its structural form. These results indicate that creative cognition in AI-supported EFL learning depends primarily on learner-controlled ideational processes rather than exposure to AI-generated content alone. The study contributes a process-oriented account of creativity in EFL contexts by specifying the cognitive pattern through which AI-related stimulation is associated with creative capacity and by clarifying how accumulated AI experience conditions this process.

Journal of Intelligence doi: 10.3390/jintelligence14070138

Authors: Yu Zhou Bin Jing Jing Zhang

As the capacity to monitor and regulate one’s own cognitive processes, metacognition is considered pivotal for mathematics learning and the development of higher-order thinking skills. However, research from a person-centered perspective remains limited regarding individual differences in metacognition and their associations with mathematics-related abilities. To address this gap, the present study surveyed 512 high school students and employed Latent Profile Analysis (LPA) to identify three distinct metacognitive profiles: “High metacognition,” “Medium metacognition,” and “Low metacognition,” accounting for 17.96%, 53.52%, and 28.52% of the sample, respectively. The BCH method revealed significant between-profile differences in computational thinking and mathematical modeling competency, with the “High metacognition” group demonstrating the strongest performance. Mediation analysis indicated that computational thinking partially mediated the relationship between metacognitive profiles and mathematical modeling competency, with the effect being most pronounced in the “High metacognition” group. These findings suggest substantial heterogeneity in students’ metacognition and highlight the potential relevance of prioritizing metacognitive cultivation in mathematics education. They further suggest that fostering computational thinking may be a productive route to strengthening mathematical modeling competency.

Journal of Intelligence doi: 10.3390/jintelligence14070137

Authors: Nermin Er Aydemir Ahmet Şahin

This study investigates the relationship between educational leadership, ICT-related factors, and students’ reading performance using the PISA 2022 dataset for Türkiye. Drawing data from 7250 students and 196 school principals, we employed multilevel structural equation modeling and Bayesian estimation. At the student level, between-school actions to sustain learning (SCHSUST) were significantly related to reading achievement. All student-level ICT variables—self-efficacy in digital competencies, practices regarding online information, and subject-related ICT use—significantly predicted reading achievement. Bayesian mediation analysis confirmed significant indirect relationships at student level, indicating that SCHSUST is associated with reading achievement primarily through ICT variables. Students’ economic, social and cultural status predicted reading achievement, indicating socioeconomic inequalities. At the school level, educational leadership (EDULEAD) has been found to be positively associated with preparedness for digital learning and school preparation for remote instruction. However, EDULEAD was negatively related to reading achievement. At the school level, all indirect relationships were insignificant. Furthermore, “economic, social and cultural status” and “academic school selectivity” emerged as the strongest predictors of reading achievement at their respective levels, indicating that the majority of reading inequality is due to substantial between-school inequality, in addition to the socioeconomic basis of digital inequality. Overall, the study highlights that meaningful ICT integration, rather than mere infrastructure provision, is associated with improved reading achievement.

Journal of Intelligence doi: 10.3390/jintelligence14070136

Authors: Constantinos M. Kokkinos Ioanna Voulgaridou Nafsika Antoniadou

Emotional intelligence and personality traits are key dimensions of individual differences contributing to developmental adjustment. This cross-sectional study investigated how trait emotional intelligence (trait EI), conceptualized as self-perceived dispositions closely embedded within personality hierarchies rather than a cognitive ability, is configured alongside the Big Five traits to influence socio-emotional functioning among preadolescents. Using a convenience sampling strategy, data were collected from 194 Greek preadolescents (aged 10–12) who completed self-report measures of trait EI, personality, and psychosocial adjustment via the Strengths and Difficulties Questionnaire (SDQ). Latent Profile Analysis (LPA) identified three distinct individual configurations: Emotionally Resourceful (54.6%), Emotionally Vulnerable (14.9%), and Emotionally Resilient (30.4%). Distal outcome testing via the Bolck-Croon-Hagenaars (BCH) three-step approach revealed that the Emotionally Resilient profile exhibited the most favorable adjustment (higher prosocial behavior, lower emotional/behavioral difficulties), whereas the Emotionally Vulnerable group displayed the highest psychosocial risks. While the reliance on cross-sectional self-reports requires cautious interpretation regarding causality, the findings underscore that preadolescent adjustment is better understood through integrated dispositional configurations rather than isolated traits, clarifying the role of trait EI within the broader personality space.

