2024_06_22_cac2e2deba2d1188427dg

Human-in-the-Loop Simulations Within Extended Reality for Preservice
用于职前教育的扩展现实中的人环模拟

Teacher Training: A Meta-Analysis and Instructional Design Guidelines
教师培训：元分析与教学设计指南

Abstract 摘要

Human-in-the-loop simulations (HITLS) within extended reality (XR) for preservice teacher training (PTT) have gained increasing attention. However, the impact of such simulations remains unclear and cannot be generalized from past generic HITLS research. Therefore, this study aimed to evaluate the effectiveness of applying HITLS within XR for PTT. Moreover, based on the NLN Jeffries simulation theory, this study systematically explored potential instructional design features across the planning, implementation, and evaluation phases of HITLS within XR for PTT. Data for this study were gathered from the Web of Science, Scopus, and Google Scholar. A comprehensive meta-analysis was conducted on findings extracted from 142 studies in 39 articles. These studies collectively involved 1564 participants and spanned the timeframe from 2013 to 2024. The results indicated that HITLS within XR had a moderate effect on PTT outcomes, with an effect size of 0.535. Regarding potentially effective instructional features, five features were identified during the planning phase: setting, graphical realism, autonomy, task variation, and simulation length. During the implementation phase, six features were identified: introduction, theoretical preparation, rehearsal, prompt, static feedback, and other scaffolds. No significant features were found during the evaluation phase. Based on these findings, this study further provided instructional design guidelines for HITLS within XR for PTT.
用于职前教师培训（PTT）的扩展现实（XR）中的人在环路模拟（HITLS）越来越受到关注。然而，这种模拟的影响仍不明确，也不能从过去的通用 HITLS 研究中归纳出来。因此，本研究旨在评估在 XR 中将 HITLS 应用于 PTT 的效果。此外，基于 NLN Jeffries 模拟理论，本研究系统地探讨了 PTT XR 中 HITLS 在计划、实施和评估阶段的潜在教学设计特点。本研究的数据来自 Web of Science、Scopus 和 Google Scholar。我们对 39 篇文章中的 142 项研究结果进行了综合荟萃分析。这些研究共涉及 1564 名参与者，时间跨度为 2013 年至 2024 年。结果表明，XR 中的 HITLS 对 PTT 结果的影响适中，效应大小为 0.535。关于潜在有效的教学特点，在计划阶段确定了五个特点：设置、图形逼真度、自主性、任务变化和模拟长度。在实施阶段，确定了六个特征：导入、理论准备、演练、提示、静态反馈和其他支架。在评估阶段没有发现重要的特征。基于这些发现，本研究进一步为 PTT XR 中的 HITLS 提供了教学设计指南。

Keywords: Human-in-the-loop simulation, Meta-analysis, Instructional design, The NLN Jeffries simulation theory
关键词人在回路中模拟元分析教学设计 NLN 杰弗里斯模拟理论

1. Introduction 1.导言

Due to the complexity and unpredictability nature of teaching, where teachers engage with students possessing diverse personalities and educational needs, preservice teachers (PT) need to gain exposure to various situations and student demographics during their teacher education, to adequately prepare for their future profession (Goldstein, 2005). Field-based experience is widely valuable in preservice teacher training (PTT), offering PTs opportunities for handling complex pedagogical situations, yet it encounters challenges including PTs' cognitive readiness, the need for more personalized guidance, and limitations in the breadth of teaching practice (Goodman, 1986; Haverly & Davis, 2024). In an effort to address the mentioned concerns yet preserve the benefits of traditional field experiences, virtual simulations emerge as a promising alternative and a valuable complement to authentic classroom experiences (Cambourne et al., 2003; Levin & Flavian, 2022; Lindberg & Jönsson, 2023). Virtual simulations have gained significant attention in medicine and health sciences (Khan et al., 2018; Yeo & Jang, 2023), and are progressively receiving recognition in PTT as well (Kaufman & Ireland, 2016). Virtual simulations can offer PTs unlimited opportunities to engage with various classroom scenarios without risking harm to real students, and involve PTs in the complex decision-making processes that in-service teachers make on a daily basis (Bradley & Kendall, 2014; Spencer et al., 2019).
由于教学工作的复杂性和不可预测性，教师要与具有不同个性和教育需求的学生打交道，因此，职前教师（PT）需要在师范教育期间接触各种情况和学生群体，为未来的职业做好充分准备（Goldstein，2005）。实地教学经验在职前教师培训（PTT）中具有广泛的价值，它为职前教师提供了处理复杂教学情境的机会，但它也遇到了一些挑战，包括职前教师的认知准备程度、更多个性化指导的需求以及教学实践广度的限制（Goodman, 1986; Haverly & Davis, 2024）。为了解决上述问题，同时保留传统现场经验的优点，虚拟模拟成为一种有前途的替代方法，也是对真实课堂经验的宝贵补充（Cambourne 等，2003；Levin & Flavian，2022；Lindberg & Jönsson，2023）。虚拟仿真在医学和健康科学领域获得了极大关注（Khan 等人，2018 年；Yeo & Jang，2023 年），在 PTT 领域也逐渐得到认可（Kaufman & Ireland，2016 年）。虚拟模拟可以为教师提供无限的机会，让他们参与到各种课堂场景中，而不必冒着伤害真实学生的风险，并让教师参与到在职教师每天所做的复杂决策过程中（Bradley & Kendall, 2014; Spencer et al.）

