Self-Monitoring Skills in Online Learning and the Role of Metacognitive Prompts: An Exploratory Study

Koval, A.A. HSE University, Moscow, Russia

Gorbunova, A.Y. HSE University, Moscow, Russia

Podolskiy, A.I. HSE University, Moscow, Russia

Getman, A.V. HSE University, Moscow, Russia

Introduction

Online learning is increasingly prevalent in modern education, including higher education (Pikhart et al., 2023), yet it presents unique challenges for learners (Pham, 2022). Unlike traditional face-to-face instruction, online learning is characterized by spatial, temporal, and material flexibility (Barak et al., 2016; Basar et al., 2021). Students can choose where and when to study, access diverse multimedia resources, and review the materials multiple times (Barak et al., 2016; Basar et al., 2021). However, this flexibility places greater demands on learners, as the learning process becomes less structured (Sistek-Chandler, 2019). Online learning also involves diminished physical presence of instructors and peers, leading to reduced real-time feedback, limited communication, and lower levels of engagement (Akpen et al., 2024; Pikhart et al., 2023; Singh et al., 2024). From the perspective of Galperin’s theory of the stage-by-stage formation of mental actions (Galperin, 1969), these challenges reflect the reduced availability of external support. In traditional classrooms, instructors naturally provide such support in the form of orienting bases—a system of cues and conditions that guide students and help them avoid inefficient trial-and-error strategies (Galperin, 1969). This orienting basis scaffolds students’ mental actions and the learning process as a whole (Podolsky, 2023). In online environments, where these supports are largely absent, students must rely more heavily on their own resources to monitor their understanding and regulate their learning. This is why students are increasingly required to engage in self-regulation (de Bruin et al., 2020).

Self-regulation is a goal-directed process in which learners actively plan, monitor, and adapt their cognition, motivation, and behavior during learning (Pintrich, 2000; Winne & Hadwin, 1998; Zimmerman, 2000). In Zimmerman’s model of self-regulated learning (SRL), this process is conceptualized as a cyclical sequence of forethought, performance, and self-reflection phases (Zimmerman, 2000). In the forethought phase, learners engage in goal setting, strategic planning, and motivational processes (e.g., self-efficacy, task value). During this phase, students establish standards against which subsequent performance can be evaluated. In the performance phase, learners actively engage in the task while employing self-monitoring and self-control. Central to this phase is self-monitoring, a metacognitive process through which learners track their understanding, progress, and task performance in real time. Self-monitoring generates ongoing feedback that informs self-control processes, enabling learners to adjust strategies, allocate effort, or redirect attention as needed (Zimmerman, 2000). Finally, in the self-reflection phase, learners evaluate their performance outcomes through self-judgments (e.g., self-evaluation, causal attribution) and self-reactions (e.g., satisfaction, adaptive/defensive responses). These reflections feed forward into subsequent forethought processes. Thus, self-monitoring plays a central, integrative role in SRL by linking performance to control processes and connecting current learning episodes with future regulation cycles.

Self-monitoring is considered cognitively demanding, as students have to learn while simultaneously monitoring their own understanding (de Bruin et al., 2020; Seufert, 2018; Zimmerman & Moylan, 2009). This challenge is amplified in online learning environments by reduced feedback, absence of peer and instructor support, and screen-based presentation, all of which impair students’ ability to accurately monitor their understanding (Ackerman & Lauterman, 2012; Mou, 2023; Viberg et al., 2020; Wong et al., 2019).

Cognitive Load Theory (CLT; Sweller et al., 1998) provides a useful framework for understanding why self-monitoring is cognitively demanding. At its core, CLT assumes a fundamental distinction between limited working memory and virtually unlimited long-term memory. Working memory enables the conscious processing of new information but is severely constrained in both capacity and duration, allowing only a small number of elements to be processed simultaneously. In contrast, long-term memory stores organized knowledge structures, or schemas, which can be retrieved and processed as single units, thereby reducing cognitive demands during learning (Sweller et al., 2019). Learning, from a CLT perspective, involves the construction and automation of these schemas, enabling more efficient processing over time.

Within this framework, cognitive load arises from the demands placed on working memory during learning. CLT distinguishes three types of cognitive load: intrinsic load (ICL), which depends on the complexity of the material and learners’ prior knowledge; extraneous load (ECL), which results from instructional design; and germane load (GCL), which reflects cognitive resources devoted to meaningful processing and schema construction (Sweller et al., 1998; Sweller et al., 2019).

Initially, GCL was conceptualized as an additional source of cognitive demand (Sweller et al., 1998). However, more recent work reconceptualizes GCL not as an additional component of total cognitive load, but as a mechanism for reallocating cognitive resources from unproductive to productive processing (Sweller et al., 2019). Despite this theoretical discussion, broad consensus has been reached that germane processing—the productive investment of working-memory resources in meaningful learning activities—is essential for learning (Kalyuga & Singh, 2015; Greenberg & Zheng, 2023). What is referred to as GCL in the present study is therefore best understood as an indicator of this productive cognitive effort, reflecting the extent to which learners engage in schema construction and meaningful processing. This operationalization is aligned with a substantial body of empirical research in which GCL has been successfully measured as a separable construct and its sensitivity to instructional manipulations has been demonstrated (Beege & Ploetzner, 2025; Costley et al., 2025; Gorbunova et al., 2026; Krieglstein et al., 2023). Treating GCL as a distinct type of load allows for a better understanding of how instructional interventions work, as GCL represents learning-relevant processing (Merkt & Bodemar, 2024).

