健康与遗传心理学研究室知识库

随着教育脱贫攻坚纵深推进，我国西部地区义务教育痛点问题已由“有学上” 转为“上好学”，特别是西部农村课堂教学质量不高，导致学生学习困难和厌学问题严重，甚至辍学。因此，如何有针对性地开展西部农村课堂教学改革，解决学生学习困难和厌学问题，是巩固拓展教育脱贫攻坚成果同乡村振兴战略高效衔接、推动义务教育实现区域均衡发展的重要举措。

示例演练学习(Learning Mathematics from Examples and by Doing , LFED)是一种学生通过考察有解例题和解决问题进行的自适应学习，将其应用于课堂教学被称为示例演练教学法。前人研究发现，示例演练教学法能提高学生学业成绩、 激发学习积极性、减轻师生负担。该方法强调学生自主学习，教师仅提供自主支持，在教学质量相对落后的西部农村课堂教学适用性方面具有巨大潜能。

本研究以西部农村初中数学教学为例，系统研究示例演练学习的成效和机制，主要包含四部分内容。

研究一旨在探讨我国西部初中数学教学方式与学生学业表现之间的城乡差异。研究基于中国教育追踪调查(China Education Panel Survey, CEPS)2013-2014 年和 2014-2015 年数据，以数学学业成绩、数学学业自我效能感作为因变量，以数学课堂教学方式(老师讲授、分组讨论和师生互动)作为自变量，以学校地理位置(城市、农村)作为调节变量，提取 2212 个有效数据。差异检验结果表明，西部地区初中数学课堂教学方式存在显著城乡差异，农村中学老师讲授、分组讨论和师生互动三种教学方式使用频率均显著低于城区中学，效应量分别为 0.32、0.48 和 0.94；西部初中生数学学业表现存在显著城乡差异，农村学生学业成绩显著低于城区学生，效应量为 0.99，农村学生学业效能感显著低于城区学生，效应量为 0.40。调节效应检验结果表明，(1)学校位置对老师讲授和学生学业成绩关系调节作用显著，对老师讲授和学生学业效能感调节作用不显著。具体来说，在城区中学，老师讲授越多，学生学业成绩越低；而在农村中学，老师讲授并不能显著预测学生学业成绩，但老师讲授越多，学生学业效能感越低。(2)学校位置对分组讨论和学生学业成绩、学业效能感关系调节作用都不显著。但在城区中学，分组讨论越多，学生学业成绩越高，学业效能感越高；而在农村中学，分组讨论并不能显著预测学生学业成绩和学业效能感。(3)学校位置对师生互动和学生学业成绩、 学业效能感关系调节作用显著。在城区中学，学生学业成绩、学业效能感会随着 师生互动增多而增加；而在农村中学，师生互动不能显著预测学生学业成绩，但能提高学生学业效能感。结果提示：过多采用教师讲授的教学方式并不能提升学生学习成绩，反而会降低农村学生学业效能感；增加以学生为主体的教学方式（分组讨论和师生互动）仅对城区中学学习成绩提升有益，提示农村中学不能照搬城 区中学的教学方式，应结合学生基础和师资情况进一步探索适用于农村数学课堂的教学方式。

基于研究一发现，研究二考察示例演练教学法对西部农村初中学生数学学业成绩、学习投入和学业自我效能感的影响，共分为三个子研究。研究二(a)、研究 二(b)考察示例演练教学法作为数学课堂干预模式对学生学业成绩、学习投入和学业自我效能感的影响；研究二(c)考察示例演练教学法作为数学课前预习干预模式对学生学业成绩、学习投入和学业自我效能感的影响。三个子研究共招募 3 个 学校 6 个班级 240 名七年级学生。以班级为单位随机分配学生到示例演练教学干预组(简称干预组)和常规教学组(简称对照组)。收集所有学生学业成绩(学业等级)、 学习投入度和学业自我效能感基线数据后，对干预组采用示例演练教学法，对照组采用常规教学方式。结果发现，示例演练教学干预对维持或提高学生数学学业自我效能感、学习投入和学业成绩具有积极效果。

研究三基于研究二 240 名学生后测数据，建立链式中介效应模型，考察示例演练学习对学生学业自我效能感的影响机制，以及自主学习需要和学习投入对两者关系的影响。结果表明：示例演练学习显著正向预测学业自我效能感，自主学习需求和学习投入在二者之间起链式中介作用，即示例演练教学法通过满足学生自主学习需求，增加学生学习投入，进而提高其学业自我效能感。

