2.1. Perceived Lightweight Teaching Quality

Perceived lightweight teaching quality refers to the students’ subjective evaluation of the effectiveness of lightweight digital resources (such as short videos and concise case studies) adopted in the teaching process in terms of reducing cognitive load, enhancing learning interest, and facilitating knowledge. This construct is rooted in cognitive load theory, but it does not directly measure cognitive load itself; rather, it measures students’ perceptions of the quality of instructional design in effectively supporting their learning. In the field of educational measurement, students’ perceptions of instructional quality have been widely used as a valid indicator of instructional effectiveness. For example, the course experience questionnaire (CEQ) evaluates the instructional performance of higher education institutions by measuring students’ perceptions of course quality ¹⁰, while the instructional skills questionnaire (ISQ) predicts students’ self-reported learning outcomes based on their perceptions of instructors’ teaching skills ¹¹. More directly relevant to the measurement content of this study, existing research indicates that students’ perceptions of the relevance of instructional content are significantly associated with the cognitive load they experience ¹², while perceptions of instructional design quality have also been found to indirectly influence learners’ engagement by affecting cognitive load ¹³. These studies provide the theoretical and empirical foundation for this study to treat “perceived quality of lightweight instruction” as a predictor of students’ learning behaviors and perceived competencies. In terms of measurement, the perceived quality of lightweight instruction scale draws upon the multimedia learning experience scale developed by Ljubojević et al. ¹ and has been contextually adapted to the specific context of lightweight instruction (short video resources, local case studies, etc.). The items cover two dimensions: cognitive load reduction (for example, “videos/case studies help understand abstract theories”) and emotional empowerment (for example, “materials make classes more vivid and engaging”). It is important to note that the perceived quality of lightweight instruction (PLTQ) and cognitive load are distinct constructs. PLTQ measures the students’ subjective evaluations of instructional design, whereas cognitive load measures the actual consumption of working memory resources during the learning process ³. This study treated PLTQ as a predictor of cognitive load rather than a substitute measure. Theoretically, students’ positive evaluations of the usefulness and engaging nature of instructional resources reflect the effectiveness of resource design in reducing extrinsic cognitive load. When students perceive videos or case studies as helpful for understanding abstract theories, this implies that these resources assist them in breaking down complex information into more manageable units, thereby freeing up working memory capacity. Empirically, prior research has shown that students’ perceptions of instructional presence within an inquiry-based community framework can effectively predict their cognitive load. In particular, instructional presence has been shown to be the best predictor of external and total cognitive load ¹⁴. Furthermore, students’ evaluations of the perceived quality of instruction have been found to significantly influence their cognitive, affective, and behavioral responses during the learning process ¹⁵. Collectively, this evidence suggests that students’ subjective evaluations of instructional design are not only indicators of their learning experiences but also key antecedents influencing the efficiency of their cognitive processing. This provides strong justification for this study’s positioning of the PLTQ as an antecedent variable predicting student behavior and competence. All items use a 5-point Likert scale. The scale demonstrated excellent internal consistency reliability in the sample of this study (Cronbach’s \(\alpha=0.962\)).

2.2. Post-Class Case Sharing Behavior

This study focused on an observable, explicit behavior in which students apply classroom learning to informal communication settings, operationalized as post-class case sharing behavior. This definition comes from the “active experimentation” stage in the experiential learning theory and serves as a key indicator of knowledge internalization and competence development. Given the multidimensional nature and measurement complexity of in-depth post-class case sharing, this study employed a single behavioral indicator for exploratory purposes: questionnaire item Q13: “After class, I have used course cases to explain management phenomena to others.” In exploratory research, it is a valid approach to use single-item measures with high content validity for specific, overt behaviors with a single target ¹⁶,¹⁷. This item is specific in wording and clear in its focus, and can effectively capture the most direct knowledge application and sharing behaviors stimulated by lightweight instruction. This operational definition aims to precisely measure the specific intermediary of “sharing behavior,” rather than encompassing all dimensions of learning transfer. Future studies should adopt more comprehensive scales to expand on this line of research.

