Learning from multiple, partially conflicting documents places high demands on both comprehension and metacomprehension. This study examined how two instructional supports—a generative task and external standards—influence metacomprehension accuracy and regulation in multiple-document reading. Building on prior research showing that both supports can independently improve metacomprehension accuracy, we investigated whether combining them yields additive benefits. A total of N = 163 university undergraduates read multiple documents on a historical controversy and were randomly assigned to one of three conditions: generating forum posts and comparing them with external standards, generating forum posts only, or viewing external standards only. Participants’ metacomprehension judgments and comprehension measures were assessed for essay performance, historical content knowledge, intertextual conflicts (contradictory information across documents), and source–content links (connections between claims and their sources). Regulation was measured via re-reading intentions. The results showed that generating forum posts alone did not improve metacomprehension accuracy. When external standards were added, learners judged their essay performance more accurately than learners who received external standards only, though no benefits emerged for the other comprehension measures. Similarly, learners in the combined condition demonstrated better regulation of source–content links than the expert-only group, whereas other regulation measures were unaffected. In addition, learners relied more strongly on their self-generated forum posts as cues to their comprehension when no external standards were provided. Together, these findings provide some support for the assumption that combining generative tasks with external standards can strengthen metacomprehension accuracy in complex writing tasks. Generative tasks alone can offer learners useful cues, but external standards appear necessary to foster more accurate metacomprehension and regulation.
Teachers play a crucial role in fostering students’ self-regulated learning (SRL) skills in the classroom. Previous research indicates that teachers adopt diverse approaches to promoting SRL, but empirical evidence on subject-specific differences remains limited. This study examines the extent to which native language and science teachers promote SRL and whether their approaches differ. To explore teachers’ SRL promotion, we analysed video recordings of their classroom behaviour, their self-reported SRL promotion, and students’ perceptions of their teachers’ SRL promotion. The study involved 28 teachers and 515 students, with data collected at two measurement points. Analyses of observational and questionnaire data revealed that teachers of both subjects promoted SRL predominantly implicitly, focusing mainly on metacognitive and cognitive strategies, while emotional regulation strategies were largely neglected. Overall, SRL promotion was implemented only to a limited extent and occurred similarly across subjects. Although comparisons of several direct and indirect SRL promotion approaches indicated some variation, no systematic subject-specific differences emerged. Moreover, correlations between the three perspectives, observations, teachers’ self-reports, and students’ perceptions, were generally low, suggesting that each perspective captures distinct and complementary aspects of SRL promotion. Taken together, these findings indicate that variation in SRL promotion may be more strongly related to individual teacher characteristics, task characteristics, or lesson type than to the subject domain. Future research on potential differences in SRL promotion would benefit from adopting multi-perspective approaches to gain a more comprehensive understanding of SRL in classroom practice, as well as from incorporating systematic variation in the observation of SRL over longer periods.
Many studies have found an association between instruction quality and digital reading performance. This study aims to investigate whether this association varies across self-regulated learning (SRL) profiles and whether instruction quality is indirectly associated with digital reading performance through SRL profiles. Participants were 12,058 fifteen-year-old students from Mainland China who participated in the Programme for International Student Assessment (PISA) 2018. Latent profile analysis was conducted to identify SRL profiles and examine their roles in the association between instruction quality and digital reading performance. Four SRL profiles emerged. Profiles High SRL and Low SRL represented the best and the weakest levels of all SRL variables, respectively. Profile High Metacognitive Knowledge was characterized by high metacognitive knowledge of reading strategies but low motivation. Profile Strong Motivation-Low Feeling of Difficulty was characterized by high motivation and low levels of feeling of difficulty and metacognitive knowledge. Stimulation of engagement, disciplinary climate, and teacher enthusiasm were positively associated with profiles exhibiting higher motivation, while teacher-directed instruction and teacher feedback were negatively associated with profiles exhibiting higher metacognition. Adaptive instruction and teacher support showed weak associations with SRL profiles. The association between instruction quality and the performance varied significantly across SRL profiles. No instruction quality variable was consistently associated with performance across all profiles. For teacher-directed instruction, stimulation of engagement, and disciplinary climate, both direct associations with digital reading performance and indirect associations via SRL profiles were statistically significant. For teacher enthusiasm, only the indirect association was significant. For adaptive instruction, teacher support, and teacher feedback, no significant indirect association was found. The findings suggest the potential moderating and mediating roles of SRL profiles in the association between instruction quality and digital reading performance, but they should be interpreted as statistical associations rather than causal effects.
