學習障礙與一般學生學業自我知覺與學業表現發展差異之研究

A Study on the Developmental Differences in Academic Self-Perception and Performance between Students with and without Learning Disabilities

過去研究已證實學障學生學業自我知覺過度樂觀，但仍欠缺從發展性的觀點考驗。本研究針對336位一般學生和268位學障學生，進行為期一年半的追蹤調查，除透過自我評估外，也透過教師評定，來檢視學障學生對學業自我效能是否隨時間有過度高估的趨勢。結果指出，一般學生與學障學生都有穩定的學業表現與學業自我效能的發展軌跡；再者，一般學生不管在學業自我效能，或學業成就都顯著高於學障學生，但是學障學生顯現學業自我效能高於自身學習表現，一般學生則相反；最後，一般學生自評的學業成就與他們教師對其學業成就的評定之間無顯著差異，學障學生自評顯著比普通班教師對其評定來得高，而且差異逐漸拉大，而相較之下，特教教師則對學障學生有較佳的評定。本研究針對結果提出研究與實務上的建議。

Previous research has established that students with learning disabilities (LD) tend to demonstrate overly optimistic perceptions of their academic capabilities compared to their actual performance. However, most studies have relied on cross-sectional data, leaving gaps in our understanding of how these self-perceptions develop over time. This longitudinal study examined the developmental trajectories of academic self-efficacy (ASE) and academic achievement (AA) among students with and without LD over an 18-month period.
The study pursued three main research questions: (1) How do the developmental trajectories of ASE and AA differ between LD and non-LD students? (2) What are the differences in developmental trajectories between the two groups regarding the discrepancy between ASE and AA? (3) How do student self-assessments compare with teacher evaluations over time?
Participants included 336 general education students (167 males) and 268 LD students (172 males) from junior high schools in Kaohsiung, Taiwan. The sampling process followed a stratified purposive approach based on four urbanization levels: low, medium-low, medium, and high urbanization. The proportion of students selected from each urbanization level matched the overall population distribution, with approximately 66.2% from highly urbanized areas, 18.6% from medium urbanized areas, 13.5% from medium-low urbanized areas, and 1.7% from low urbanized areas.
For general education students, sampling was conducted at the class level by selecting one class per grade from each chosen school. For LD students, selection was based on official identification according to Taiwan’s special education regulations, considering gender distribution, urban-rural location, and grade level to ensure representative sampling. All LD students received 10–13 hours of resource room support weekly.
Data collection occurred at three time points over 18 months: November 2019, May 2020, and November 2020. General education students completed questionnaires during morning study periods, while LD students completed them during resource room sessions. Teachers assisted with reading when necessary, and completion time averaged 30 minutes. For absent students, teachers facilitated make-up sessions. No participants dropped out or transferred during the study period.
For teacher evaluations, 11 homeroom teachers assessed general education students, each evaluating 23–30 students. For LD students, 236 general education teachers and 187 special education teachers participated, with general education teachers typically assessing 1–2 students and special education teachers evaluating 1–3 students each. There was no teacher turnover during the study period. Teachers completed evaluations during their free time using detailed rating guidelines provided by researchers.
The study employed two primary instruments to assess academic achievement and self-efficacy. The Academic Performance Rating Scale (APRS) was used to measure academic achievement through 19 items across three subscales. Both students and teachers completed this assessment using a 5-point Likert scale. The Chinese version of the APRS underwent comprehensive translation and validation procedures, including initial translation by two bilingual experts, back-translation verification, expert review by three scholars, pilot testing with 20 first-grade LD students, and field testing with 185 general education students. The validation process demonstrated strong psychometric properties, with internal consistency reliability (Cronbach’sα) ranging from .81 to .95 across different raters and time points, and test-retest reliability over two weeks ranging from .88 to .95. Factor analysis supported the scale’s construct validity, with Kaiser-Meyer-Olkin sampling adequacy of 0.82 and Bartlett’s test of sphericity of 676.63 (p < .001), revealing a single factor that explained 49.97% of variance.
The second instrument was the Self-Efficacy Questionnaire for Children (SEQ-C), from which only the academic self-efficacy subscale containing 7 items was utilized. This scale also employed a 5-point Likert format and underwent the same rigorous translation and validation process as the APRS. The Chinese version demonstrated satisfactory psychometric properties, with Cronbach’sαranging from .80 to .83 across time points and item reliability coefficients between .46 and .61. Factor analysis revealed a single factor structure explaining 48.69% of variance, with KMO sampling adequacy of .65 and Bartlett’s test of sphericity of 109.70 (p < .001). Sample items included questions about students’confidence in their ability to focus during lessons and complete academic tasks successfully.