Journal of Intelligence doi: 10.3390/jintelligence14070135

Authors: Nurullah Aykaç

Drawing on research on self-regulated learning, metacognition, and cognitive aspects of writing, this study examines whether structured free writing can function as a scaffold for developing writing self-regulation, affective engagement, and genre-based performance. While free writing has traditionally been conceptualized as a creativity-enhancing or pre-writing activity, its role in fostering self-regulatory and metacognitive processes associated with intelligent performance remains underexplored. A convergent parallel mixed-methods design was employed with preservice teachers enrolled in a seven-week writing workshop in higher education. Quantitative data included validated measures of writing self-regulatory efficacy and attitudes toward free writing, along with analytic assessments of informative and narrative writing performance. Qualitative data were obtained through semi-structured interviews examining participants’ regulation of their writing processes. Findings indicate significant improvements in self-regulatory efficacy, more adaptive attitudes toward writing, and gains in genre-specific performance. Qualitative results further suggest that structured free writing supports metacognitive awareness, reduces writing-related anxiety, and enhances sustained cognitive engagement. By conceptualizing free writing as a mechanism for developing self-regulatory capacities underlying intelligent behavior, this study contributes to research on individual differences in learning and highlights the integration of cognitive and affective processes in academic performance.

Journal of Intelligence doi: 10.3390/jintelligence14070134

Authors: Adrian Büchli Jens F. Beckmann Stefan J. Troche

Dynamic testing (DT) aims to overcome a limitation of static assessment by providing learning opportunities as part of the test procedure so as to allow an individual’s latent potential to be evaluated rather than relying on a mere snapshot of their performance. Despite strong theoretical support, DT remains underutilized in research and applied settings. Thus, this study analyzes whether the Figural Analogies Learning Test (FALT), a newly developed short-term learning test (a DT procedure), is suitable for the assessment of latent potential in school-aged children. Mode of delivery for learning opportunities (i.e., feedback) was experimentally manipulated to investigate whether a digital or social modality results in differential effects on learning gains. A further aim was to establish whether—by capturing test takers’ latent potential—the FALT has incremental predictive utility in relation to academic achievement. The study randomly assigned 279 Swiss schoolchildren to one of three conditions: digital FALT, social FALT, or a static analogies task, all based on the same item set. Intervention effectiveness was evaluated using a static retest two weeks later. In both FALT conditions (digital and social), children achieved higher scores at the repeated assessment than in the static control condition. Moreover, socially mediated feedback produced significantly larger score gains than digitalized feedback. The three conditions did not differ in their ability to predict academic achievement and, thus, showed no incremental predictive utility for the FALT. However, an explorative exclusion of children with high grade averages indicated an advantage for the digital FALT relative to its static counterpart. The results provide tentative evidence that while both digitally and socially mediated feedback enhance learning performance, the latter is more effective. We discuss the absence of incremental predictive utility, specifically in the context of self-regulated learning, as well as implications for the design and diagnostic use of learning tests in educational contexts.

Journal of Intelligence doi: 10.3390/jintelligence14070133

Authors: Aljawharah Fahad Aljubilah Khaled Ahmed Abdel-Al Ibrahim Ahmad Al-Adwan Sayed M. Ismail Anwar Hammad Al-Rashidi Khalid Abdullah Alotaibi

Drawing on mindset theory, self-regulated learning theory, learner autonomy, and adaptive intelligence theory, this study tested an integrated structural model of the associations between academic growth mindset and self-perceived adaptive intelligence among Jordanian undergraduates. In this study, self-perceived adaptive intelligence refers to students’ perceived wisdom-related, social/practical, creative, and uncertainty-navigation tendencies; it is not an objective or performance-based measure of cognitive ability. The hypothesized sequential mediation structure was retained, but it was estimated and interpreted as a set of theoretically ordered indirect associations through learning motivation, metacognitive awareness, self-regulated learning strategies, and academic adaptation. Learner autonomy was examined as a moderator of the association between self-regulated learning strategies and academic adaptation. Data were obtained from 640 undergraduate students enrolled in Jordanian universities and analyzed using partial least squares structural equation modeling with WarpPLS 8.0. Academic growth mindset was positively associated with learning motivation and metacognitive awareness. Both constructs were positively associated with self-regulated learning strategies, which, in turn, were positively associated with academic adaptation; academic adaptation, in turn, was positively associated with self-perceived adaptive intelligence. The theoretically ordered sequential indirect associations through the motivational and metacognitive routes were statistically significant, whereas learner autonomy did not significantly moderate the association between self-regulated learning strategies and academic adaptation. Because the data were single-wave and self-reported, the term “sequential” refers to the theory-imposed ordering of paths in the statistical model, not to an observed temporal or developmental progression. Accordingly, the findings represent structural associations and do not establish causal sequencing. The findings contribute to intelligence and higher-education research by distinguishing domain-specific academic adaptation from broader self-perceived adaptive intelligence informed by Sternberg’s adaptive intelligence framework.