This study investigates the relationship between self-assessment and self-regulated learning (SRL) in upper elementary education, focusing on the distinctions between product and process-based self-assessment and the role of self-assessment accuracy. A four-week intervention involving 187 students explored whether self-assessment practices enhanced SRL across planning, monitoring, control, and reflection phases. Results revealed that practicing a combination of product- and process-based self-assessment led to greater improvements in self-assessment accuracy over time, compared to practicing only product-based assessment. However, no direct improvements in SRL behaviors were observed, suggesting that self-assessment alone may not suffice to foster comprehensive SRL development. Findings highlight the nuanced relationship between self-assessment accuracy and SRL, suggesting that self-assessment may need to be complemented by additional forms of SRL support to yield broader improvements in SRL.
Self-assessment (SA) is central to self-regulated learning, yet scalable instruments for capturing learners’ self-reported SA practices in domain-specific writing contexts remain limited. This study developed and validated the Writing Self-assessment Practice Questionnaire for EFL Learners (WSAPQ-EFL), designed to measure how learners report establishing evaluative standards, monitoring writing quality, and forming judgments during English writing self-assessment. Drawing on process-sensitive qualitative findings from a preliminary phase using think-aloud protocols and stimulated-recall interviews (Ou, 2026), the questionnaire was administered to 1,180 Chinese university English majors in the psychometric validation phase. Exploratory and confirmatory factor analyses on independent subsamples (EFA, N = 661; CFA, N = 519) supported a 61-item, 14-factor structure organized within three stages: Determining Standards, Monitoring/Cue-seeking, and Cognitive Judgment. This structure captures reported dimensions of self-regulated learning and dual-process-informed judgment by distinguishing learners’ reported reliance on criterion sources, internal monitoring cues such as processing fluency and sense of familiarity, and three judgment patterns involving intuitive reliance, consistent intuitive–rational alignment, and intuitive–rational conflict. A theory-driven hierarchical regression on an independent criterion-validity sample (N = 687) showed that the final 14-factor model explained 44.0
Learning independently involves not only cognitive but also emotional, motivational, and metacognitive challenges, particularly for children still developing their self-regulated learning skills. Consequently, children’s learning experiences can fluctuate rapidly within one learning activity, and assessing these experiences without disrupting learning remains an important research aim. In the present study, the feasibility of unobtrusively collecting two forms of trace data—pen pressure and facial expressions—from young learners in authentic classroom contexts was investigated, and the validity of the resulting indicators was tested. Data was analysed from two classroom studies with over 580 third- and fourth-grade children in Germany, who learned mathematics content in a tablet-based digital learning environment using digital pens and integrated webcams. Pen pressure was recorded in three phases of the learning task, facial expressions following feedback were coded for valence, and linear mixed-effects models tested how these indicators varied with item accuracy and difficulty. Mean and variance in pen pressure increased with item difficulty and after positive feedback, suggesting pen pressure indicators reflect both cognitive load and arousal. More varied pen pressure on difficult items that were answered incorrectly may reflect the metacognitive experience of uncertainty. Children’s facial expressions were rated more positively after positive feedback than error feedback and when they self-reported higher valence, while nuanced effects of item difficulty pointed to attributional processes shaping children’s emotional reactions. Overall, the findings demonstrate that pen pressure and facial expressions can be collected unobtrusively in real classroom contexts and provide valuable insights into young learners’ experiences.
The mere act of providing a judgment of learning (JOL) influences subsequent memory performance. Such JOL reactivity depends on features of the learned material, as shown by numerous studies demonstrating that providing JOLs tends to enhance memory for related word pairs but not unrelated word pairs. Consequently, the mechanisms underlying existing accounts of positive reactivity predict the effect should generalize across semantically related word pairs broadly. In contrast, we propose that providing JOLs prompts elaboration, but JOL reactivity is conditional on the amount of test-relevant elaboration that cues afford. That is, positive reactivity for related word pairs is not due to their semantic relatedness per se, but rather, occurs because semantically related pairs tend to afford deeper elaboration. Across three preregistered experiments (N = 375), we test these competing mechanistic assumptions by examining whether positive reactivity arises for synonyms–a word pair type that is characterized by high semantic relatedness but only affords shallow elaboration on that relation due to high similarity. Despite being highly related, we find no JOL reactivity for synonym pairs, significantly differing from the positive reactivity observed for distinct-related word pairs. That is, JOL reactivity does not uniformly generalize across all semantically related word pairs. Further, supporting an afforded elaboration mechanism, we find that a novel, continuous metric of the amount of elaboration word pairs afford moderates JOL reactivity. These results emphasize the need for refinement of major accounts of JOL reactivity and offer a pathway for transferring positive JOL reactivity to real-world learning.