2.3. Perceived Competence

Perceived competence refers to students’ overall subjective evaluation of improvements in knowledge, skills, and beliefs through course learning. Its theoretical framework is derived from the classic training evaluation model while integrating the perspective of self-determination theory (SDT), focusing primarily on the effectiveness of learning at the behavioral and outcome levels ⁵, particularly on enhancing learners’ competence ¹⁸. In terms of measurement, this study employed a seven-item scale (Q14–Q20) covering multiple dimensions, including knowledge mastery, confidence in application, problem-solving, and professional identity. The application and transfer dimension correspond to Kirkpatrick’s “behavioral level” concept ⁵; the self-efficacy dimension is based on Bandura’s self-efficacy theory ¹⁹; and the professional identity dimension draws on the operationalization approach of the occupational identity scale. The scale uses a 5-point Likert scale (1 \(=\) no improvement, 5 \(=\) significant improvement), and the mean of all items is calculated as the score for perceived competence. This scale demonstrated excellent reliability in this study (Cronbach’s \(\alpha=0.976\)), indicating its ability to consistently measure the construct.

2.4. Academic Major (Moderator/Grouping Variable)

AM’s influence essentially comes from the professional cognitive system it forms. Schemas determine how individuals filter, interpret, and integrate information ⁶. In this study, we posited perceived case relevance as a key mediating variable. For instance, due to schema differences, students in financial management may generally perceive a management case from Dream of the Red Chamber as being weakly relevant to their core domains of financial analysis and risk control. This initial low relevance perception would systematically undermine their subsequent learning engagement and transfer willingness, rather than impairing the efficiency of translating positive teaching experiences into behavior. Therefore, this paper proposes a chain-mediation model (as shown in Fig. 1): AM influences perceived case relevance, which in turn influences perceived lightweight teaching quality, which further influences post-class case sharing behavior, and ultimately influences perceived competence.

2.5. Control Variables

To account for other potential influences, this study included two control variables:

1. ・

   Self-Assessed Performance: We controlled for students’ self-reported mastery of course content (Q3). The theoretical basis for this is that cognitive load theory posits that learners’ prior knowledge is a key factor in moderating internal cognitive load—learners with extensive prior knowledge can process more information automatically, thereby freeing up cognitive resources for deep processing ³. By controlling for self-assessed performance, we can partially eliminate the confounding effect of differences in prior knowledge on the results.

2. ・

   Course Semester: This is controlled using a dummy variable based on the student’s actual semester of enrollment (current semester or previous semester, corresponding to questionnaire item Q4). The theoretical basis for this is the memory decay effect (Ebbinghaus’s forgetting curve)—the longer the time since the course was taken, the lower the retention rate of course content. By controlling for the course period, we can, to some extent, mitigate the potential influence of memory decay and differences in teaching cycles on the research results.

3.1. Perceived Lightness of Instruction and Perceived Competence

Cognitive load theory indicates that optimized instructional design can effectively reduce external cognitive load during the learning process, thereby freeing up more cognitive resources for knowledge construction and in-depth processing. Lightweight instruction is based on the principles of multimedia learning cognition and cognitive load optimization ²⁰, incorporating lightweight teaching resources, such as short videos and localized cases, to visualize and contextualize abstract theories, and aiming to enhance the fluidity and immersive quality of the learning process. This approach has been proven to boost multimedia learning efficiency effectively. The facilitating effect of this instructional model on students’ perceived competence can be analyzed in two dimensions. First, lightweight resources can directly stimulate students’ learning interest and intrinsic motivation by lowering the cognitive threshold. According to SDT, when the learning environment supports students’ autonomy and competence, their intrinsic motivation is more likely to be stimulated. When students perceive learning materials as easy to understand and engaging, their emotional experiences become more positive, laying a psychological foundation for active knowledge absorption and internalization ¹⁵. Second, based on experiential learning theory, effective learning requires a complete cycle of concrete experience, reflective observation, abstract conceptualization, and active experimentation. The rich scenarios and case studies provided by lightweight instruction create high-quality concrete experiences for students, bridging theory and practice to enhance their confidence in mastering and applying knowledge (that is, competence). Empirical research has demonstrated the positive effects of lightweight or similar teaching strategies on learning engagement and knowledge acquisition. From a theoretical perspective, cognitive load theory offers an explanation at the cognitive level, suggesting that optimized design frees up cognitive resources; SDT provides an explanation at the motivational level, emphasizing the fulfillment of the need for autonomy and competence ²¹. Therefore, we propose H1:

1. H1:

   Perceived lightness of instruction positively influences students’ perceived competence.