Journal of Intelligence doi: 10.3390/jintelligence14070131

Authors: Carolina Bello-Correas Teresa Alzás Laura Alonso-Díaz

This systematic review, following PRISMA 2020 guidelines, analyzes scientific literature on bullying and cyberbullying among adolescents (aged 12–16) in educational settings, focusing on ICT, smartphone hyperconnectivity, and emotional education. An exhaustive search across Web of Science, Scopus, PubMed, Dialnet, CSIC, SciELO, and Google Scholar identified 34 empirical studies. A narrative synthesis was performed due to methodological heterogeneity. The synthesized evidence suggests that cyberbullying frequently acts as a persistent extension of school violence, where continuous digital exposure makes it difficult for victims to emotionally disconnect. Empirical data indicate a concerning correlation between prolonged bullying and psychosocial distress, including self-harming behaviors and suicidal ideation. Furthermore, results highlight systemic gaps: heightened vulnerability is reported among girls and LGBTQ+ students, alongside disparities between public and state-subsidized schools regarding institutional involvement and emotional support resources. Educational implications suggest reactive protocols are insufficient. Evidence supports that systematic emotional education, enhancing socio-emotional skills like impulse control, empathy, self-esteem and emotional regulation, acts as a key protective factor. Consequently, fostering “digital emotional intelligence” emerges as a promising preventive educational strategy to protect adolescents’ well-being in hyperconnected environments.

Journal of Intelligence doi: 10.3390/jintelligence14070132

Authors: María Auxiliadora Robles-Bello Francisca Barba-Colmenero Jonathan Vinicio Camino-Alarcón Nieves Valencia-Naranjo

Background: Sensory processing differences are highly prevalent in children with autism spectrum disorder (ASD) and may shape how they interact with their environment and acquire new knowledge. The present study investigated developmental changes in learning potential and cognitive functioning in preschool children with ASD presenting a sensory seeking profile with typical sensitivity. In addition, the study examined which components of early learning potential predicted intellectual functioning two years later. Methods: A longitudinal design was employed with 44 preschool boys with ASD assessed at ages four and six. The potential for learning through dynamic assessment, intelligence and sensory processing patterns was assessed. Paired sample t-tests were conducted to examine developmental changes, and hierarchical regression analyses were used to identify predictors of intellectual functioning at age six. Results: Significant improvements were observed in general cognitive ability, classification ability, and perspective-taking skills across the two-year period. Regression analyses revealed that visual memory, sequential pattern completion, classification ability, and perspective taking significantly predicted intellectual functioning at age six, jointly explaining 51% of the variance in IQ scores. Conclusions: The findings suggest that visuospatial processing and pattern-based reasoning may play a central role in cognitive development among preschool children with ASD who exhibit active sensory engagement.

Journal of Intelligence doi: 10.3390/jintelligence14070130

Authors: Eduardo Melguizo-Ibáñez Pilar Puertas-Molero Gabriel González-Valero José Manuel Alonso-Vargas

Physical activity-based methodologies have been proposed as educational tools to improve executive functions, although their impact on early years and primary education has not yet been fully assessed. This study aims (a) to conduct a systematic review of intervention programmes applying physically active methodologies to inhibitory control in early childhood and primary education and (b) to calculate the overall effect size of intervention programmes applying physically active methodologies to inhibitory control. A systematic search was conducted in the Scopus, Web of Science, Eric and PsycINFO databases. The search yielded a total of 16 scientific articles that formed the basis of the quantitative synthesis. The meta-analysis showed a small overall effect size (g = 0.25; 95% CI [0.11; 0.39] p = 0.0006). Meta-regression analyses indicated that the type of intervention and session length were not significantly associated with the effect size. It was found that the number of sessions was significantly associated with the effect size (β = 0.0017; p = 0.010). As a conclusion, physically active methodologies may be associated with small improvement in inhibitory control; the certainty of evidence is very low. Confidence in these findings is limited, and the true effect may differ.