There is a long history of attempts to simulate classroom situations to complement field experience. Early simulations incorporated printed materials and role plays conducted by either students or tutors (McGarr, 2021), as well as manually controlled film sequences featuring non-interactive animated content comprising sound, text, or images (Dalgarno et al., 2016). With technological advancements, systems for PTT evolved with improved user interface design and sophisticated algorithms, gradually increasing interactivity. These advancements have
模拟课堂情境以补充实地经验的尝试由来已久。早期的模拟包括由学生或辅导员进行的印刷材料和角色扮演（McGarr，2021 年），以及手动控制的电影序列，其特点是由声音、文本或图像组成的非交互式动画内容（Dalgarno 等人，2016 年）。随着技术的进步，PTT 系统在改进用户界面设计和复杂算法的基础上不断发展，交互性逐渐增强。这些进步
progressed to the point where extended reality (XR) technologies, such as virtual reality (VR), mixed reality (MR), and augmented reality (AR), contribute to virtual simulations with more sophisticated graphic design and multimodal interactive features, bolstering authenticity, engagement, and immediacy (Stanney et al., 2023). For PTT, a pure XR environment lacking virtual human interaction fails to simulate real classroom scenarios, restricts interactivity, and consequently limits PTs' development of teaching skills such as interpersonal competency. Therefore, Human-in-the-Loop Simulations (HITLS) within XR for PTT are gaining attention.
随着虚拟现实（VR）、混合现实（MR）和增强现实（AR）等扩展现实（XR）技术的发展，虚拟仿真的图形设计和多模态交互功能更加复杂，增强了真实性、参与性和即时性（Stanney 等人，2023 年）。对于 PTT 而言，缺乏虚拟人类互动的纯粹 XR 环境无法模拟真实的课堂场景，限制了互动性，从而限制了 PT 的教学技能（如人际交往能力）的发展。因此，针对 PTT 的 XR 中的 "人在环中模拟"（HITLS）越来越受到关注。

In HITLS, there is a pronounced emphasis on the human interacting with virtual humans, and the overall validity and utility of HITLS largely depend on human behaviors. Virtual humans are typically driven by either avatar technology (virtual humans controlled by both humans and computers) (Ely et al., 2018; Hirsch et al., 2023) or agent technology (virtual humans controlled solely by computers) (Nussli & Oh, 2014; DeSantis et al., 2023). Virtual humans adapt and complete interaction activities in response to human behaviors. This interactive and adaptive nature makes HITLS well-suited for addressing large and semi-structured problems characterized by unpredictability, where human involvement plays a crucial role (Rothrock & Narayanan, 2011). Teaching activities in classroom settings can be complex and unpredictable, presenting significant challenges for preservice teachers. In this regard, HITLS proves highly beneficial for PTT. HITLS enables PTs to engage with simulated classroom environments and interact with virtual humans depicting different roles. PTs can practice teaching until educational goals are met, gaining hands-on experience across various teaching scenarios in a safe and controlled environment.
在 HITLS 中，明显强调人与虚拟人的互动，HITLS 的整体有效性和实用性在很大程度上取决于人的行为。虚拟人通常由虚拟化身技术（由人类和计算机共同控制的虚拟人）（Ely 等人，2018 年；Hirsch 等人，2023 年）或代理技术（仅由计算机控制的虚拟人）（Nussli 和 Oh，2014 年；DeSantis 等人，2023 年）驱动。虚拟人根据人类的行为进行调整并完成交互活动。这种互动和自适应的特性使 HITLS 非常适合解决大型和半结构化的问题，这些问题的特点是不可预知，人的参与在其中发挥着至关重要的作用（Rothrock 和 Narayanan，2011 年）。课堂教学活动可能既复杂又难以预测，这给职前教师带来了巨大挑战。在这方面，HITLS 证明对 PTT 大有裨益。HITLS 使教师能够参与模拟课堂环境，并与扮演不同角色的虚拟人进行互动。实习教师可以进行教学练习，直到达到教学目标，从而在安全可控的环境中获得各种教学场景的实践经验。