3.2. Mediating Role of Post-Class Case Sharing Behavior

Although perceived lightness of instruction enhances learning experiences, the transition from “positive perception” to “competence enhancement” is not automatic. The Kirkpatrick evaluation model indicates that training effectiveness progresses through four levels: reaction, learning, behavior, and results. The transfer at the “behavioral level” is crucial for measuring whether learning outcomes translate into practice. According to the triadic interaction model of social cognitive theory, environmental, behavioral, and individual cognitive factors interact ²². In lightweight instruction, positive perceived lightness of instruction (environmental support) reduces the psychological cost of knowledge retrieval and application, thereby promoting students’ behaviors of externally sharing learned knowledge (for example, explaining concepts to others using cases). Their success directly and powerfully enhances students’ confidence in their knowledge mastery and application abilities—that is, it elevates their “perceived competence.” Without this crucial behavioral conversion step, positive learning experiences may remain confined to the affective level. Theoretically, Kirkpatrick’s four-level evaluation model positions the “behavioral level” as the pivotal point for transformation ⁵; Bandura’s social cognitive theory explains the interaction between environment, behavior, and outcomes ²²; and Kolb’s experiential learning theory identifies “active experimentation” as the final stage of the learning cycle ⁴. Therefore, we propose H2:

1. H2:

   Post-class case sharing behavior mediates the effect of perceived lightness of instruction on perceived competence.

3.3. Mechanism of Academic Major: Hypothesis of Front-End Cognitive Differences

AM shapes students’ unique knowledge schemas and problem awareness ⁶. In the context of lightweight instruction, these discipline-specific schemas may influence the perception–behavior–competence pathway through two distinct mechanisms.

The first is a moderation of the conversion process (H3a): even with equally positive perceptions of lightweight instruction, students’ motivation and efficiency in translating perception into practice (for example, sharing behavior) are stronger when cases align with their professional schemas (for example, business administration students analyzing general management cases). This reflects individual differences in how environmental support fulfills basic psychological needs, as outlined by SDT ²³. The second is the induction of a front-end cognitive difference (H3b): AM shapes the initial strength of the cognitive connection that students form with instructional cases. The theoretical foundation of front-end cognitive differences lies in schema theory. Schemas are abstract knowledge structures that store experiences and determine how individuals understand and interpret new information ⁶. When new information closely matches existing schemas, individuals can rapidly assimilate it, resulting in smooth cognitive processing and active engagement; when the match is low, individuals must engage in accommodation, which increases cognitive effort and may even lead to a sense of alienation. From a broader perspective on motivation, SDT suggests that intrinsic motivation is fostered when learning activities support individuals’ basic psychological needs for autonomy, competence, and relatedness ²¹. Applying this theoretical framework to the context of this study, the scarce identification of finance majors with comprehensive management cases essentially stems from a lack of relevance between the teaching cases and their “professional identity.” This inhibits the fulfillment of their need for “relatedness,” thereby affecting their learning engagement and intrinsic motivation. Therefore, pre-cognitive differences arise not only from schema matching but also from learners’ professional identity and motivational engagement. Consequently, the cognitive gap at the front end stems not only from schema matching but is also linked to learners’ professional identity and motivational engagement.

In the context of this study, the professional schemas of business administration students encompass comprehensive management knowledge, such as organizational behavior, team management, and leadership, exhibiting a high degree of schema fit with comprehensive management case studies, such as Dream of the Red Chamber and Romance of the Three Kingdoms. In contrast, the professional schemas of financial management students center on financial analysis, risk control, and financial statement interpretation, resulting in a naturally lower perceived relevance to comprehensive management case studies. This disparity in schema compatibility constitutes an “initial cognitive threshold” that constrains teaching effectiveness. We operationalize this as “perceived case relevance (PCR),” defined as the degree to which students subjectively perceive the connection between teaching cases and their professional knowledge systems. It is worth emphasizing that while “upstream cognitive differences” use professional background as an operational variable, their theoretical implications extend far beyond disciplinary classification itself. Professional background is a comprehensive manifestation of the cognitive schemas, problem awareness, and knowledge networks formed during the process of professional socialization; it serves as an observable proxy variable for front-end cognitive differences. In addition to professional background, front-end cognitive differences may also stem from individual-level factors, such as personality traits, prior educational experiences, interests, and cognitive styles, which will be discussed in the section on research limitations.