Journal of Intelligence doi: 10.3390/jintelligence14070129

Authors: Roberto Burro Giada Vicentini Giada Mignolli Anna Maria Meneghini Daniela Raccanello

The ability to effectively use different emotion regulation (ER) strategies across situations—i.e., ER flexibility (ERF)—plays a key role in emotional intelligence and social functioning. However, it has rarely been studied in young people, partly due to the lack of age-appropriate measurement instruments. This study investigated children’s perceptions of ERF by adapting an adult scale and proposing the 10-item Emotion Regulation Flexibility Questionnaire for Children (ERFQ-C). A sample of 346 children aged 9–12 years was recruited, who completed the ERFQ-C along with questionnaires assessing emotion malleability and school-related well-being. First, the analyses confirmed the scale’s unidimensional structure and measurement invariance across sex and school level. Second, the scale showed adequate fit to the Rasch model, and item scores were converted into Rasch scores. Third, path analyses indicated that emotion malleability was positively associated with ERF perception, which, in turn, was positively associated with school-related well-being. Notwithstanding the limitations of self-report data, the instrument can be used in future research to advance theoretical understanding of the intertwinement between perceived ERF and other constructs related to children’s functioning (e.g., emotional intelligence) and inform interventions aimed at improving youth well-being.

Journal of Intelligence doi: 10.3390/jintelligence14070128

Authors: Sarah K. C. Dygert Andrew F. Jarosz

Resolving misrepresentations is key when faced with ambiguous information. For example, problem-solvers may misrepresent the constraints of a problem, while readers may misrepresent parts of a sentence. This work investigates how a shared cognitive process might facilitate the restructuring of representations during complex tasks. Students from Mississippi State University participated in studies demonstrating that similar restructuring processes are necessary for successful creative problem-solving and garden path sentence comprehension. Experiment I demonstrated that successful creative problem-solving correlates with successful garden path sentence resolution after removing the variance of non-ambiguous sentence comprehension. Experiment II demonstrated that successful garden path sentence comprehension predicts creative problem-solving, even after controlling for working memory capacity and fluid intelligence. Furthermore, this relationship was unique to creative problem-solving, as garden path comprehension had no bearing on analytic problem-solving success. Results are taken as evidence for a shared cognitive process that aids in revising misrepresented information across certain contexts. This work provides an understanding of how individuals grapple with inconsistencies between the goals of their tasks and their representations of the world.

Journal of Intelligence doi: 10.3390/jintelligence14070127

Authors: Xueheng Zhou Honggang Liu

This study examined learner engagement in AI-supported informal digital learning of English (IDLE) by integrating L2 selves and control-value appraisals within a Control-Value Theory framework. Drawing on data from 653 Chinese university students, the study employed structural equation modeling (SEM) to test the associations among ideal L2 self, ought-to L2 self, control appraisal, intrinsic value appraisal, extrinsic value appraisal, and learner engagement. It further used fuzzy-set qualitative comparative analysis (fsQCA) to identify configurational pathways associated with high engagement. The SEM results showed that both ideal L2 self and ought-to L2 self were positively associated with the three types of control-value appraisals, which were in turn positively associated with engagement. Mediation analysis indicated that control appraisal, intrinsic value appraisal, and extrinsic value appraisal mediated the associations between L2 selves and engagement. The fsQCA results revealed six configurations associated with high engagement, suggesting that engagement may be linked to multiple motivational-appraisal patterns. These findings highlight the importance of both average associations and configurational diversity in understanding AI-supported IDLE engagement.