From a CLT perspective, self-monitoring can be understood as an additional layer of cognitive processing that operates alongside primary learning activities. Prior research shows that self-monitoring may affect different types of cognitive load in distinct ways (Seufert, 2018). Self-monitoring may increase ECL by introducing additional regulatory processing demands, exert limited influence on ICL because task complexity remains unchanged, and enhance GCL by supporting learning-relevant cognitive activities (Seufert, 2018). 

Students with higher self-monitoring skills are better at managing cognitive load and demonstrate more effective self-monitoring during learning (Glogger-Frey et al., 2016; Schunk, 1982; Seufert, 2018; Sidi & Ackerman, 2024; Zimmerman & Schunk, 2011). Specifically, students with high self-monitoring skills tend to acquire, organize and use information more efficiently, enabling them to regulate their study time effectively, allocate cognitive resources more strategically and persist in the face of difficulty (Shi & Qu, 2022). In contrast, students with low self-monitoring skills show weaker regulatory control, resulting in poorly planned and less purposeful cognitive activity, even when their cognitive abilities are relatively high (Shi & Qu, 2022). From a resource perspective, these differences reflect that higher self-monitoring skills provide learners with greater metacognitive resources for managing cognitive demands and optimizing learning outcomes (Karpov et al., 2017). It is also worth noting that according to CLT, fostering self-monitoring skills through instructional interventions may be inherently limited (Sweller et al., 2019). However, empirical research suggests that self-monitoring skills can be developed through extended instructional interventions, typically requiring several weeks or months of sustained practice (e.g., Dignath et al., 2023; Huff & Nietfeld, 2009; Zimmerman & Schunk, 2011).

Beyond cognitive load, self-monitoring skills are also consistently associated with positive learning outcomes (e.g., Jung et al., 2024; Schraw et al., 2006; Zimmerman & Schunk, 2011). However, the strength of their relationship with academic performance is less consistent, with meta-analytic evidence suggesting effects that range from weak to moderate (Credé & Phillips, 2011). This variability likely reflects multiple contextual and methodological influences, indicating that further investigation could help refine our understanding of these relationships. 

To address the challenges students face with self-monitoring, instructional tools can be designed to activate self-monitoring during learning (Zimmerman, 2000). In online learning  environments, metacognitive prompts are widely used (Guo, 2022). They are specifically designed to support self-monitoring by directing students’ attention to their understanding, progress, or strategy use (Bannert & Mengelkamp, 2013; Berthold et al., 2007; Müller & Seufert, 2018). They function by signaling when and how students should monitor themselves, thereby activating monitoring processes that might otherwise not be initiated spontaneously (Berthold et al., 2007; Müller & Seufert, 2018). Compared to more elaborate self-monitoring tools, metacognitive prompts are considered simple tools because they require minimal effort and do not necessitate real-time instructor involvement (Engelmann et al., 2021).

Prior research shows that self-monitoring support tools, including metacognitive prompts, improve students’ self-monitoring skills (Dignath et al., 2023) and, when gradually withdrawn, function as metacognitive scaffolding (Azevedo & Hadwin, 2005). However, the interaction between students’ existing self-monitoring skills and metacognitive prompts remains underexplored. From a theoretical perspective, students with stronger self-monitoring skills would be expected to use prompts more efficiently (Nelson & Narens, 1990; Zimmerman, 2000), as their higher skills enable them to recognize the relevance of prompts and translate monitoring information into effective regulatory actions. Moreover, the academic benefits of any self-monitoring support depend not only on prompts but on students’ capacity for reflective processing and strategic adjustment (Guo, 2022). These considerations suggest that metacognitive prompts may be particularly beneficial for students with stronger self-monitoring skills, facilitating more efficient management of cognitive load and improved academic performance.

On the other hand, an alternative theoretical framework suggests a different prediction. According to Galperin’s theory of the stage-by-stage formation of mental actions (Galperin, 1969), cognitive processes, including self-monitoring, develop through the internalization of external, materialized actions. From this perspective, metacognitive prompts can be understood as a form of external scaffolding that provides learners with an orienting basis for action. As learners repeatedly engage with such prompts, the externally guided monitoring processes may gradually become internalized, transforming into autonomous self-monitoring skills (Talyzina, 1981). This view implies that prompts could be particularly beneficial for students with weaker self-monitoring skills, as they offer structured support that compensates for underdeveloped internal resources. The resource approach to metacognition (Karpov et al., 2017) further extends this idea. According to this view, learners with lower self-monitoring skills possess fewer internal metacognitive resources, making them more dependent on external support. Metacognitive prompts may therefore serve as a substitutional resource—temporarily compensating for deficient internal resources—thereby reducing performance gaps between high- and low-skilled students. Thus, contrary to the amplification hypothesis, these theoretical perspectives predict that metacognitive prompts may function as a compensatory mechanism, benefiting weaker students more than stronger ones.

Overall, despite the theoretical and practical importance of self-monitoring skills in online learning, several gaps remain. First, it is unclear how self-monitoring skills relate to different types of cognitive load and academic performance in online learning environments. Second, while metacognitive prompts are widely used to support self-monitoring in online learning, it remains unknown whether such prompts amplify existing skill differences (benefiting stronger students) or compensate for them (benefiting weaker students). These competing predictions can be operationalized as distinct interaction patterns. The amplification hypothesis would be supported by a positive interaction between self-monitoring skills and prompt condition, indicating that prompts strengthen the relationship between self-monitoring skills and outcomes, thereby benefiting higher-skilled students more. Conversely, the compensatory hypothesis would be supported by a negative interaction, indicating that prompts weaken the relationship between self-monitoring skills and outcomes, thereby reducing the advantage of higher-skilled students and benefiting lower-skilled students more. To address these gaps, the present study investigated two research questions.

