掌握同伴交往的问题解决能力是学前儿童社会情感学习的重要目标。当前儿童同伴交往问题频发，校园欺凌层出不穷，无不揭示出在学前期培育儿童同伴交往的问题解决能力的必要性和紧迫性。评估与干预是发展学前儿童同伴交往的问题解决能力的两大枢纽，但传统范式下的评估与干预存在局限性，亟待革新。推进数字化是儿童评估和教育教学改革的主旋律。作为数字化的技术支撑，人机交互因兼顾互动性、趣味性及便捷性等优势，为革新学前儿童评估与干预提供新思路。本研究以“学前儿童同伴交往的问题解决能力评估和干预”为话题，选择“人机交互”作为评估与干预的技术支撑，依据人际认知问题解决理论和具身认知理论，采用内容分析法、德尔菲法、情境访谈法、问卷调查法、教育实验法等，剖析了人机交互赋能学前儿童同伴交往的问题解决能力评估与干预的价值和方略，基于人机交互研制了学前儿童同伴交往的问题解决能力评估工具，应用评估工具考察了其发展特点，并采用人机交互实施了教育干预。主要研究结果如下：

首先，人机交互赋能学前儿童同伴交往的问题解决能力评估与干预。剖析人机交互下评估与干预的价值意蕴和实现方略是本研究的首要工作，为整项研究扎实学理性基础。研究分析了学前儿童同伴交往的问题解决能力评估与干预为何需要被革新、人机交互为何能够予以革新等内容，发现当前学前儿童在传统评估与干预范式中的“具身性”存在缺位与式微之嫌，而人机交互下的评估与干预则有助于回归和增进其“具身性”。研究认为，“身体涉入”“情境加持”和“关系互动”是激活学前儿童“具身性”、实现人机交互评估与干预的三大路径，为后续学前儿童同伴交往的问题解决能力评估与干预提供了可行的路向启发。

其次，人机交互下学前儿童同伴交往的问题解决能力评估工具。研究以教育指引、国际经验以及相关干预项目和评估工具为综合依据，初步构建了本土学前儿童同伴交往的问题解决能力评估指标。采用德尔菲专家咨询法，邀请了17位专家对初步构建的评估指标进行了两轮次修订，最终将同伴问题归因、行为结果预测、多元方法生成、行为策略决策确定为同伴交往的问题解决能力评估指标。在此基础上，基于人机交互研制了学前儿童同伴交往的问题解决能力评估工具，并邀请137名学前儿童参加预测试，通过项目分析、信度检验、效度检验等发现，该评估工具具有良好的区分度和信效度。

再者，人机交互评估下学前儿童同伴交往的问题解决能力发展特点。描述学前儿童的发展特点是教育干预实施前的重要环节。应用所研制的评估工具对314名学前儿童实施评估，结果发现：学前儿童同伴交往的问题解决能力整体存在年龄效应，但性别效应不显著；非独生子女发展优势明显。各维度年龄主效应显著，仅“行为策略决策”存在性别效应。此外，学前儿童同伴交往的问题解决能力面临一些困境：惯于将同伴问题的发生归因于他者，忽视对自身及其他客观要素的思考；偏向预测社交行为的外显性、即时性结果，对隐匿性、滞后性结果预测较少；关于同伴问题的解决方法以消极性、中性为主，积极性解决方法的掌握数量不多；关于解决同伴问题的策略决策建设性程度不高。总体看，当前学前儿童同伴交往的问题解决能力发展不容乐观，亟待实施教育干预予以改善和提升。