Journal of Intelligence doi: 10.3390/jintelligence14070126

Authors: Minbo Li Yunyi Zhao Xin Shan Xiaofei Du

Although artificial intelligence (AI) is reshaping music education, outcome-oriented quantitative syntheses remain relatively limited. This mixed-methods review examined the effects of AI-assisted music learning on learning motivation, self-regulated learning (SRL), and creative performance, while identifying learner-, task-, and time-related boundary conditions and clarifying how AI support is implemented and experienced. A three-level meta-analysis was used for quantitative integration, complemented by a qualitative synthesis of implementation pathways and learner experiences. Results showed positive trends across all three core domains, with different levels of statistical support and substantial heterogeneity. Learning motivation showed the most robust evidence (g = 1.28), creative performance showed a larger but highly heterogeneous effect (g = 1.21), and SRL showed a preliminary positive trend (g = 0.57). The task complexity × prior ability interaction provided tentative, directional evidence for learner–task fit, mainly for motivational outcomes. Dose-related analyses suggested a possible asynchronous pattern: motivational gains may emerge rapidly in the short term, whereas gains in higher-order cognition may become more evident under sustained intervention. Qualitative synthesis identified three AI implementation pathways—evaluative feedback, generative support, and adaptive personalization—suggesting that effectiveness depends less on technological complexity itself than on aligning AI roles with task demands and learner needs. Future research should strengthen long-term designs, deepen SRL-related evidence, and examine the adaptive effects of different AI roles across diverse music learning contexts.

Journal of Intelligence doi: 10.3390/jintelligence14070125

Authors: Aylin Akinlar

This study presents a bibliometric and thematic analysis of intercultural intelligence research in digitally mediated contexts. Using a dataset of 92 journal articles indexed in the Web of Science Core Collection between 2006 and April 2026, the study maps the intellectual structure, collaboration patterns, and thematic evolution of the field. Data were analyzed using the Bibliometrix R package (version 5.3.0) and Biblioshiny through performance analysis, science mapping, co-citation analysis, collaboration networks, and thematic mapping techniques. The findings indicate that research output has expanded considerably since the mid-2010s, reflecting the growing importance of intercultural competencies in digitally connected environments. The intellectual structure of the field remains strongly anchored in cultural intelligence theory while increasingly incorporating themes related to communication, trust, global virtual teams, leadership, and digitally mediated collaboration. Thematic evolution analyses reveal a gradual shift from traditional cultural intelligence constructs toward broader concerns associated with virtual interaction, digitally mediated collaboration, and emerging AI-supported communication environments. The results also demonstrate a highly collaborative and internationally connected research landscape, although notable geographical imbalances remain. By providing a systematic overview of the field, this study contributes to a deeper understanding of how intercultural intelligence scholarship is evolving in response to digital transformation and offers directions for future interdisciplinary research.

Journal of Intelligence doi: 10.3390/jintelligence14070124

Authors: Caterina Padulo Anna Cascone Francesco De Crescenzo Onofrio Gigliotta Beth Fairfield

The verbal working memory training proposed by Borella and co-authors found specific and transfer effects among older adults. However, the effective training lengths needed to maximize transfer effects are not yet clear. Also, far-transfer effects related to psychological well-being and subjective health are still under debate. The present study aimed to assess gains and transfer effects of a modified version of the WM training protocol by Borella and co-authors by comparing the original three-session (1 h each) version to a modified eighteen-session (1 h each) version. Our results confirmed the already demonstrated specific cognitive effects that seem to increase as the number of sessions increases. Regarding psychological well-being and subjective health, we found that while even three sessions of training can diminish reported loneliness and negative affective states, the longer training significantly improves the subjective perception of general health, suggesting that longer working memory training may be particularly fruitful in promoting well-being and successful aging.

Journal of Intelligence doi: 10.3390/jintelligence14070123

Authors: Ángela Lori Eva María Padilla-Muñoz Francisco Javier Cano-García Antonino Raffone

In recent decades, the educational landscape has undergone rapid transformations, highlighting emotional intelligence (EI) and mindfulness as key constructs for students’ well-being and adaptation. This study examines the impact of a structured 12-week school-based mindfulness programme on emotional intelligence in a cross-national sample of students. The sample comprised 207 participants (52.7% from Spain and 47.3% from Italy). In Spain, the mean age was 8.65 years (SD = 0.36) in primary education and 13.71 years (SD = 0.49) in secondary education; in Italy, the mean age was 8.09 years (SD = 0.30) and 13.33 years (SD = 0.43), respectively. Using a quasi-experimental repeated-measures design, EI was assessed at pre-test, post-test, and 5-month follow-up with the EQ-i:YV and analysed using General Linear Models. In line with these objectives, results indicated no significant increase in global EI (p > .05), although small, domain-specific effects emerged (η2p ≈ .03–.09). Italian primary students showed improvements in General Mood and Interpersonal Competence, while Italian secondary students displayed delayed gains in Stress Management. No significant effects were observed in Spanish subsamples or in Adaptability and Intrapersonal Competence. These findings indicate modest, context-dependent effects, with limited and context-dependent educational implications.