RQ1: How are students’ self-monitoring skills related to different types of cognitive load and academic performance in online learning?

RQ2: How do metacognitive prompts moderate the relationship between students’ self-monitoring skills, cognitive load, and academic performance? 

Methods

Participants and Procedure

The sample comprised 40 Russian undergraduate students (77.5% female, age range 19-23 years, median = 20) enrolled in an online minor course. The students represented diverse academic backgrounds across the humanities, social sciences, and technical fields.

Participants completed a 40-minute asynchronous online self-study task. They studied a text on questionnaire theory (Rogozin & Ipatova, 2020) and prepared for a subsequent test. Each participant received an individual copy of the text in Google Docs and was permitted to annotate the material. Participants were randomly assigned to either a prompt condition (n = 20) or a control condition (n = 20). Participants in the prompt condition received metacognitive prompts embedded after each of the six substantive sections of the text. Adapted from Berthold et al. (2007), these prompts encouraged learners to evaluate their understanding and connect the content to prior knowledge by asking three questions: (1) “Which two or three key ideas from this text do I understand best?”; (2) “Overall, how confident am I in my understanding of this text? (0–100%)”; and (3) “To what extent was this content already familiar to me? (very familiar / partly familiar / not familiar)”, followed by a brief prompt to reflect on prior knowledge: “If I already knew something about this topic, what prior knowledge or experiences did I draw on?”. These prompts were designed to support self-monitoring during asynchronous online study.

After doing the learning task, students had 20-25 minutes to complete a questionnaire and a test. Following that, the participants took part in a debriefing session. The study was approved by the institutional ethics committee, and the informed consent was obtained from all participants. 

Measures

Different types of cognitive load (ICL, ECL, GCL) were measured using a validated 15-item questionnaire developed by Krieglstein et al. (2023). Responses were recorded on a 9-point Likert scale ranging from 1 (strongly disagree) to 9 (strongly agree). Cronbach’s alpha coefficients indicated good internal consistency for ICL (α = .74), ECL (α = .88), and GCL (α = .86). Cognitive load was measured immediately after the learning task and before the test to capture the cognitive demands of the learning activity rather than the assessment.

Self-monitoring skills were measured using items from the Comprehension Monitoring subscale of the Metacognitive Awareness Inventory (MAI; Schraw & Dennison, 1994), in the validated Russian adaptation (Perikova & Byzova, 2022). Responses were recorded on a 5-point Likert scale ranging from 1 (totally disagree) to 5 (absolutely agree). Cronbach’s alpha also indicated good reliability (α = .71). Self-monitoring skills were measured after the learning task and the test. This construct was treated as a stable individual characteristic that is unlikely to change as a result of a short experimental intervention. This timing is consistent with recommendations in SRL research to prevent SRL measurement tools from inadvertently promoting any self-regulation, including self-monitoring (Dörrenbächer-Ulrich et al., 2021). Administering the questionnaire prior to the learning task was avoided to prevent artificially heightened self-monitoring in the control group.

Academic performance was measured using a 10-item multiple-choice comprehension test, with each question having one correct answer. It assessed students’ acquisition and retention of key concepts from the learning material, thereby capturing their immediate learning outcome. When carefully constructed, multiple-choice tests provide a reliable and valid basis for inferences about academic performance (Brown & Abdulnabi, 2017). 

Analytical Strategy

The analysis was done in R. Linear regression analyses were conducted to examine the relationships between students’ self-monitoring skills and cognitive load, as well as academic performance (RQ1). To test whether metacognitive prompts have a potential to moderate the relationships between self-monitoring skills and the outcome variables, multiple linear regression models including interaction terms were estimated (RQ2). Prompt condition was coded as a dichotomous variable, with 0 representing the control condition (no prompts) and 1 representing the prompt condition. This coding was used in all regression analyses for RQ2 to examine the main effect of prompts and their interaction with self-monitoring skills. The interaction term served as a test of whether prompts amplify or compensate for individual differences in self-monitoring skills. To control for multiple comparisons, False Discovery Rate (FDR) correction with a threshold of q < .05 was used (Benjamini & Hochberg, 1995). This approach was chosen over more conservative corrections (e.g., Bonferroni) to balance Type I error control with statistical power (Nakagawa, 2004).

Prior to regression analyses, Spearman’s correlation was used to check for multicollinearity among predictors and to describe the relationships among all study variables. Spearman’s correlation was selected because all variables were measured on ordinal Likert-type scales, which do not meet the interval-level assumptions required for Pearson correlation. No correlations exceeded the recommended threshold of r > .90 (Tabachnick & Fidell, 2006), indicating no multicollinearity concerns. Other regression assumptions were also evaluated, and no other substantial violations were detected. Although the sample size was modest, it was adequate for the estimated models given the limited number of predictors, and prior evidence that linear regression yields reliable coefficient estimates in small samples when assumptions are met (Austin & Steyerberg, 2015; Harrell, 2015).

Results

Descriptive statistics (see Table 1) and correlation analysis (see Table 2) for all variables are presented. 