最后，人机交互下学前儿童同伴交往的问题解决能力干预。研究采用了社交机器人辅助教育干预的模式，设置了“教师使用机器人”干预组(n=27)、“教师单独实施”干预组(n=20)和控制组(n=23)，实施了为期8周的干预活动，并通过实验组控制组前后测检验发现，社交机器人对学前儿童同伴交往的问题解决能力整体影响成效显著，“教师使用机器人”干预组同伴交往的问题解决能力后测总分显著高于前测，且显著高于“教师单独实施”干预组和控制组的后测总分。在各维度上，“教师使用机器人”干预组的同伴问题归因、多元方法生成、行为策略决策的后测得分显著高于前测，且显著高于其他两组后测得分。但在行为结果预测上，其后测得分与“教师单独实施”干预组无差异。总之，相比教师单独实施干预，机器人辅助干预的模式在提升学前儿童同伴交往的问题解决能力上更具优势，对多数维度的促进更明显，但在行为结果预测的干预上未表现出特殊作用。

基于上述结果，研究对人机交互评估与干预的价值和方略、人机交互下同伴交往的问题解决能力评估与干预等进行了综合讨论，并提出三点建议：发挥评估作用，回应学前儿童同伴交往的问题解决能力发展需求；应用人机交互，提升学前儿童同伴交往的问题解决教育质量；以点带面，推进人机交互赋能学前儿童社会情感能力的评估与教育。此外，本研究总结了三个创新点，即探究了人机交互下学前儿童同伴交往的问题解决能力评估与干预的价值及方略，研制了人机交互评估工具并揭示了学前儿童同伴交往的问题解决能力发展特点，探询了人机交互对学前儿童同伴交往的问题解决能力的影响。未来仍需对人机交互下学前儿童同伴交往的问题解决能力评估与干预作深入研究。

Cultivating the peer interactional problem-solving skills in preschool children is an inherent need to promote their social and emotional learning. The frequent occurrence of peer interactional problems among preschool children and the constant occurrence of peer bullying on campus all demonstrate the necessity and key of cultivating the peer interactional problem-solving skills in the early stages. Assessment and intervention are the two major hubs for promoting the development the peer interactional problem-solving skills in preschool children. However, traditional paradigms of assessment and intervention have limitations, making it difficult to achieve precise assessment and efficient intervention for preschool children. Promoting digitization is the main theme of current child development assessment and educational reform. As a typical means of achieving digitalization, human-machine interaction has become a new paradigm for the assessment and intervening in preschool children due to its advantages of balancing interactivity, fun, and convenient application. Therefore, this study focuses on the assessment and intervention of peer interactional problem-solving skills in preschool children, selecting "human-machine interaction" as the technical support and implementation method for assessment and intervention, with the theories of Interpersonal Cognitive Problem-Solving, and Embodied Cognition as the theoretical foundations, use the study methods such as Delphi, Situational Interviews, Questionnaire Surveys, and Educational Experiments. This study analyzed the value and strategy of the assessment and intervention of peer interactional problem-solving skills in preschool children assisted by human-machine interaction, developed the assessment tool of the peer interactional problem-solving skills in preschool children assisted by human-machine interaction, applied the tool to examine the development characteristics and shortcomings of the peer interactional problem-solving skills in preschool children, implemented the education intervention assisted by human-machine interaction. The main research findings are as follows:

The value and strategy of assessment and intervention of the peer interactional problem-solving skills in preschool children assisted by human-machine interaction. The primary task of this study is to systematically clarify the value of assessing and intervening in the peer interactional problem-solving skills of preschool children under the human-computer interaction, which lays a solid academic and rational foundation for the entire study. Firstly, an overall analysis was conducted on why the assessment and intervention of the peer interactional problem-solving skills need to be innovated, and why human-machine interaction technology can be innovated. It was found that the "embodiment" of preschool children in the traditional assessment and intervention paradigm is missing and declining, while the assessment and intervention under human-machine interaction can help to return and enhance the "embodiment". Meanwhile, this study suggests that "physical involvement", "situational reinforcement", and "relational interaction" are the three major pathways to activate the "embodiment" of preschool children and achieve the assessment and intervention assisted by human-machine interaction, providing specific guidance for the assessment and intervention of the peer interactional problem-solving skills in preschool children.