To clarify whether AM exerts its influence through a moderating or mediating mechanism, we propose the following set of comparable hypotheses:

1. H3a (Moderation Hypothesis):

   AM moderates the effect of perceived lightweight teaching quality on post-class case sharing behavior. Specifically, compared to students majoring in financial management, perceived quality of lightweight teaching has a stronger positive effect on the sharing behavior of students majoring in business administration.

1. H3b (Mediating Mechanism Hypothesis):

   AM influences subsequent variables through the mediating role of perceived case relevance. Specifically, business administration students’ perceived case relevance is significantly higher than that of financial management students (H3b1), and this difference is transmitted through the chain “perceived lightweight teaching quality \(\to\) post-class case sharing behavior” to perceived competence (H3b2).

Based on the above analysis, this study tested the aforementioned hypotheses simultaneously by constructing a chain mediation model (Fig. 1) (AM \(\to\) perceived case relevance \(\to\) perceived lightweight teaching quality \(\to\) post-class case sharing behavior \(\to\) perceived competence). The test results will reveal the actual roles of the two mechanisms: if the interaction term in H3a is significant, it supports the moderating mechanism; if the chained indirect effect in H3b2 is significant, it supports the front-end cognitive differences as the core transmission mechanism; and if neither is significant, it indicates that the scope of professional background’s influence needs to be redefined.

4.1. Research Sample and Data Collection

This study conducted an exploratory analysis of the mechanisms within a specific teaching context. The organizational behavior course was selected because it is a foundational course that inherently attracts students from multiple disciplines, providing an ideal setting to observe issues of disciplinary adaptation. The use of localized cases, such as Dream of the Red Chamber and Romance of the Three Kingdoms, represents a typical practice of lightweight instruction. This study recruited undergraduates majoring in business administration and financial management who took the organizational behavior course during the 2023–2024 and 2024–2025 academic years as research subjects. Cluster sampling was conducted, and data were collected through questionnaire surveys. The survey was conducted both online and offline at the end of the course to ensure thorough responses and a high completion rate. In total, 157 questionnaires were distributed. To maximize data quality and minimize interference from extraneous variables, screening question Q24 (“Did the teacher explain temperament types using Dream of Red Mansions during the semester when this course was offered?”) was implemented. This process excluded nine questionnaires from participants who did not fully engage in the case-based teaching, ensuring that all analyzed samples received an equivalent level of instructional intervention. Question Q23 (“To ensure questionnaire quality, please select ‘Somewhat Agree’”) was embedded as an attention check item. Eight questionnaires that failed screening were excluded, effectively eliminating instances of careless responses. By examining the correlation coefficient (\(r = -0.501\), \(p < 0.001\)) between reverse-scored item Q12 (some case studies lack strong relevance to key concepts) and forward-scored item Q6 (video/real-life examples helped me grasp abstract theories), we validated the data and excluded 24 contradictory-response questionnaires. After a rigorous data cleaning process, 116 valid samples were obtained, with a valid recovery rate of 73.9%. The sample composition is described in Table 1.

There were 116 valid samples, of which 36 were business administration majors (31.0%) and 80 were financial management majors (69.0%). Although the sample sizes of the two groups were unbalanced, the independent samples \(t\)-test and subsequent bootstrap analysis demonstrated good robustness to sample size differences. Moreover, as this study primarily focused on elucidating the underlying mechanisms, the distribution of this sample did not affect the reliability of the core conclusions.

Although the official course nature designation differed between the two programs, both courses used closed-book examinations for their final assessments in practice. Consequently, there were no substantive differences in student assessment pressure or grading methodology. To avoid controlling for nonexistent confounders, course nature was not included as a control variable, allowing for a more isolated examination of the AM’s effect.

4.2. Variable Measurement and Operationalization

All core variables were assessed on a 5-point Likert scale (1 \(=\) strongly disagree / no improvement to 5 \(=\) strongly agree / significant improvement). The operational definitions and measurement details for each variable are summarized in Table 2.