HITLS, especially when deployed within XR environments, is capable of enhancing much more perceived realism and presence, and has drawn increasing attention in research studies. However, empirical research has produced mixed findings, with some studies supporting its effectiveness while others show opposing results. Therefore, a comprehensive meta-analysis is essential to systematically examine the effects of HITLS within XR for PTT.
HITLS 尤其是在 XR 环境中部署时，能够增强更多的感知真实感和临场感，在研究中引起了越来越多的关注。然而，实证研究的结果喜忧参半，一些研究支持其有效性，而另一些研究则显示出相反的结果。因此，有必要进行全面的荟萃分析，以系统地研究 HITLS 在 XR 中对 PTT 的影响。

Moreover, the mere of use of cutting-edge technology does not inherently ensure an improved teaching or learning process if the instructional design is not sound enough (Clark, 1994; Tam, 2000). In the context of HITLS within XR for PTT, implementing HITLS without considering its educational implications may lead to issues such as increased mental load, misconceptions, nonsensical practices, and conflicts of interest. Therefore, the significance of instructional designs cannot be overstated, as evidenced by a growing body of research (Cohen et al., 2020; Lu et al., 2023) in this field. Nevertheless, the instructional designs in these studies often appeared as isolated features, involving the singular application of instructional features such as orientation, prompt, debrief activities, or various combinations of these features. Few studies systematically considered these features with robust pedagogical rationale support that guides which features should be prioritized (Reiser, 2001). Additionally, there is a lack of validation for the effectiveness of instructional designs. Consequently, the credibility of these results is called into doubt, making it challenging to replicate in future applications of HITLS within XR for PTT.
此外，如果教学设计不够合理，仅仅使用尖端技术并不能从本质上确保改进教学或学习过程（Clark，1994；Tam，2000）。就 PTT XR 中的 HITLS 而言，在不考虑其教育意义的情况下实施 HITLS 可能会导致诸如增加心理负担、误解、无意义的实践和利益冲突等问题。因此，教学设计的重要性怎么强调都不为过，该领域越来越多的研究（Cohen 等人，2020 年；Lu 等人，2023 年）证明了这一点。尽管如此，这些研究中的教学设计往往表现为孤立的特征，涉及教学特征的单一应用，如定向、提示、汇报活动，或这些特征的各种组合。很少有研究系统地考虑了这些特点，并提供了强有力的教学理论支持，以指导哪些特点应优先考虑（Reiser，2001 年）。此外，教学设计的有效性也缺乏验证。因此，这些结果的可信度值得怀疑，这使得在未来的 PTT XR 中应用 HITLS 具有挑战性。

Therefore, the present meta-analysis held a twofold purpose: firstly, to validate the effectiveness of HITLS within XR for PTT, and secondly, to systematically explore instructional design features worth considering at each phase of HITLS within XR-planning, implementation, and evaluation-guided by the NLN Jeffries simulation theory (Jeffries, 2021). These features were then coded as moderator variables to explore potentially effective ones at each phase, aiding in the decision-making processes of design, implementation, and evaluation of HITLS within XR for PTT. Such an analysis aimed to promote the use of HITLS within XR for PTT and provide
因此，本荟萃分析具有双重目的：首先，验证针对 PTT 的 XR 中 HITLS 的有效性；其次，在 NLN 杰弗里斯模拟理论（Jeffries，2021 年）的指导下，系统地探索 XR 中 HITLS 的每个阶段（规划、实施和评估）值得考虑的教学设计特征。然后将这些特征作为调节变量进行编码，以探索每个阶段潜在的有效特征，从而帮助 PTT 在 XR 中设计、实施和评估 HITLS 的决策过程。这种分析旨在促进在 PTT XR 中使用 HITLS，并提供
valuable insights into effective instructional designs for its implementation. The following research questions were posed:
为实施有效的教学设计提供有价值的见解。我们提出了以下研究问题：

RQ1: What is the impact of HITLS within XR on PTT outcomes?
问题 1：XR 中的 HITLS 对 PTT 结果有何影响？

RQ1:Which instructional design features demonstrate potential effectiveness during the planning, implementation, and evaluation phase of HITLS within XR for PTT, respectively?
问题 1：在 PTT XR 中的 HITLS 计划、实施和评估阶段，哪些教学设计特点分别显示出潜在的有效性？