Two experiments investigated the effects of task instructions and environmental setting on undergraduate students’ multiple document reading strategies and their use of information in written messages. Participants studied a set of four digital documents under a moderate time constraint in order to write a message either for an academic or a personal purpose. Half of the participants completed the task in a university room, whereas the other half completed the task from home. Working from home for a personal purpose decreased time on task and participants’ attention to the most informative documents. Students working for a personal purpose were more likely to refer to themselves and to articulate an opinion on the topic of interest in their messages, whereas those working for an academic purpose made more references to the information sources. Experiment 2 replicated most of the findings with a modified procedure in which participants wrote their messages from memory. In addition, Experiment 2 found that students working for an academic purpose and/or at the university had a more accurate memory for who said what in the documents. The results support a purposeful approach of reading whereby self-regulation is governed by readers’ understanding of explicit and implicit task demands.
Recent research shows that asking learners to make immediate judgments of learning (JOLs) during learning can improve their later memory performance for word pairs, words, and images. Whereas this so-called positive JOL reactivity effect has sparked much discussion on its educational relevance, empirical tests of the generalizability of immediate JOL reactivity to educationally relevant materials remain rare. The experiments reported in this study examined whether making immediate JOLs improves test performance across three exemplary types of educationally relevant materials. In three experiments, participants studied key-term definitions (Experiment 1), pairs of country outlines and country names (Experiment 2), or pairs of animal images and species names (Experiment 3), made immediate JOLs during studying or did not make JOLs, and were tested on their memory in short-answer tests (Experiments 1 to 3) or multiple-choice tests (Experiment 1). Results revealed that making immediate JOLs did not improve test performance. On the contrary, Experiment 2 and an integrative data analysis indicated that making JOLs reduced test performance when difficulty was high. These results provide no evidence that beneficial effects of immediate JOLs generalize to the learning of educationally relevant materials, highlighting the need for further research to understand why positive JOL reactivity does not emerge in this context.
While knowledge construction (KC) and group-level regulation are known to be critical for successful collaborative problem solving (CPS), the dynamic interplay between them remains underexplored. To capture the complex KC-regulation interplay, this study examined the sequential and temporal patterns of KC and group-level regulation during a four-week CPS task. Video data from 24 undergraduate students in six teams were coded and analyzed using process mining to visualize and compare interaction flows across high-, mid-, and low-performing groups. Results showed that KC was frequent across all groups, but performance differences were explained by how regulation was temporally positioned within epistemic activity. High-performing groups selectively embedded regulation within epistemic sequences. The groups used socially shared task understanding as a bridge back into sustained knowledge construction and maintained reciprocal cycles that returned to elaborative KC. In contrast, mid- and low-performing groups showed truncated or locally confined regulatory loops that were weakly reintegrated into KC. Performance gaps were most pronounced during the solution ideation phase, indicating that regulatory effectiveness is phase-sensitive in extended CPS. These findings contribute a process-oriented account of KC–regulation interplay and inform the design of phase-sensitive supports for collaborative learning.
Although metacognitive scaffolding is recognized as an effective strategy to enhance collaborative programming learning, its effectiveness may vary with task complexity. This quasi-experimental study examined the differential effects of metacognitive scaffolding across programming tasks of low, medium, and high complexity, involving 195 seventh-grade students from a Taiwanese junior high school (experimental group: 97; control group: 98). A 3 (task complexity: low, medium, high) × 2 (metacognitive scaffolding: with vs. without) mixed-factorial design was employed. The results revealed a significant interaction between task complexity and metacognitive scaffolding on students’ computational thinking tendency and learning motivation, with the scaffolding effects becoming more pronounced as task complexity increased. No significant interaction effects were found for metacognitive awareness or programming learning achievement. These findings suggest that teachers should adjust metacognitive scaffolding intensity according to task complexity and leverage challenging tasks to promote students’ higher-order thinking in collaborative programming.