Table 1

Descriptive Statistics 

	M	Median	SD	Min	Max	Skewness	Kurtosis
Self-monitoring skills	3.69	3.75	.65	2.33	5.0	- .29	- .45
ICL	3.36	3.20	.91	1.80	5.6	.43	- .49
ECL	3.70	3.80	1.39	1.20	6.2	.01	- 1.00
GCL	4.92	5.20	.99	1.40	6.4	- 1.78	3.47
Academic performance	7.20	7.00	1.83	4.00	10.0	.08	- 1.01

Table 2 

Correlation Analysis (Spearman)

	Self-monitoring skills	ICL	ECL	GCL	Academic performance
Self-monitoring skills	–
ICL	-.445**	–
ECL	-.263*	.464**	–
GCL	.238	-.336*	-.386*	–
Academic performance	.342*	-.281	-.158	-.289	–

Notes. *p < .05, **p < .01

RQ1. How are students’ self-monitoring skills related to different types of cognitive load and academic performance in online learning?

After applying FDR correction for multiple comparisons (see Table 3), self-monitoring skills showed marginal associations with ICL (b = -.48, adjusted p = .054), ECL (b = -.72, adjusted p = .054), and academic performance (b = .92, adjusted p = .054). Self-monitoring skills were unrelated to GCL (b = .09, adjusted p = .703).

Table 3

Linear Regression Models Results for the Relationships between Self-Monitoring Skills, Cognitive Load (ICL, ECL, GCL), and Academic Performance

	ICL	ECL	GCL	Academic performance
Intercept	5.12  (<.001***)	6.37  (<.001***)	4.56  (<.001***)	3.81 (.025*)
Self-monitoring skills	-.48 (.033*)	-.72(.033*)	.09(.703)	.92(.041*)
FDR-adjusted p-value for self-monitoring skills	.054†	.054†	.703	.054†
Adjusted R2	.09	.09	.003	.08
Partial R2	.11	.11	.004	.11

Notes. † p < .10 (trend), *p < .05, ***p < .001; values in parentheses are original p-values

RQ2: How do metacognitive prompts moderate the relationship between students’ self-monitoring skills, cognitive load, and academic performance? 

After applying FDR correction for multiple comparisons (see Table 4), significant interactions were found between self-monitoring skills and metacognitive prompt use for ICL (b = .99, adjusted p = .024), ECL (b = 1.65, adjusted p = .016), and academic performance (b = -2.34, adjusted p = .016). The overall model for GCL was not significant (p = .103). Given this, and consistent with recommendations for interpreting interaction effects (Aiken & West, 1991), we do not interpret the GCL interaction term and exclude it from further discussion.

To assess the robustness of the interaction effects for ICL, ECL, and academic performance, bias-corrected and accelerated (BCa) bootstrap confidence intervals based on 5,000 resamples were computed (Efron & Tibshirani, 1994). The 95% BCa bootstrap confidence intervals supported all three interaction effects: ICL (CI [.163, 1.878]), ECL (CI [.828, 2.765]), and academic performance (CI [-3.867, -.965]). As none of the confidence intervals included zero, the interaction effects appeared robust to resampling.

A significant positive main effect of prompt condition on ECL was also found (b = .99, p < .001). This main effect should not be interpreted in isolation, as it reflects the average across all participants and is qualified by the significant interaction, which shows that prompts were beneficial primarily for students with lower self-monitoring skills.

Table 4

Multiple Linear Regression Models Results for the Interactions Between Self-Monitoring Skills and Metacognitive Prompt Use Predicting Cognitive Load and Academic Performance

	ICL	ECL	GCL	Academic performance
Intercept	3.16 (<.001***)	3.16 (<.001***)	5.09  (<.001***)	7.42 (<.001***)
Self-monitoring skills	-1.11(.002 **)	-1.8(<.001***)	.82(.04 *)	2.39(<.001***)
Metacognitive prompt use (prompt condition)	.34(.19)	.99(< .001**)	-.29(.33)	-.32(.53)
Interaction term	.99(.024*)	1.65(.008**)	-1.15(.024*)	-2.34(.008**)
FDR-adjusted p-value for the interaction term	.024*	.016*	.024*	.016*
BCa CI for interaction term	[.163, 1.878]	[.828, 2.765]	–	[-3.867, -.965]
Model F	4.24(.012*)	7.32(< .001**)	2.22(.103)	4.4(.009**)
Adjusted R2	.19	.33	.09	.21
Partial R2	.26	.38	.16	.27

Notes. *p < .05, **p < .01, ***p < .001; interaction terms represent the Self-Monitoring skills × Metacognitive prompt use effect; values in parentheses are original p-values

Simple slope analyses (see Table 5) indicated that in the control group, higher self-monitoring skills were associated with lower ICL (b = -1.11, 95% CI [-1.78, -.44]), lower ECL (b = -1.81, 95% CI [-2.74, -.88]), and higher academic performance (b = 2.39, 95% CI [1.06, 3.72]). In contrast, in the prompt group, these relationships were no longer statistically significant for ICL (b = -.12, 95% CI [-.64, .40]), ECL (b = -.16, 95% CI [-.89, .57]), or academic performance (b = .05, 95% CI [-.99, 1.09]). Consistent with the significant interaction effects, these results indicate that metacognitive prompts weakened the association between students’ self-monitoring skills and these learning outcomes, reducing the extent to which performance and cognitive load depended on students’ initial self-monitoring skills. 