研究四结合眼动追踪技术与口语报告技术，进一步考察示例演练学习中的线索效应以及示例演练学习的认知加工机制。研究四(a)考察示例演练学习中的线索效应，共招募 2 个学校 64 名八年级学生，最终有效样本 50 个。自变量为学习材料是否有线索，共两个水平(有线索、无线索)，将 50 名学生随机纳入线索组或对照组，对两组学生眼动指标(AOI 注视时间比例、AOI 注视次数比例、AOI 首次注视时长)、主观评价得分(认知负荷、自我效能感)和测验成绩得分(演练测验和 迁移测验)进行独立样本 t 检验。结果表明，线索会显著影响学生对示例注意分配和学习效果，线索组比对照组在示例产生式上注视时间更长且注视次数更多，示 例学习后演练测验和迁移测验成绩也更高。研究四(b)基于研究四(a)线索组 16 名 学生眼动数据和新招募的 6 名八年级学生口语报告数据，共同探讨示例演练学习的认知加工机制。自变量为学生学业水平(高分组、低分组)。研究共分为两步， 首先对 16 名学生眼动指标(示例注视时长、演练注视时长、迁移测验注视时长)、 主观评价得分(认知负荷、自我效能感)和测验成绩得分(演练测验和迁移测验)进 行秩和检验并分别对比两组学生在解决不同难度几何问题(简单图形问题和复杂图形问题)的注视轨迹图和注视热点图。(1)秩和检验结果表明，学业水平会显著影响学生迁移测验注视时长、迁移测验成绩和认知负荷。具体表现为：高分组比低分组在迁移测验中解题效率更高、认知负荷更低。(2)对比两组学生在不同难度几何问题的注视热点图和注视轨迹图可以发现：从示例-演练-迁移测试，学生的注意力都分布在跟解题有关的线索区，三个阶段注视时间越来越短，注视点个数 越来越少；当测验问题为简单图形问题时，两组学生注视时长差异不明显；当测 验问题为复杂图形问题时，低分组学生在示例学习阶段和迁移测验阶段注视时长明显大于多于高分组。其次，对 6 名学生口语报告陈述进行编码、转译后，对比两组学生解决几何证明问题思维过程。结果发现：学生在演练阶段和迁移测验阶段解决几何问题时，主要基于对前一阶段产生式的条件建构和优化，表现为模式再认过程，并受到图形复杂程度和学生学业水平影响。口语报告分析与眼动追踪 实验结果相互印证，共同表明：学生的示例演练学习分为条件建构和条件优化两个阶段，能否再认问题情境中的条件线索是影响问题解决效率的关键因此，需要加强对产生式条件的学习。

综上，在教学实践层面，本研究结果证明示例演练学习在我国初中数学课堂教学具有较好的可行性和适用性，可有效维持或提高学生的数学学业成绩、学习投入和自我效能感，为解决因教学方式落后而导致的学生学习困难和厌学等问题提供心理学的解决方案；在理论研究层面，本研究结果支持基于自我决定理论解释示例演练教学法在真实课堂教学情境中的心理行为机制，提出示例演练学习的社会认知学习理论框架，为后续研究提供理论借鉴；在研究方法层面，本研究结果支持用眼动追踪技术和口语报告技术共同探究示例演练学习认知加工机制的必要性和重要性，为示例演练学习相关研究提供方法学参考。除此以外，本研究 进一步探讨了示例演练学习的线索效应，为优化示例演练学习材料供实证依据。

With the deepening of the education poverty alleviation campaign, the pain point of compulsory education in the western region of China has changed from "having a school" to "having a good school", which is specifically manifested in the following: the quality of classroom teaching in the rural areas of the western region is not high, which has led to students' learning difficulties and serious boredom of learning, and even dropping out of school. The problem is that the quality of classroom teaching in rural areas in western China is not high, leading to serious learning difficulties and boredom among students, who may even drop out. Therefore, how to carry out targeted classroom teaching reform in rural areas in the western region, and how to solve the problem of students' learning difficulties and boredom with schooling, is an important measure for consolidating and expanding the results of the education poverty alleviation and rural revitalization strategies, as well as for promoting the achievement of balanced development of compulsory education in the region.

Learning Mathematics from Examples and by Doing (LFED) is a kind of adaptive learning that combines learning by example and learning by doing, where students learn by examining solved examples and solving problems. Previous studies have found that applying this method to classroom teaching can improve students' academic performance, stimulate learning motivation, and reduce the burden on teachers and students. The method emphasizes students' independent learning, and teachers only provide independent support, which has great potential in terms of applicability to classroom teaching in rural secondary schools in the western region where the quality of teaching is relatively backward.