2. Literature review 2.文献综述

2.1. Current HITLS systems within XD for PTT
2.1.目前在 XD 内用于 PTT 的 HITLS 系统

Currently, some HITLS systems within XD for PTs are available, falling into two main categories: avatar-mediated interactive training, and agent-mediated interactive training. Avatars and agents in HITLS both represent virtual humans but differ in terms of control (Mulvaney et al., 2024). Avatar-mediated training is computer- and human-based, such as Mursion, which blends digital puppetry with AI technology in a "human-in-the-loop" approach. Avatars display basic facial and gesture expressions powered by AI, while real human actors guide real-time verbal conversations (Nagendran et al., 2013). Agent-mediated training, exemplified by systems like Evelyn, is solely computer-based, with virtual agents driven by large language models communicating in natural language and simulating various student characteristics, without human intervention (Dai, 2023). For a deeper understanding of current HITLS systems within XR for PTT, this study analyzed several typical HITLS systems within XR from several perspectives, including XR environment, system interactivity, environment fidelity, system interactivity, behavioral fidelity, avatar/agent activity, and avatar/agent tasks (see Appendix A).
目前，XD 中的一些 HITLS 系统可供 PT 使用，主要分为两类：以虚拟人为媒介的交互式训练和以代理为媒介的交互式训练。HITLS 中的阿凡达和代理都代表虚拟人，但在控制方面有所不同（Mulvaney 等人，2024 年）。以阿凡达为媒介的培训以计算机和人类为基础，如 Mursion，它以 "人在回路中 "的方式将数字木偶与人工智能技术相结合。阿凡达显示由人工智能驱动的基本面部和手势表情，而真人演员则引导实时语言对话（Nagendran 等人，2013 年）。以伊夫林（Evelyn）等系统为例，以代理为中介的培训完全以计算机为基础，虚拟代理由大型语言模型驱动，以自然语言进行交流，并模拟各种学生特征，无需人工干预（Dai，2023 年）。为了更深入地了解目前 XR 中用于 PTT 的 HITLS 系统，本研究从多个角度分析了 XR 中的几个典型 HITLS 系统，包括 XR 环境、系统交互性、环境保真度、系统交互性、行为保真度、虚拟化身/代理活动以及虚拟化身/代理任务（见附录 A）。

XR environment typically includes AR, VR, and MR. AR overlays virtual elements onto the real world, while VR normally immerses users in a completely virtual environment using a headset. MR combines elements of both, integrating virtual and real-world aspects (Suryodiningrat et al., 2023). Preferences for supported hardware vary across XR environments, influencing the level of immersion and learner satisfaction with the HITLS experience (Hudson et al., 2019). In current HITLS systems within XR for PTT, avatar-mediated interactive training usually occurs in MR environments with semi-immersion (using electronic large screens), while agent-mediated interactive training is commonly found in VR environments, offering both fully immersive (using headsets) and non-immersive experiences (on small screens like computers or mobile devices). AR environments also host agent-mediated interactive training, mostly in a non-immersive manner, although there are fewer related applications.
XR 环境通常包括 AR、VR 和 MR。AR 将虚拟元素叠加到现实世界中，而 VR 通常使用头戴式设备让用户沉浸在完全虚拟的环境中。MR 结合了两者的元素，整合了虚拟和现实世界的各个方面（Suryodiningrat 等人，2023 年）。不同的 XR 环境对支持硬件的偏好各不相同，这影响了 HITLS 体验的沉浸程度和学习者的满意度（Hudson 等人，2019 年）。在当前用于 PTT 的 XR HITLS 系统中，以虚拟人为媒介的交互式培训通常发生在半沉浸式（使用电子大屏幕）的 MR 环境中，而以代理为媒介的交互式培训通常发生在 VR 环境中，提供完全沉浸式（使用耳机）和非沉浸式体验（在电脑或移动设备等小屏幕上）。AR 环境也提供以代理为中介的互动培训，主要是非沉浸式培训，但相关应用较少。