Table 5

Simple Slopes for the Relationship between Self-Monitoring Skills and Learning Outcomes by Condition

	Group	b	SE	95% CI
ICL	Control	-1.11	.33	[-1.78, -.44]
	Prompt	-.12	.26	[-.64, .40]
ECL	Control	-1.81	.46	[-2.74, -.88]
	Prompt	-.16	.36	[-.89, .57]
Academic performance	Control	2.39	.66	[1.06, 3.72]
	Prompt	.05	.51	[-.99, 1.09]

Discussion

The present study investigated how self-monitoring skills are related to different types of cognitive load and academic performance in online learning environments (RQ1), and whether metacognitive prompts, the most widely used self-monitoring support in online settings, moderate these relationships (RQ2).

For RQ1, after FDR correction, higher self-monitoring skills showed marginal associations with lower ECL and higher academic performance, consistent with prior work linking self-monitoring to more effective self-regulated learning (Schraw et al., 2006; Zimmerman, 2000; Zimmerman & Schunk, 2011). This pattern is theoretically consistent with the assumption that better self-monitoring helps students detect misunderstandings earlier and avoid inefficient strategies, freeing resources that would otherwise be spent on extraneous processing (Seufert, 2018). Two findings diverged from our expectations: self-monitoring skills were also marginally associated with lower ICL and were unrelated to GCL. The trend for a negative association with ICL may reflect differences in perceived task difficulty—higher-skilled learners may organize and “chunk” information more effectively, making the material feel less complex even when task demands are unchanged (Löffler et al., 2016). The null effect for GCL contrasts with prior findings suggesting that effective self-monitoring should increase GCL by promoting deeper engagement with the material (Dignath et al., 2023; Guo, 2022; Seufert, 2018). This discrepancy may be explained by a cost-benefit perspective on cognitive load in online learning, which states that in this environment students are more prone to invest additional cognitive effort in deeper processing only when the perceived benefits outweigh the associated cognitive costs (Skulmowski & Xu, 2021). In the present study, students with higher self-monitoring skills may have primarily used these skills to reduce ICL and ECL, while the learning context provided insufficient incentive or opportunity to invest additional cognitive effort in elaborative processing, leading to a null effect for GCL (Gorbunova et al., 2026).

For RQ2, metacognitive prompts were associated with moderated relationships between self-monitoring skills, ICL, ECL, and academic performance. Contrary to the assumption that prompts would strengthen these relationships, the observed pattern was compensatory. The associations between self-monitoring skills and learning outcomes were weaker when metacognitive prompts were used, potentially making successful learning less dependent on students’ initial self-monitoring skills. From a theoretical perspective, prompts may externalize the metacognitive process, thus reducing differences between high- and low-skilled students (Galperin, 1969; Karpov et al., 2017). Prompts may have functioned as external scaffolds that provided an orienting basis for action and filled internal resource deficits, benefiting weaker students more. At the same time, prompts may not have boosted the advantages of higher skills because they target basic self-monitoring, not the elaborative processing that more skilled learners might need. However, this interpretation is speculative, and it is also possible that prompts did not merely compensate for skill deficits but may have interfered with the learning processes of higher-skilled students. Specifically, the significant interaction suggests that for students with higher self-monitoring skills, the additional structure provided by prompts may have been redundant or even disruptive, potentially reducing their performance and increasing their perceived cognitive load. Thus, while prompts may support lower-skilled students, they may not be beneficial and could even be counterproductive for students who are already capable of effective self-monitoring. 

Regarding the GCL interaction, it should not be interpreted as evidence of a reliable effect. Although the interaction term was significant in isolation, the overall model was not significant. At the same time, prior research indicates that increases in GCL typically require more demanding forms of self-monitoring support, such as learning journals or generative tasks, than the brief comprehension checks used in this study (Guo, 2022). 

Conclusion

This exploratory study examined self-monitoring skills in online learning, their relationship with different types of cognitive load (ICL, ECL, GCL) and academic performance, and whether metacognitive prompts have the potential to moderate these relationships. After FDR correction, self-monitoring skills showed marginal associations with lower ICL and ECL and better performance, while the relationship with GCL in this study was insignificant. Metacognitive prompts appeared to act as a compensatory mechanism rather than an amplifier of existing skills: when prompts were used, students’ outcomes tended to depend less on their pre-existing self-monitoring skills, potentially reducing skill-based differences in cognitive load and performance. However, these findings are preliminary and require replication with larger samples.

The findings suggest several tentative implications for instructional design. First, given that self-monitoring skills showed marginal associations with lower cognitive load and better performance, fostering these skills through explicit instruction or practice opportunities may help students manage the cognitive demands of online learning more effectively. Second, because metacognitive prompts appear to reduce students’ reliance on their existing self-monitoring skills, they may serve as an equity-oriented support for students with weaker self-monitoring skills. Finally, when the instructional goal is to promote deeper processing and improve academic performance, metacognitive prompts should be combined with more complex self-monitoring support (e.g., learning journals or generative tasks) rather than used solely as brief comprehension checks.

Limitations

Several limitations should be considered when interpreting these findings. First, this study was exploratory in nature, and the results should be regarded as preliminary. The sample size was modest (N = 40), which limits statistical power. This may explain why the associations between self-monitoring skills and cognitive load and performance became marginal after FDR correction. According to Tabachnick and Fidell (2006), such a sample size provides sufficient power for detecting only large effects (R² ≥ .26), while adequate detection of medium effects (R² ≈ .13) typically requires larger samples. In linear regression models, the observed effects were not large (partial R² = .11), suggesting that marginal p-values after FDR correction reflect the limited sample size rather than the absence of meaningful relationships. The modest sample size also limited the stability of the interaction effects, with a minimally adequate cases-to-predictors ratio of 10:1 (Harrell, 2015). Bootstrap resampling (5,000 iterations) provided some reassurance; however, this does not substitute for adequate power. Replication with larger samples is needed to confirm the reliability of these findings.