The study employed sophisticated statistical techniques for data analysis. Initial preliminary analyses using SPSS 24.0 included descriptive statistics, t-tests, and correlation analyses to examine group differences and relationships among variables. The primary analytical approach utilized latent growth modeling to examine within-group developmental trajectories for ASE and AA separately. These models included three observation indicators representing the three time points, with parameters set at 0 (2019/12), 0.5 (2020/05), and 1 (2020/12) to assume linear growth. Additionally, conditional latent growth modeling was employed to compare trajectories between groups by adding conditional variables to the basic latent growth models. For group comparisons, general education students were coded as 0 and LD students as 1, while dummy coding was used to distinguish between student self-evaluations and different teacher ratings for evaluation comparisons.
Model fit was assessed using multiple indices including chi-square tests, CFI (threshold > .90), RMSEA (threshold < .08), and SRMR (threshold < .05). When necessary, models were modified by allowing correlations between residuals based on modification indices while maintaining theoretical justification. The comprehensive analysis plan enabled examination of both individual growth trajectories and group differences while accounting for the nested structure of the longitudinal data.
1. Stable Trajectories: Both groups demonstrated stable developmental trajectories in ASE and AA over the 18-month period. For general education students, the initial ASE score was 3.35 with a non-significant growth rate of 0.02, while their AA showed an initial score of 3.45 with a non-significant decline of -0.06. LD students showed an initial ASE of 2.97 with a non-significant decline of -0.13, and an AA initial score of 2.82 with a non-significant decline of -0.07.
2. Group Differences: General education students consistently outperformed LD students in both ASE and AA. The difference in ASE between groups showed an initial gap of -0.23 that remained stable over time. The AA gap was larger at -0.48 and also remained consistent throughout the study period.
3. Self-Perception Patterns: LD students consistently overestimated their academic abilities, with ASE scores exceeding their AA by approximately 0.15 points. In contrast, general education students slightly underestimated their abilities, with ASE scores about 0.12 points below their AA.
4. Teacher Evaluations: For general education students, self-assessments aligned closely with teacher evaluations, showing no significant differences. However, LD students’self-assessments were significantly higher than their general education teachers’evaluations (difference of 0.26), and this gap widened over time. Special education teachers’evaluations of LD students were higher than general education teachers’ratings and closer to students’self-assessments, with an initial difference of -0.26 that showed a significant increase of 0.23 over time.
The findings support and extend previous research showing that LD students maintain overly optimistic academic self-perceptions despite lower actual achievement. This pattern remained stable throughout the study period, suggesting it may serve as an enduring psychological mechanism rather than a temporary developmental phase.
Two primary theoretical explanations emerge for this phenomenon. The cognitive immaturity hypothesis suggests that LD students lack the metacognitive skills necessary for accurate self-assessment, leading to persistent miscalibration between perceived and actual abilities. The self-protective hypothesis posits that inflated self-perceptions serve as a psychological buffer against the emotional impact of repeated academic failures, helping maintain motivation and engagement in academic tasks despite challenges.
The divergent evaluations between general and special education teachers highlight important considerations in assessing LD students’academic performance. Special education teachers’more positive evaluations may reflect their specialized training and deeper understanding of LD students’capabilities and challenges. Alternatively, these differences might indicate varying assessment standards or expectations between general and special education settings.
These findings have significant implications for educational practice. While maintaining positive self-perceptions may protect LD students’psychological well-being, significant miscalibration could impede effective learning strategies and academic progress. Educational interventions should aim to help LD students develop more accurate self-assessment skills while preserving their academic motivation and emotional well-being. This might include structured feedback systems, metacognitive training, and targeted support for developing realistic academic goals.
The study also emphasizes the need for better coordination between general and special education teachers in evaluating LD students’performance. Professional development focusing on standardized assessment practices and understanding of learning disabilities could help reduce evaluation disparities and provide more consistent feedback to students. Future research should examine the long-term impact of various intervention strategies and investigate the role of additional factors such as classroom environment, peer relationships, and parental involvement in shaping LD students’academic self-perceptions and achievement.