In this study, we conducted a systematic research on the effectiveness and mechanism of LFED by taking junior high school mathematics teaching in rural areas of the western region as an example, which contains four main parts.

The first study aims to explore the urban-rural differences between the teaching methods of middle school mathematics in western China and the academic performance of students. Based on data from the China Education Panel Survey (CEPS) in 2013- 2014 and 2014-2015, the study used three indicators, namely, mathematics academic achievement, perceived mathematics learning pressure, and academic self-efficacy, as the dependent variables of students' academic performance in mathematics, and mathematics classroom teaching modes (teacher teaching, group discussion, and teacher-student interaction) as the independent variables, and school geographic location (urban, rural) as the moderating effect, group discussion and teacher-student interaction) as the indicators of independent variables, and the geographical location of the school (urban, rural) as the moderating variables, 2212 valid data were extracted. The results of the difference test show that there is a significant urban-rural difference in the teaching methods of middle school mathematics in the western region. The frequency of using three teaching methods, namely teaching by rural middle school teachers, group discussions, and teacher-student interaction, is significantly lower than that of urban middle schools, with effect sizes of 0.32, 0.48, and 0.94, respectively; There is a significant urban-rural difference in the mathematical academic performance of middle school students in western China. The academic performance of rural students is significantly lower than that of urban students, with an effect size of 0.99. The academic efficacy of rural students is significantly lower than that of urban students, with an effect size of 0.40. The results of the moderation effect test indicate that (1) school location has a significant moderating effect on the relationship between teacher teaching and student academic performance, but has no significant moderating effect on teacher teaching and student academic efficacy. Specifically, in urban middle schools, the more teachers teach, the lower the academic performance of students; In rural middle schools, teaching by teachers does not significantly predict students' academic performance, but the more teachers teach, the lower their academic efficacy. (2) The moderating effect of school location on group discussions, student academic performance, and academic efficacy is not significant. But in urban high schools, the more group discussions, the higher the academic performance and self-efficacy of students; In rural middle schools, group discussions cannot significantly predict students' academic performance and academic efficacy. (3) The school location has a significant moderating effect on the interaction between teachers and students, as well as the relationship between student academic performance and academic efficacy. In urban middle schools, students' academic performance and self-efficacy will increase with the increase of teacher-student interaction; In rural middle schools, teacher-student interaction cannot significantly predict students' academic performance, but it can improve their academic efficacy. The results indicate that excessive use of teacher taught teaching methods does not improve student academic performance, but rather reduces the academic efficacy of rural students; Adding student-centered teaching methods (group discussions and teacher-student interaction) is only beneficial for improving academic performance in urban middle schools. It is suggested that rural middle schools should not copy the teaching methods of urban middle schools, and should further explore teaching methods suitable for rural mathematics classrooms based on the student foundation and teacher situation.

Based on the findings of Study 1, Study 2 examines the impact of LFED based on adaptive production learning theory on the mathematical academic performance, learning engagement, and academic self-efficacy of rural middle school students in the Western region. The study is divided into three sub-studies, among which Study 2a and Study 2b examine the impact of LFED as a mathematical classroom intervention mode on student academic performance, learning engagement, and academic self-efficacy; Study 2c examines the impact of LFED as a pre-class math preview intervention model on student academic performance, learning engagement, and academic self-efficacy. Three sub-studies recruited 240 seventh-grade students from 6 classes in 3 schools. On a class-by-class basis, students are randomly assigned to the LFED group (the intervention group) and the routine teaching group (the control group). After collecting baseline data on mathematics scores (or academic levels), learning engagement, and academic self-efficacy from all students, the intervention group was given LFED learning mode, while the control group was given a routine teaching mode. The results showed that the LFED intervention had a positive effect on maintaining or improving students' mathematical academic self-efficacy, learning engagement, and academic performance.

Based on the posttest data of 240 students in Study 2, Study 3 established a chain mediation effect model to investigate the mechanism of the influence of LFED on students' academic self-efficacy, as well as the influence of the need for self-directed learning and learning input on the relationship between the two. The results showed that the use of LFED had a significant positive predictive effect on academic selfefficacy, and independent learning needs and learning input played a chain mediating role between the two, i.e., the use of LFED could satisfy the independent learning needs of the students, and was conducive to increasing the learning input of the students, which in turn improved the academic self-efficacy of the students.