4.3. Explanation and Limitations Regarding the Validity of Single-Item Measurement

To measure post-class case sharing behavior (PCSB), this study used a single item (Q13: “After class, I used course cases to explain management phenomena to others”). We recognize that multi-item scales better capture behavioral multidimensionality (for example, frequency and depth). However, for a specific, overt, and singular core behavior, a well-targeted single-item measure is a valid and established approach in exploratory research ¹⁶,¹⁷. Item Q13 in this study was directly derived from the “active experimentation” stage in Experiential Learning Theory. Preliminary validity tests indicate that Q13 shows moderate-to-strong correlations with perceived competence (\(r = 0.52\)) and perceived lightweight teaching quality (\(r = 0.45\)). A major limitation is that a single item cannot distinguish the frequency, depth, proactiveness, or recipients of sharing, which may introduce measurement error. Subsequent research should employ multidimensional behavioral scales or combine them with ecological measurement methods, such as behavioral logs, to conduct robustness tests.

4.4. Data Analysis Methods

This study employed SPSS 30.0 and the PROCESS macro (Models 4 and 7) for data analysis, with Bootstrap sampling set at 5000 iterations. The analysis sequentially performed reliability and validity tests and descriptive statistics and employed Model 4 to examine the main and mediating effects (H1 and H2). Model 7 was used to test the moderating effect (H3a), while an independent samples \(t\)-test verified the front-end difference of H3b.

5.1. Descriptive Statistics and Preliminary Analysis

5.2. Reliability and Validity Testing

Before hypothesis testing, the reliability and validity of the multi-item scale were assessed. The Cronbach’s \(\alpha\) coefficient of PLTQ Scale (Q6–Q11) was 0.962, and that of PC Scale (Q14–Q20) was 0.976 (Table 4). All standardized factor loadings exceeded 0.7, and Cronbach’s \(\alpha\) exceeded 0.9, indicating excellent internal consistency of the scales. The AVE values for both latent variables exceeded 0.750, significantly surpassing the 0.5 benchmark, indicating excellent convergent validity.

Exploratory factor analysis (EFA) revealed that the KMO value was 0.921, and Bartlett’s test of sphericity was significant (\(\chi^{2} = 2263.661\), \(p < 0.001\)) (Table 5). The extracted two-factor structure (perceived lightness of instruction and perceptions of competence) was clear, with cumulative variance explained reaching 90.0%. All item factor loadings exceeded 0.70, indicating good construct validity.

5.3. Hypothesis Testing

5.3.1. Main Effects and Mediating Effects Testing (H1 and H2)

To explore the internal mechanisms underlying the significant positive impact of PLTQ on PC, this study further incorporated PCSB as a mediating variable into the model. Mediation effects were examined using Model 4 of the SPSS macro program process (with a bootstrap sample size of 5000). Using the Bootstrap method provided by Hayes ²⁴, we verified the mediating role of PCSB between PLTQ and PC. The path coefficients between PCSB, PLTQ, and PC are shown in Fig. 2.

As shown in Table 6, the bias-corrected 95% bootstrap confidence interval for the indirect effect of PLTQ on PC via PCSB did not include zero. This result confirms that PCSB significantly mediates the relationship between PLTQ and PC. The direct effect (0.644) and indirect effect (0.142) accounted for 81.9% and 18.1% of the total effect (0.786), respectively.

Table 6. Decomposition of total, direct, and mediating effects.

	Effect value	SE	LLCI	ULCI	Effect size	Conclusion
Total effect	0.786	0.098	0.591	0.981		Significant
Direct effect	0.644	0.091	0.464	0.824	81.9%	Significant
Mediating effect	0.142	0.046	0.045	0.228	18.1%	Significant

The standardized value of the indirect effect was 0.11. On the original scale (1–5 points), for every 1-point increase in perceived quality of lightweight teaching, perceived competence indirectly increased by 0.142 points through the mediating effect of post-class case-sharing behavior, accounting for 3.6% of the mean perceived competence score (3.910). This indirect effect accounts for 18.1% of the total effect, representing a small-to-moderate effect size, and is significant in terms of revealing the theoretical mechanisms and explaining individual differences.