Journal of Intelligence doi: 10.3390/jintelligence14070122

Authors: Dominik Weber Nicolas Becker Frank M. Spinath Marco Koch

In applied diagnostics, scores from multiple intelligence tests are often combined by simple arithmetic averaging. Although convenient, this practice is statistically imprecise: it neglects (a) that tests are positively correlated but not identical, each capturing somewhat different aspects of intelligence, and (b) that combining correlated measures reduces variance. Consequently, arithmetic means underestimate intelligence above the IQ scale center and overestimate it below the center. Such distortions can lead to serious misjudgments in high-stakes assessment, such as in educational placement or forensic evaluations of legal culpability. To obtain exploratory evidence on the prevalence of simple averaging practice, we surveyed n = 75 psychologically educated individuals familiar with IQ metrics. In response to a case vignette requiring the combination of several IQ scores, 58 participants (77.33%) applied simple averaging, and none considered intercorrelations and variance reduction. Therefore, the aim of this article was to propose the application of a more valid method for combining scores (as well as corresponding reliabilities and validities) of multiple tests. To validate the method, we performed Monte Carlo simulations, covering a wide range of test characteristics. Results showed virtually perfect accuracy (r = 1.00, p < .001), and the outcomes were robust against variations in the number of tests to be combined, score distributions, and intercorrelations. To facilitate adoption in practical diagnostics, we developed and introduced an open-source R package and an accompanying open-access Shiny web application.

Journal of Intelligence doi: 10.3390/jintelligence14070121

Authors: Daniela Cecic-Mladinic María del Carmen Carcelén-Fraile Diana Marcela Aristizábal-García Noelia Vigil-Torres

Self-regulation is a foundational developmental skill in early childhood that supports academic readiness, social competence, and long-term health outcomes. Physical activity has been proposed as a modifiable behavior that may contribute to self-regulation development; however, evidence in preschool-aged children remains inconsistent and has primarily focused on cognitive outcomes. This systematic review and meta-analysis aimed to examine the association between physical activity and self-regulation across domains in early childhood using cross-sectional evidence. The review was conducted in accordance with PRISMA guidelines and registered in PROSPERO (CRD420261361512). Searches were conducted in PubMed, ERIC, CINAHL, and Web of Science. Fifteen studies met the inclusion criteria and were included in the qualitative synthesis and meta-analysis. A random-effects meta-analysis revealed a small but statistically significant positive association between physical activity and self-regulation (r = 0.099, 95% CI = 0.064–0.133, p < 0.001), although substantial heterogeneity was observed (I2 = 91.1%). Subgroup analyses showed significant associations for both cognitive and non-cognitive self-regulation outcomes. Sensitivity analyses confirmed the robustness of the findings, and no substantial publication bias was detected. Physical activity may represent a promising modifiable factor that supports self-regulation development in early childhood, although longitudinal and experimental research is needed to clarify causal relationships.

Journal of Intelligence doi: 10.3390/jintelligence14070120

Authors: Yixuan Xu Hanwei Wu

English proficiency is vital for non-native speakers’ career development, yet classroom instruction alone cannot meet practical demands, making informal digital learning of English (IDLE) increasingly important. Artificial intelligence (AI), with conversational and multimodal functions, offers new opportunities for IDLE. However, existing research on AI-mediated IDLE has predominantly focused on language majors and often relied on a single methodological lens, neglecting STEM undergraduates and the complex network dynamics among motivational factors. However, research has largely focused on language majors, leaving STEM majors underexplored. Guided by the Hedonic-Motivation System Adoption Model (HMSAM), this study analyzed data from 413 Chinese STEM majors using partial least squares structural equation modeling (PLS-SEM, SmartPLS 4.0) and psychological network analysis (PNA, R 4.5.3). PLS-SEM results showed that enjoyment was the strongest direct predictor of AI-IDLE, followed by focused immersion, perceived usefulness, and curiosity. Control contributed indirectly via focused immersion, while boredom was non-significant. Perceived ease of use influenced AI-IDLE only through cognitive and emotional pathways. The model explained 58.1% of the variance. PNA further identified enjoyment, focused immersion, and control as central nodes, while the link between perceived usefulness and AI-IDLE was non-significant. These findings suggest that Chinese STEM undergraduates’ AI-IDLE is primarily driven by intrinsic hedonic motivations rather than utilitarian evaluations. The study provides empirical support for designing AI tools that enhance enjoyment and control to foster STEM students’ extracurricular English engagement.