Study 4 combined eye-tracking technology with protocol analysis technology to further examine the cue effect in LFED and the cognitive processing mechanism of LFED. Study 4(a) examined the cue effect in LFED, recruiting a total of 64 eighthgrade students from two schools, with a final valid sample of 50. The independent variable was whether the learning material was cued or not, and there were two levels (cued, uncued). 50 students were randomly included in either the cued group or the control group, and the eye-movement metrics (proportion of time spent on AOI gaze, proportion of number of times spent on AOI gaze, and length of the first time spent on AOI gaze), the subjective evaluation scores (cognitive load, self-efficacy), and the quiz performance scores (rehearsal quiz and transfer quiz) were tested with independent sample t-tests. The results showed that cues significantly affected students' attention allocation to and learning of the exemplars, with the cued group having longer and more frequent gaze times on the exemplar-generating formula and higher scores on the rehearsal and transfer tests after exemplar learning than the control group. Study 4(b) is based on eye movement data from 16 students in the cued group of Study 4(a) and protocol data from 6 newly recruited 8th graders to explore the cognitive processing mechanisms of LFED jointly. The independent variables were students' academic levels (high and low subgroups). The study was divided into two steps, firstly, a rank-sum test was conducted on the eye movement indexes (example gaze duration, rehearsal gaze duration, transfer test gaze duration), subjective evaluation scores (cognitive load, selfefficacy), and test performance scores (rehearsal test and transfer test) of 16 students and compared the gaze trajectory graphs and attention hotspot maps of the two groups of students solving geometrical problems with different levels of difficulty (simple graphical problems and complex graphical problems). The results of the rank-sum test show that academic level significantly affects students' gaze duration, migration test performance, and cognitive load in the migration test. The results showed that students in the higher group were more efficient in solving the migration test, had lower cognitive load, and had a higher rate of solving the test correctly than students in the lower group. Comparing the attention hotspots and attention trajectories of the two groups of students in geometric problems of different levels of difficulty, it can be found that: from the example-exercise-transfer test, the students' attention is distributed in the cue area related to solving the problem, and the attention time in the three phases is getting shorter and shorter, with fewer and fewer attention points; the difference in the attention duration between the two groups of students is not obvious when the test problem is a simple graphical problem; and the difference between the two groups of students in the complex graphical problem is not obvious when the test problem is a complex graphical problem. When the quiz problem was a complex graphical problem, the difference between the two groups was not significant in the rehearsal stage, and the students in the lower group had significantly more gaze duration in the example learning stage and the transfer quiz stage than the higher group. Second, after coding and translating the oral report statements of the six students, the two groups of students' thinking processes in solving geometric proof problems were compared. It was found that: students solved geometric problems in the rehearsal stage and in the transfer test stage mainly based on the conditional construction and optimization of the generative formula of the previous stage, which manifested itself as a process of schema rerecognition, and that this process was influenced by the complexity of the graphs and the students' level of schooling. The results of protocol analysis and the results of the eye-tracking experiments corroborate each other, and together they suggest that students' LFED is divided into two phases: conditional construction and conditional optimization and that being able to sensitively re-recognize conditional cues in the problem situation is the key to the efficiency of problem solving, and that the learning of generative conditioning needs to be strengthened.

In summary, at the level of teaching practice, the results of this study prove that LFED is feasible and applicable in the teaching of junior high school mathematics classrooms in rural areas of western China, which can effectively maintain or improve students' academic performance, learning engagement and self-efficacy in mathematics, and provide psychological solutions to solve the problems of students' learning difficulties and boredom in rural areas in the western region due to the backwardness of the teaching mode; at the level of theory and research, the results of this study support the mechanisms based on self-determination theory and self-efficacy theory to explain the role of LFED in real classroom teaching contexts. At the theoretical level, the results of this study support the mechanisms based on self-determination theory and self-efficacy theory to explain the role of LFED in real classroom teaching situations, find that students' independent learning needs and learning engagement play an important role in LFED and put forward a theoretical framework of the social learning theory of LFED, which provides theoretical references for the subsequent related studies. At the methodological level, the results of this study support the necessity and importance of exploring the cognitive processing mechanism of LFED with eye-tracking technology and protocol analysis technology, which provides a methodological reference for subsequent studies related to example rehearsal learning. In addition, this study further explored the cueing effect in LFED, which provides empirical evidence for further optimizing LFED materials.