Self-regulation develops rapidly in early childhood and predicts school readiness, school adjustment, and academic achievement. However, children from low-SES families struggle with self-regulation, putting them at risk for school success. While research indicates that parental behaviors significantly influence children’s self-regulation, most interventions are classroom-based. The study aims to examine the effects of the parent-focused intervention on parental self-regulation, parents’ co-regulation, parental stimulation, and children’s self-regulation in Turkish low-SES families. Using a quasi-experimental pretest-posttest control-group design, 66 mothers and their 4- to 6-year-old children from low-SES families were assigned to two groups: an intervention and a control group. The intervention developed by the researchers includes six sessions organized into three modules. The assessment instruments included the “Me as a Parent Scale,” “Home Learning Index,” and “Emotion Regulation Checklist,” all of which were completed by the mothers. The researchers also administered the “The Head-Toes-Knees-Shoulder Revised (HTKS-R)” to the children and the “Parent-Child Challenge Task” to the mother-child dyads. Following the intervention, mothers demonstrated significant improvements in parental self-regulation, parental stimulation, and co-regulation. Children’s behavioral regulation also improved, and those outcomes were observed at the six-month follow-up within the intervention group, while there were no significant differences in children’s emotion regulation between the groups. Overall, the findings demonstrated that a parent-based intervention can enhance behavioral regulation in children from low-SES families. The study highlights the importance of early childhood self-regulation for future development and suggests promoting parent-focused programs, particularly in non-Western settings, to enhance children’s self-regulation.
Knowledge exploration refers to the process of seeking information to deepen one's understanding of new topics or concepts. It involves engaging with unfamiliar and challenging stimuli, such as new objects, events, or ideas, with the goal of integrating new experiences and expanding one's knowledge base. However, the mechanisms behind exploration remain underexplored, particularly in academic settings. This study utilized an experimental framework involving 104 high school students, who answered 15 multiple-choice questions of varying difficulty and subsequently reported their epistemic emotions. They were then given the opportunity to explore the answers. The findings indicate that difficult tasks significantly increased knowledge exploration. Moreover, the study identified that accuracy feedback significantly moderates this relationship. The within-person path analysis revealed that epistemic emotions, particularly curiosity and confusion, were found to mediate the impact of task difficulty on knowledge exploration. However, boredom did not emerge as a significant mediator. These results highlight the importance of task design in educational settings, suggesting that educators and curriculum designers can strategically leverage task difficulty to promote active engagement and in-depth understanding among students. By incorporating challenging tasks into the curriculum, they can stimulate students' curiosity and motivation to explore the knowledge domain.
This study examines how the order of rubric performance levels (high-to-low vs. low-to-high) and four types of feedback (no feedback, rubric-based, product, and process) influence university students’ self-regulated learning (SRL) and self-efficacy during an academic task. A total of 120 first-year students completed two landscape analysis tasks supported by rubrics while verbalizing their thoughts through think-aloud protocols. Self-efficacy was measured before and after the experimental session. Students predominantly produced positive self-regulatory actions across the five task phases, displaying a W-shaped pattern that reflected shifts in task demands: higher regulatory engagement during rubric reading and written analysis, and lower engagement during oral description phases. Negative actions were less frequent and concentrated in the initial rubric reading and first written analysis. Neither performance-level order nor feedback type significantly affected positive or negative SRL actions; the only significant within-subjects factor was task phase. Self-efficacy decreased across all conditions, suggesting a recalibration of confidence in response to task demands and rubric criteria. These findings indicate that moment-to-moment SRL engagement is primarily shaped by the nature of the task phase rather than by the specific scaffolding configuration, and underscore the importance of process-oriented methods for capturing dynamic regulatory patterns in rubric-supported learning environments.