Journal of Intelligence doi: 10.3390/jintelligence14070119

Authors: Simge Karakaş Mısır

The present study investigates the semantic structure, dominant themes, and temporal evolution of research on giftedness in early childhood through a comparative topic modeling approach. A final analytic sample (n = 518) of peer-reviewed journal articles indexed in the Scopus and Web of Science databases was analyzed. Three topic modeling methods, Latent Dirichlet Allocation (LDA), Structural Topic Modeling (STM), and BERTopic, were systematically compared using multiple evaluation metrics. BERTopic demonstrated the strongest overall performance, producing approximately 11% higher coherence than STM and approximately 34% higher coherence than LDA. In terms of diversity, it achieved 14% to 17% greater thematic variety and, according to the Gini coefficient, revealed a 58% to 60% more balanced thematic distribution. BERTopic-based analyses identified five major thematic axes: Socio-Linguistic Development and Family Context, Psychometric Intelligence, Identification, and Cognitive Differences, Program Access, Identification, and Educational Equity, Early Academic Skills and Cognitive Development, and Creativity, Higher-Order Thinking, and Enrichment Programs. Thematic mapping and topic similarity analysis were used to examine the semantic structure of the field, while linear regression-based trend analysis over the 1918–2026 publication period showed that family context, socio-linguistic development, and equity-related themes have gained increasing importance over time, whereas psychometric identification largely maintained its central position within the field. These findings indicate that the field is moving toward a more inclusive, semantically grounded, and equity-oriented perspective. However, they should be interpreted in light of the study’s reliance on article abstracts, the sensitivity of BERTopic clustering parameters, and the use of linear trend modeling.

Journal of Intelligence doi: 10.3390/jintelligence14070118

Authors: Wenjin Zhang Yoed N. Kenett Mujie Ma Peiduo Liu Wenjing Yang

Scientific creativity depends not only on what learners know but also on how knowledge is organized, yet evidence in scientific domains remains limited. Across two experiments, we examined whether knowledge structures predict scientific divergent thinking and whether knowledge form shapes creativity through its effects on these structures. Undergraduate students completed the Scientific Divergent Application Task (SDAT), which requires generating novel applications based on learned scientific principles. In Pilot Experiment (n = 39), semantic memory networks constructed from participants’ knowledge representations revealed that network connectivity and efficiency were positively associated with better SDAT performance, whereas recall accuracy was unrelated. In Experiment 2 (n = 126), holding informational content constant, knowledge forms significantly influenced performance: both associative and schematic knowledge promoted scientific divergent thinking more effectively than case-based knowledge, whereas schematic knowledge further demonstrated advantages in originality and knowledge network connectivity and efficiency. These findings suggest that knowledge organization, rather than retention alone, underlies individual differences in scientific creativity, and that schematic presentation may support creative application by fostering more efficient knowledge structures.

Journal of Intelligence doi: 10.3390/jintelligence14070117

Authors: Rui Wang Kastoori Kalaivanan Jiani Ren Shen-Hsing Annabel Chen Chew Lee Teo

Matrix reasoning tasks remain among the most widely used instruments for assessing abstract reasoning and are often assumed to be culturally neutral. However, this assumption has been challenged by studies reporting significant cross-cultural variation in performance on nonverbal matrix reasoning tasks, even when groups show comparable performance on verbal measures of general cognitive ability. One plausible reason is that many matrix reasoning tasks rely primarily on accuracy-based performance metrics while providing limited insight into response timing and task engagement during problem solving. The present study examined the Matrix Reasoning Item Bank (MaRs-IB), a new online matrix reasoning instrument integrating both accuracy and response time, in 458 participants from China and the UK. Results demonstrated strong psychometric properties across both cultural contexts, while also revealing systematic between-group differences in overall task performance. Chinese participants were generally slower but more accurate, whereas UK participants responded more quickly with lower overall accuracy. Rather than reflecting a classical speed–accuracy trade-off, these patterns may indicate cross-cultural variation in persistence, deliberative engagement, and the metacognitive regulation of cognitive effort during reasoning tasks. In particular, Chinese participants allocated more time before responding and persisted longer on challenging task items, whereas UK participants demonstrated relatively faster responding and shorter response times on more challenging items. These findings suggest that cross-cultural differences in matrix reasoning performance may reflect not only differences in observed performance levels, but also variation in how participants allocate time and sustain engagement during cognitively demanding tasks.