A significant total effect of PLTQ on PC was observed. After introducing PCSB into the model, the direct effect of PLTQ on PC remained significant. Furthermore, the bootstrap confidence interval for the indirect effect via PCSB did not include zero. These results indicate that PCSB plays a significant partial mediating role in the relationship between PLTQ and PC. Thus, H1 and H2 were supported.

5.4. Robustness Testing

To examine the robustness of the research model and rule out course assessment format (examination vs. evaluation) as a potential alternative explanation, we investigated the moderating effect of course characteristics. First, PROCESS Model 1 ²⁴ was used to test the moderating effect on the total effect path (PLTQ \(\to\) PC). The results showed a nonsignificant interaction effect (\(b=-0.078\), \(p=0.733\), 95% bootstrap CI \([-0.581,0.302]\)). Second, to further assess robustness, we tested its moderation of the first stage of the mediation path (PLTQ \(\to\) PCSB). However, the model did not converge stably with the control variables included, precluding a reliable estimation. Nevertheless, the null finding for the total effect path moderation already excludes course nature as a significant moderator of the primary relationship. In summary, course nature did not significantly moderate the PLTQ \(\to\) PC path. This result rules out assessment format as a viable alternative explanation, thereby strengthening the claim that “disciplinary adaptation” is the core explanatory variable. The nonsignificant moderating effect may be attributed to the fact that both graded and nongraded courses employed identical lightweight case libraries (Q24 shows a 92% core case usage rate) and identical final assessment formats. Thus, the formal distinction in course labels did not capture the core mechanisms underlying effectiveness conversion.

The empirical findings support the core causal pathway “PLTQ \(\to\) PCSB \(\to\) PC” (H1 and H2). More importantly, hypothesis testing revealed that AM does not influence the overall conversion efficiency by modulating this pathway (H3a rejected) but rather by inducing an upstream difference in PCR (H3b1 supported). Furthermore, this upstream difference did not propagate further to PC (H3b2 rejected). This precisely pinpoints the origin of the “professional adaptation dilemma”—the cognitive starting point—and reveals that it does not necessarily lead to differentiation at the endpoint. This finding corroborates the substantive consistency of the assessment formats discussed in Section 3, enabling dual confirmation that the inter-professional differences identified in this study can be robustly attributed to the differences in professional cognitive schema shaped by disciplinary backgrounds, rather than superficial situational factors. Additionally, the insignificant moderating effect of course nature effectively rules out the alternative explanation of assessment pressure, thereby enabling the reverse validation that AM and the resultant differences in professional cognitive schema are the more stable and profound root causes underlying the differentiated effectiveness of lightweight instruction.

6.1. Research Findings

This study integrated cognitive load theory, experiential learning theory, and social cognitive theory. Using a sample of 116 business and management students, it constructed a model combining mediation and chained mediation to systematically analyze the mechanisms through which the effectiveness of lightweight teaching is translated into outcomes, as well as its boundary conditions. The main findings of this study are as follows:

1. (a)

   PLTQ significantly positively impacts students’ PC. Data analysis indicates that students’ positive experiences with lightweight instructional design can be directly and effectively translated into recognition of their own competence improvement (\(b=0.786\), \(p<0.001\)). This conclusion reaffirms that optimizing instructional design through “light-asset” resources, such as short videos and localized case studies, effectively reduces extraneous cognitive load, offering a viable pathway to enhance students’ learning gain and sense of competence. This provides direct evidence for practical reforms in business and management courses in the context of new liberal arts construction.

2. (b)

   PCSB acts as a partial intermediary between PLTQ and PC. The findings indicate that PLTQ not only directly influences PC but also indirectly promotes competence development by motivating students to apply and share their knowledge in post-class scenarios (indirect effect value \(=\) 0.142, 95% bootstrap CI \([0.045, 0.228]\)). This indicates that the behavioral transformation of “learning for application” is the key mechanism through which lightweight instruction achieves profound impact.