In addition, the measurement of self-monitoring skills after the learning task may have introduced state-dependent influences, as scores could reflect not only stable individual differences, but also reactions to task difficulty and test performance. This timing, while consistent with common practice in SRL research to avoid reactivity effects (Dörrenbächer-Ulrich et al., 2021), nevertheless limits our ability to draw causal inferences. Furthermore, the assessment format, a short multiple-choice test, may have been insufficiently sensitive to differences in deeper understanding, potentially reducing observable academic performance effects even when learning processes differed. Future research should employ more diverse learning tasks and assessments that capture transfer or elaborative understanding. 

Further studies could also compare different types of prompts, including AI-generated adaptive prompts tailored to individual learners’ comprehension levels, and investigate how their use interacts with longer-term development of self-monitoring skills to better understand how compensatory support can transition into sustained self-regulated learning.

Ethics Statement

This study was conducted in accordance with the ethical standards of the Institutional Review Board of the National Research University Higher School of Economics (HSE IRB). The project titled “Supporting Self-Monitoring in Online Learning: Impact on Cognitive Load and Learning Outcomes” was reviewed and approved at the HSE IRB meeting held on October 23, 2025. The IRB confirmed that the research conforms fully to ethical guidelines and may be implemented in its current form.

All participants were informed about the purpose of the study, the procedures involved, and their right to withdraw at any time without penalty. Written informed consent was obtained from all participants prior to their involvement in the research. Confidentiality and anonymity of the data were ensured throughout the study.

Author Contributions

A.A. and A.Y. conceived of the idea. A.A. and A.V. collected the data. A.A. developed the theory and performed the computations. A.Y. and A.V. verified the analytical methods. A.Y. supervised the findings of this work. A.I. contributed to the conceptual explanation of the findings and assisted in the preparation of the manuscript’s final draft. All authors discussed the results and contributed to the final text of the manuscript. 

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgements

This work/article is an output of a research project (HSE-BR-2025-010) implemented as part of the Basic Research Program at HSE University.

References

Ackerman, R., & Lauterman, T. (2012). Taking reading comprehension exams on screen or on paper? A metacognitive analysis of learning texts under time pressure. Computers in Human Behavior, 28(5), 1816–1828. https://doi.org/10.1016/j.chb.2012.04.023

Aiken, L.S., & West, S.G. (1991). Multiple regression: Testing and interpreting interactions. Sage Publications.

Akpen, C.N., Asaolu, S., Atobatele, S., Okagbue, H., & Sampson, S. (2024). Impact of online learning on student’s performance and engagement: A systematic review. Discover Education, 3(1), 205. https://doi.org/10.1007/s44217-024-00253-0

Austin, P.C., & Steyerberg, E.W. (2015). The number of subjects per variable required in linear regression analyses. Journal of Clinical Epidemiology, 68(6), 627–636. https://doi.org/10.1016/j.jclinepi.2014.12.014

Azevedo, R., & Hadwin, A.F. (2005). Scaffolding self-regulated learning and metacognition: Implications for the design of computer-based learning environments. Instructional Science, 33(5–6), 367–379. https://doi.org/10.1007/s11251-005-1272-9

Bannert, M., & Mengelkamp, C. (2013). Scaffolding hypermedia learning through metacognitive prompts. In R. Azevedo & V. Aleven (Eds.), International handbook of metacognition and learning technologies (pp. 171–186). Springer. https://doi.org/10.1007/978-1-4419-5546-3_12

Barak, M., Hussein-Farraj, R., Dori, Y. (2016). On-campus or online: Examining self-regulation and cognitive transfer skills in different learning settings. International Journal of Educational Technology in Higher Education, 13, 35. https://doi.org/10.1186/s41239-016-0035-9

Basar, Z.M., Mansor, A.N., Jamaludin, K.A., & Alias, B.S. (2021). The effectiveness and challenges of online learning for secondary school students – a case study. Asian Journal of University Education, 17(3), 119. https://doi.org/10.24191/ajue.v17i3.14514

Beege, M., & Ploetzner, R. (2025). Learning from interactive video: The influence of self-explanations, navigation, and cognitive load. Instructional Science, 53(1), 99–119. https://doi.org/10.1007/s11251-024-09693-5

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B (Statistical Methodology), 57(1), 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x

Berthold, K., Nückles, M., & Renkl, A. (2007). Do learning protocols support learning strategies and outcomes? The role of cognitive and metacognitive prompts. Learning and Instruction, 17(5), 564–577. https://doi.org/10.1016/j.learninstruc.2007.09.007

Brown, G.T.L., & Abdulnabi, H.H.A. (2017). Evaluating the quality of higher education instructor-constructed multiple-choice tests: Impact on student grades. Frontiers in Education, 2, 24. https://doi.org/10.3389/feduc.2017.00024

Costley, J., Kapuza, A., Gorbunova, A., & Shcheglova, I. (2025). How adding structure reduces complexity: More interconnections in concept maps do not increase cognitive load. Instructional Science, 53(5), 1243–1262. https://doi.org/10.1007/s11251-025-09736-5

Credé, M., & Phillips, L.A. (2011). A meta-analytic review of the Motivated Strategies for Learning Questionnaire. Learning and Individual Differences, 21(4), 337–346. https://doi.org/10.1016/j.lindif.2011.03.002

de Bruin, A.B.H., Roelle, J., Carpenter, S.K., Baars, M., & Efg-Mre (2020). Synthesizing Cognitive Load and Self-Regulation Theory: A theoretical framework and research agenda. Educational Psychology Review, 32(4), 903–915. https://doi.org/10.1007/s10648-020-09576-4