Metacognitive regulation, which encompasses both monitoring and control processes, is thought to be of great importance to learning and academic success. Despite its importance, most research focuses on metacognitive monitoring, excluding metacognitive control. Moreover, its developmental trajectory, particularly in the context of arithmetic, remains underexplored. This study examined the development of metacognitive regulation in arithmetic. We performed a cross-sectional investigation in 7-8-year-olds (n = 114), 10-11-year-olds (n = 102), adolescents (n = 99), and adults (n = 102). Participants completed two arithmetic tasks. In one task we measured post-error slowing (PES) as an indicator of metacognitive control. In the other task, we used retrospective confidence judgments as an indicator of metacognitive monitoring. Results demonstrate that people slow down after committing an error across all age groups, with the magnitude of PES decreasing with age. Although the ability to monitor one’s performance initially appeared to improve with age, this effect disappeared when task performance was controlled for. Notably, monitoring correlated with arithmetic performance across all age groups, while PES showed only associations with task performance and metacognitive monitoring in 7-8-year-olds and not in older age groups. These results indicate that although individuals slow down after making errors, the limited effectiveness of this strategy in improving performance appears to diminish with age. People might apply other, perhaps proactive, control strategies, as they get older. These results deepen our understanding of the development of metacognitive regulation in relation to academic performance and can inform strategies to effectively support children’s learning and academic success.
Cognitive aging involves both progressive decline and dynamic adaptive reorganization in lexical processing. While older adults benefit from early language experience and accumulated lexical knowledge, the metacognitive mechanisms supporting this adaptation remained unclear. Through a lexical decision task and an eye-tracked sentence reading task, we examined how metacognitive monitoring and control operate in lexical processing between the two age groups. Following the theoretical framework of metacognition, early eye-movement indices were treated as markers of online monitoring, whereas later and global measures indexed strategic control. Compared to young adults (18–26 years), older adults (64–74 years) showed reduced overall processing efficiency but heightened sensitivity to the age of acquisition, indicating robust monitoring of processing difficulty for late-acquired words. In sentence reading, older adults strategically leveraged contextual predictability, demonstrating reduced total fixation duration and regressions under high-predictability conditions. In contrast, young adults demonstrated a selective reliance on context for more demanding lexical items. These findings suggest that aging is characterized not only by decline but also by adaptive metacognitive regulation, whereby older adults flexibly recruit experiential knowledge and contextual cues to compensate for their processing constraints.
Students skilled in self-regulated learning (SRL) can plan, monitor, and reflect on their learning, using strategy knowledge (SK) to select and apply strategies. Teachers often struggle to assess SRL and rely on non-diagnostic cues, particularly when their knowledge of SRL is limited. Biased evaluations could affect how teachers support students and impact their performance. This study investigated teachers’ accuracy in judging students’ SK across five SRL components (cognition, metacognition, time management, management of the learning environment, and motivation regulation) by comparing teachers’ SK judgments with students’ SK test scores. Specifically, we explored judgment biases related to students’ academic achievement, gender, and socioeconomic status. Further, teacher characteristics that may contribute to biased judgments were examined (knowledge of SRL, self-concept, experience, and gender stereotypes). The sample included 200 students and 69 teachers. Teachers’ relative judgment accuracy was assessed using correlations with cluster-robust standard errors. Multilevel models were used to analyze how student characteristics affected teachers’ SK judgments. Cross-level interactions were employed to investigate the impact of teacher variables on judgment biases. Results indicated low accuracy and component-specificity of teachers’ judgments. Furthermore, their judgments were affected by achievement and gender biases. These may be attributed to a halo effect, in which teachers’ overall impression of a student influenced their SK judgments. Teachers’ knowledge of SRL, self-concept, and experience were not associated with reduced biases, but were (trend-wise) associated with stronger biases, albeit not consistently across components. These findings question teachers’ ability to distinguish between students’ SRL skills and their overall impressions of students.
Digital reading has been widely linked to more superficial processing and higher frequency of multitasking than print reading. These aspects have been extensively examined in single-text research, indicating detrimental effects on reading comprehension and metacomprehension, particularly when reading on digital and when engaging in multitasking. Nevertheless, much less is known about how these factors interact and influence readers in multiple-text contexts. Given that today’s readers are regularly exposed to multiple, and often conflicting, sources of information, it is crucial to identify factors that may obstruct accurate self-monitoring and understanding. This study examined how reading media (print vs. tablet) and multitasking while reading affected intertextual comprehension and metacomprehension accuracy among 151 Spanish secondary school students (from grade 8 to 10). Our findings show that multitasking hindered intertextual comprehension more in the print condition, whereas tablet readers demonstrated no notable decline. Unexpectedly, neither reading media nor multitasking significantly influenced metacomprehension scores, pointing to a general difficulty among students in accurately assessing their own understanding. These results suggest that students’ expectations about different reading media may shape how they experience and manage distractions. Furthermore, this study shows that metacognitive overconfidence persists in multiple-text scenarios, as students tend to overlook conflicting information without being aware.