3. (c)

   AM influences PCR to create initial cognitive differences (H3b1 is supported), but these differences are not transmitted to PC via the “PLTQ \(\to\) PCSB” pathway (H3b2 is not supported). Business administration students had significantly higher PCR scores than financial management students (mean difference \(=\) 0.479, Cohen’s \(d=0.505\)); however, this initial difference did not lead to further differentiation in PC (there were no significant differences between the two groups across any PC dimensions). This finding reveals the true boundaries of the “major-fit dilemma”: the problem lies at the starting point, not in the process, and this initial difference can be bridged through instructional design.

In summary, the functional boundary model constructed in this study (Fig. 3) not only validates the core transformation pathway of “PLTQ \(\to\) PCSB \(\to\) PC” but also precisely locates the “front-end cognitive gap” at the starting point of the learning process, confirming that different starting points do not necessarily lead to divergence at the endpoint.

6.2. Research Implications

6.2.1. Theoretical Contributions

The primary theoretical contributions of this study are twohold.

1. (a)

   Precisely identifying the location and boundaries of the “front-end cognitive gap” to deepen our understanding of the mechanisms underlying the effectiveness of differentiated instruction. Existing research has largely focused on the direct or moderating effects of academic background on learning outcomes but it has lacked a detailed deconstruction of its underlying mechanisms. By distinguishing between the “front-end cognitive differences” and “process moderation” mechanisms and employing a chained mediation model for testing, this study found that AM does indeed result in significant initial cognitive disparity (PCR), but this disparity is not further transmitted to PC through the “PLTQ \(\to\) PCSB” chain. The theoretical significance of this finding lies in its precise identification of the root cause of the major-majority mismatch at the starting point of the learning process, rather than within the process itself; simultaneously, it reveals the crucial boundary condition that “differences at the starting point \(\ne\) differences at the endpoint.” This study combined schema theory with SDT to explain the mechanisms underlying the formation of front-end cognitive differences from both cognitive and motivational perspectives. This provides a new theoretical perspective for understanding the pathways through which other stable individual characteristics, such as prior knowledge and cognitive styles, exert their effects.

2. (b)

   Providing new boundary conditions for the application of cognitive load theory in differentiated teaching contexts. This study found that even when information presentation has been “streamlined,” if the content does not sufficiently align with learners’ professional cognitive schemas, its effectiveness will face a “cognitive threshold” in the initial stage. More importantly, once students cross this threshold, streamlined instruction consistently promotes the development of competencies among students from different disciplines (both H1 and H2 are significant). This suggests that when applying cognitive load theory in interdisciplinary teaching contexts, the focus should be on the boundary condition of initial cognitive schema alignment, rather than assuming that “streamlined” designs will automatically bridge all differences.

6.3. Research Limitations and Future Prospects

This study had several limitations, highlighting directions for future research. First, the sample was drawn from a single course at the same university, and the data were cross-sectional, limiting the external validity of the conclusions and the ability to draw causal inferences. Nevertheless, the core findings—namely, the significant effect of AM on PCR and the pathway through which PLTQ influences PC via PCSB—remain robust within the current sample. Future research could validate these findings through a longitudinal design involving multiple universities and courses. Second, there are limitations in the measurement of key variables. The use of a single-item measure for post-class case-sharing behavior and the absence of a specialized cognitive load scale may introduce measurement error. However, even with such errors, H1, H2, and H3b1 remain significant, indicating that these effects are robust. The indirect effect (0.142) identified in this study should be regarded as preliminary exploratory evidence. Future studies could employ multidimensional behavioral scales and cognitive load measures to assess behavioral transfer and cognitive processing more accurately. Finally, while this study focused on professional background as an operationalized variable of the cognitive schema, it did not measure individual-level factors, such as personality traits, interests, or cognitive styles, limiting a comprehensive explanation of the sources of “upstream cognitive differences.” Additionally, while the key nonsignificant finding (H3b2 not supported) revealed the boundary condition that “differences at the starting point \(\ne\) differences at the endpoint,” the underlying cognitive mechanisms remain unclear. Future research could incorporate individual difference variables while controlling for professional background and further analyze the interaction mechanisms between subject-specific effects and individual effects by measuring indicators such as study time and effort.

Despite these limitations, as an exploratory first step in deconstructing the professional adaptation mechanisms of lightweight teaching, this study’s proposed “front-end cognitive gap” framework and “diagnosis–design–optimization” model still provide a clear theoretical foundation and design starting point for subsequent research.