Dignath, C., Van Ewijk, R., Perels, F., & Fabriz, S. (2023). Let learners monitor the learning content and their learning behavior! A meta-analysis on the effectiveness of tools to foster monitoring. Educational Psychology Review, 35(2), 62. https://doi.org/10.1007/s10648-023-09718-4

Dörrenbächer-Ulrich, L., Weißenfels, M., Russer, L., & Perels, F. (2021). Multimethod assessment of self-regulated learning in college students: Different methods for different components? Instructional Science, 49(1), 137–163. https://doi.org/10.1007/s11251-020-09533-2

Efron, B., & Tibshirani, R. (1994). An introduction to the Bootstrap. Chapman & Hall/CRC. https://doi.org/10.1201/9780429246593

Engelmann, K., Bannert, M., & Melzner, N. (2021). Do self-created metacognitive prompts promote short- and long-term effects in computer-based learning environments? Research and Practice in Technology Enhanced Learning, 1, 16. https://doi.org/10.1186/s41039-021-00148-w

Galperin, P.Ya. (1969). Stages in the development of mental acts. In M. Cole & I. Maltzman (Eds.) A handbook of contemporary Soviet psychology (pp. 249-273). Basic Books.

Glogger-Frey, I., Gaus, K., & Renkl, A. (2016). Learning from direct instruction: Best prepared by several self-regulated or guided invention activities? Learning and Instruction, 51, 26–35. https://doi.org/10.1016/j.learninstruc.2016.11.002

Gorbunova, A., Adamovich, K.A., Savelyev, A., & Costley, J. (2026). Behavior patterns characterize students’ choices and relate to cognitive load and performance in learner-controlled environments. The Internet and Higher Education, 69, 101073. https://doi.org/10.1016/j.iheduc.2025.101073

Greenberg, K., & Zheng, R. (2023). Revisiting the debate on germane cognitive load versus germane resources. Journal of Cognitive Psychology, 35(3), 295–314. https://doi.org/10.1080/20445911.2022.2159416

Guo, L. (2022). The effects of self-monitoring on strategy use and academic performance: A meta-analysis. International Journal of Educational Research, 112, 101939. https://doi.org/10.1016/j.ijer.2022.101939

Harrell, F.E. (2015). Regression modeling strategies: With applications to linear models, logistic and ordinal regression, and survival analysis (2nd ed.). Springer. https://doi.org/10.1007/978-3-319-19425-7

Huff, J.D., & Nietfeld, J.L. (2009). Using strategy instruction and confidence judgments to improve metacognitive monitoring. Metacognition and Learning, 4(2), 161–176. https://doi.org/10.1007/s11409-009-9042-8

Jung, J., Shin, Y., Chung, H., & Fanguy, M. (2024). The effects of pre-training types on cognitive load, self-efficacy, and problem-solving in computer programming. Journal of Computing in Higher Education, 37(3), 809–830. https://doi.org/10.1007/s12528-024-09407-3

Kalyuga, S., & Singh, A. (2015). Rethinking the boundaries of cognitive load theory in complex learning. Educational Psychology Review, 28(4), 831–852. https://doi.org/10.1007/s10648-015-9352-0

Karpov, A.A., Karpov, A.V., Karabushchenko, N.B., & Ivashchenko, A.V. (2017). The interconnection of learning ability and the organization of metacognitive processes and traits of personality. Psychology in Russia: State of the Art, 10(1), 67–79. https://doi.org/10.11621/pir.2017.0105

Krieglstein, F., Beege, M., Rey, G.D., Sanchez-Stockhammer, C., & Schneider, S. (2023). Development and validation of a theory-based questionnaire to measure different types of cognitive load. Educational Psychology Review, 35(1), 9. https://doi.org/10.1007/s10648-023-09738-0

Löffler, E., Von Der Linden, N., & Schneider, W. (2016). Influence of domain knowledge on monitoring performance across the life span. Journal of Cognition and Development, 17(5), 765–785. https://doi.org/10.1080/15248372.2016.1208204

Merkt, M., & Bodemer, D. (2024). Learning with videos: Do task instructions and the availability of a pause button matter? Journal of Computer Assisted Learning, 40(6), 2856–2871. https://doi.org/10.1111/jcal.13044

Mou, T.-Y. (2023). Online learning in the time of the COVID-19 crisis: Implications for the self-regulated learning of university design students. Active Learning in Higher Education, 24(2), 185-205. https://doi.org/10.1177/14697874211051226

Müller, N.M., & Seufert, T. (2018). Effects of self-regulation prompts in hypermedia learning on learning performance and self-efficacy. Learning and Instruction, 58, 1–11. https://doi.org/10.1016/j.learninstruc.2018.04.011

Nakagawa, S. (2004). A farewell to Bonferroni: The problems of low statistical power and publication bias. Behavioral Ecology, 15(6), 1044–1045. https://doi.org/10.1093/beheco/arh107

Nelson, T.O., & Narens, L. (1990). Metamemory: A theoretical framework and new findings. The Psychology of Learning and Motivation, 26, 125–141. https://doi.org/10.1016/S0079-7421(08)60053-5