The assessment tool for the peer interactional problem-solving skills in preschool children assisted by human-machine interaction. It initially constructed the assessment indicators of local preschool children's peer interaction problem-solving skills based on the educational policy, international experience, relevant intervention projects and assessment tools. Using the Delphi expert consultation method, inviting 17 experts to form a consulting group, two rounds of revisions were made to the initially constructed assessment indicators. The assessment indicators for the peer interactional problem-solving skills of preschool children were determined as peer problem attribution, behavioral outcome prediction, multivariate method generation, and behavioral strategy decision-making. And then, the assessment tool were developed assisted by human-machine interaction. 137 preschool children were invited to participate in a pre-test. Through project analysis, reliability testing, validity testing, etc., it was found that the assessment tool has good discrimination and reliability validity.

The developmental characteristics and shortcomings of the peer interactional problem-solving skills in preschool children. Describing the developmental profile of preschool children is an important step before implementing the educational interventions. The assessment tool were conducted on 314 children aged 3-6. The results showed that there was an age effect and an "only child" effect of the peer interactional problem-solving skills, but the gender effect was not significant. The main effect of age is significant in all dimensions, gender effects are only presented in the behavioral strategy decision-making. In addition, preschool children were accustomed to attributing peer problems to others, neglecting thinking about oneself and other objective factors. They are biased towards predicting explicit and immediate results of social behavior. And limited mastery of proactive strategies for solving peer problems. The constructive level of behavioral strategy decision-making urgently needs to be improved. Overall, the development of peer interactional problem-solving skills among preschool children is not optimistic, and there is an urgent need to implement educational interventions.

The educational intervention on the peer interactional problem-solving skills among preschool children assisted by human-machine interaction. To overcome the shortcomings of the past educational interventions, the pattern of social robot assisted intervention was adopted to enhance children's peer interactional  problem-solving skills. Based on the educational experimental method, a social robot assisted intervention group (n=27), a teacher alone intervention group (n=20), and a control group (n=23) were set up, an 8-week intervention activity was implemented, through the experimental design of the pre- and post-tests, we had found that the overall impact of social robot assisted dialogic reading on children's skill to solve peer interactional problems is significant. The total score of the robot group in the post-test is significantly higher than that in the pre-test, and significantly higher than the total score of the teacher group and the control group in the post-test. In terms of local impact, the post-test scores of peer problem attribution, multivariate method generation, behavioral strategy decision-making, and behavioral outcome prediction in the robot group were significantly higher than those in the pre-test. Among them, the first three dimensions were significantly higher than the post-test scores of the teacher group and the control group. However, there was no significant difference between the robot group and the teacher group in predicting behavioral outcomes in the post-test scores. Overall, compared to the teaching by teachers, social robot assisted interventions has an advantage in improving the overall level of peer interactional problem-solving skills, and have a more significant promoting effect on most dimensions of peer interactional problem-solving skills. However, it has not shown a special effect on the behavior outcome prediction skill.

Based on the above results, the value and the practice of human-machine interaction assessment and intervention on the peer interactional problem-solving skills in preschool children were discussed and analyzed, and three suggestions were proposed: playing the role of assessment to respond to the development needs of the peer interactional problem-solving skills in preschool children; utilizing the advantages of human-machine interaction intervention to enhance the quality of the education for preschool children’s peer interactional problem-solving skills; promoting the innovation of assessment and education for preschool children’s social emotional ability assisted by human-machine interaction. In addition, three innovative points of this study were identified: the value and strategy of the assessment and intervention of the peer interactional problem-solving skills in preschool children assisted by human-machine interaction were analyzed, the characteristics of the peer interactional problem-solving skills in preschool children were revealed based on the human-machine interaction assessment, the impact of human-machine interaction on peer interactional problem-solving skills in preschool children was explored. Further in-depth research is needed in the future on the assessment and intervention of peer interactional problem-solving skills among preschool children assisted by human-computer interaction.