Perikova, E. I., & Byzova, V. M. (2022). Faktornaia struktura russkoiazychnoi versii oprosnika "Metakognitivnaia vkliuchennost v deiatelnost" [Factor structure of the Russian version of the "Metacognitive Awareness Inventory"]. Kulturno-istoricheskaia psikhologiia [Cultural-Historical Psychology], 18(2), 116–126. https://doi.org/10.17759/chp.2022180213

Pikhart, M., Klimova, B., Cierniak-Emerych, A., & Dziuba, S. (2023). A comparative analysis of perceived advantages and disadvantages of online learning. Behavioral Sciences, 13(3), 262. https://doi.org/10.3390/bs13030262

Pintrich, P.R. (2000). The role of goal orientation in self-regulated learning. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation (pp. 451–502). Academic Press. https://doi.org/10.1016/B978-012109890-2/50043-3

Pham, N.S. (2022). The effectiveness of teaching and learning online: A study on HUFI English-majored students. International Journal of TESOL & Education, 2(3), 1–12. https://doi.org/10.54855/ijte.22231

Podolsky, A.I. (2023). Vstupitelnoe slovo [Foreword]. In P. Ya. Galperin, Psikhologiia: predmet i metod. Izbrannye psikhologicheskie trudy [Psychology: Subject and method. Selected psychological works] (pp. 13–32). Moscow University Press.

Rogozin, D., & Ipatova, A. (2020). Metodicheskii audit monitoringovykh sotsiologicheskikh issledovanii [Methodological audit of monitoring social research]. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3663255

Schraw, G., Crippen, K.J., & Hartley, K. (2006). Promoting self-regulation in science education: Metacognition as part of a broader perspective on learning. Research in Science Education, 36(1–2), 111–139. https://doi.org/10.1007/s11165-005-3917-8 

Schraw, G., & Dennison, R.S. (1994). Assessing metacognitive awareness. Contemporary Educational Psychology, 19(4), 460–475. https://doi.org/10.1006/ceps.1994.1033

Schunk, D.H. (1982). Progress self-monitoring. The Journal of Experimental Education, 51(2), 89–93. https://doi.org/10.1080/00220973.1982.11011845 

Seufert, T. (2018). The interplay between self-regulation in learning and cognitive load. Educational Research Review, 24, 116–129. https://doi.org/10.1016/j.edurev.2018.03.004

Shi, Y., & Qu, S. (2022). Analysis of the effect of cognitive ability on academic achievement: Moderating role of self-monitoring. Frontiers in Psychology, 13, 996504. https://doi.org/10.3389/fpsyg.2022.996504

Sidi, Y., & Ackerman, R. (2024). Opting out as an untapped resource in instructional design: Review and implications. Educational Psychology Review, 36(2), 41. https://doi.org/10.1007/s10648-024-09879-w

Singh, S., Loh, J. Y., & Lin, P.K.F. (2024). Impact of online learning experience during COVID-19: Exploring the relationship between academic resilience, academic self-efficacy, and perceived social support. In M. Zangeneh (Ed.), Essentials in health and mental health: Unlocking the keys to wellness (pp. 39–56). Springer Nature. https://doi.org/10.1007/978-3-031-56192-4_3

Sistek-Chandler C.M. (2019). Exploring online learning through synchronous and asynchronous instructional methods. In Advances in mobile and distance learning book series. IGI Global. https://doi.org/10.4018/978-1-7998-1622-5

Skulmowski, A., & Xu, K.M. (2021). Understanding cognitive load in digital and online learning: A new perspective on extraneous cognitive load. Educational Psychology Review, 34(1), 171–196. https://doi.org/10.1007/s10648-021-09624-7

Sweller, J., van Merriënboer, J.J., & Paas, F.G. (1998). Cognitive architecture and instructional design. Educational Psychology Review, 10, 251–296. https://doi.org/10.1023/A:1022193728205

Sweller, J., Van Merriënboer, J.J., & Paas, F. (2019). Cognitive architecture and instructional design: 20 years later. Educational Psychology Review, 31(2), 261–292. https://doi.org/10.1007/s10648-019-09465-5

Tabachnick, B.G., & Fidell, L.S. (2006). Using multivariate statistics (5th ed.). Pearson.

Talyzina, N.F. (1981) The psychology of learning. Progress.

Viberg, O., Khalil, M., & Baars, M. (2020). Self-regulated learning and learning analytics in online learning environments: A review of empirical research. In Proceedings of the Tenth International Conference on Learning Analytics & Knowledge (LAK '20), 524–533. https://doi.org/10.1145/3375462.3375483

Winne, P.H., & Hadwin, A.F. (1998). Studying as self-regulated learning. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.), Metacognition in educational theory and practice (pp. 277–304). Lawrence Erlbaum Associates.

Wong, J., Baars, M., Davis, D., Van Der Zee, T., Houben, G.J., & Paas, F. (2019). Supporting self-regulated learning in online learning environments and MOOCs: A systematic review. International Journal of Human–Computer Interaction, 35(4-5), 356–373. https://doi.org/10.1080/10447318.2018.1543084

Zimmerman, B.J. (2000). Attaining self-regulation: A social cognitive perspective. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation (pp. 13–39). Academic Press. 

Zimmerman, B.J., & Moylan, A.R. (2009). Self-regulation: Where metacognition and motivation intersect. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.), Handbook of metacognition in education (pp. 299–315). Routledge.

Zimmerman, B.J., & Schunk, D.H. (2011). Self-regulated learning and performance: An introduction and an overview. In B. J. Zimmerman & D. H. Schunk (Eds.), Handbook of self-regulation of learning and performance (pp. 1–12). Routledge.