System interactivity is another crucial aspect of XR environments, as it has the potential to enhance learners' sense of presence (Chang, 2017). When examining the interactivity of HITLS, two primary components should be taken into account: technological requirement and interaction type. Avatar-mediated interactive training systems often involve PTs interacting with avatars via large screens, requiring substantial screen space, and facilitating natural interactions such as voice and gestures. For agent-mediated interactive training systems, in non-immersive environments, PTs interact with agents using smaller screens like smartphones and desktops. While in fully immersive environments, PTs typically wear Head-Mounted Displays (HMDs) and use controllers to interact with agents. Interaction types in agent-mediated interactive training systems include haptic interactions like menu choices or natural interactions like voice and gestures, varying based on the system's level
系统交互性是 XR 环境的另一个重要方面，因为它有可能增强学习者的临场感（Chang，2017）。在研究 HITLS 的互动性时，应考虑两个主要因素：技术要求和互动类型。以头像为媒介的交互式培训系统通常涉及参与者通过大屏幕与头像进行交互，需要大量的屏幕空间，并促进语音和手势等自然交互。对于以代理为中介的交互式培训系统，在非沉浸式环境中，学员使用较小的屏幕（如智能手机和台式机）与代理互动。而在完全沉浸式环境中，参与者通常佩戴头戴式显示器（HMD）并使用控制器与代理进行交互。以代理为中介的交互式培训系统中的交互类型包括菜单选择等触觉交互或语音和手势等自然交互，根据系统的级别而有所不同。
of intelligence. Regarding environment fidelity, it refers to the degree of realism in the virtual environment, including aspects such as the graphical realism of the environment and virtual humans, as well as virtual humans' representation in terms of numbers and status. The current HITLS systems within XD for PTT often prioritize higher graphical realism for both the environment and virtual humans, although they may not achieve photo-real or human-like levels. However, debate exists regarding whether high graphical fidelity enhances or impedes learning (Wages et al., 2004; Bossard et al., 2008). Therefore, further exploration is needed to investigate whether graphical realism of HITLS has an impact on PTT. Furthermore, in current HITLS systems, virtual humans often portray students, parents, and other roles. Avatar-mediated systems usually feature 1 to 8 avatars, while agent-mediated systems may include over 25 agents, better reflecting real classroom sizes and potentially enhancing environment fidelity.
智能。关于环境保真度，它指的是虚拟环境的逼真程度，包括环境和虚拟人的图形逼真度，以及虚拟人在数字和状态方面的表现。目前用于 PTT 的 XD 中的 HITLS 系统通常优先考虑环境和虚拟人的较高图形逼真度，尽管它们可能无法达到照片般逼真或类似人类的水平。然而，关于高图形逼真度是会增强还是会阻碍学习的争论依然存在（Wages 等人，2004 年；Bossard 等人，2008 年）。因此，需要进一步探讨 HITLS 的图形逼真度是否会对 PTT 产生影响。此外，在当前的 HITLS 系统中，虚拟人通常扮演学生、家长和其他角色。以虚拟人为媒介的系统通常有 1 到 8 个虚拟人，而以代理为媒介的系统可能包括超过 25 个代理，从而更好地反映真实教室的规模，并有可能提高环境的逼真度。

Behavioral fidelity in HITLS refers to the realism and consistency of virtual human behaviors. When combined with graphical realism, it contributes to believability, fostering interaction between learners and virtual humans (Groom et al., 2009). Enhancing virtual humans' human-like appearance involves two main considerations: Behavior realism assesses if the avatar acts autonomously like a real human, while affect realism encompasses emotional aspects, including the avatar's feelings, thoughts, and non-verbal expressions like facial and gesture expressions. Delamarre et al. (2021) identified three main ways of generating virtual humans' behaviors and expressions: Scenario-based, where behaviors and expressions are pre-programmed and computer-controlled; Wizard of

-based, where a human actor controls the behaviors based on predefined instructions; and Model-based, which uses computational models or large language models to simulate student behaviors and expressions. Typical avatar-mediated interactive systems primarily employ the WOz-based method. While agents in typical agent-mediated interactive systems are either model-based or scenario-based, depending on the system's intelligence level. Several theories offer predictions regarding how different levels of behavioral fidelity of virtual humans affect human responses. The Maximization Theory indicates that increased realism improves interactions (Jafar et al., 2010), while the Uncanny Valley Theory warns that too much realism may lead to discomfort (Johnston et al., 2020). Therefore, further research is needed on how this affects PTT in XR simulations.
HITLS 中的行为逼真度是指虚拟人行为的真实性和一致性。当与图形逼真度相结合时，它有助于提高可信度，促进学习者与虚拟人之间的互动（Groom 等人，2009 年）。增强虚拟人的类人外观涉及两个主要考虑因素：行为逼真度评估虚拟人是否像真人一样自主行动，而情感逼真度则包括情感方面，包括虚拟人的情感、思想以及面部和手势等非语言表达。Delamarre 等人（2021 年）指出了生成虚拟人行为和表情的三种主要方式：基于场景，即行为和表情由预先编程并由计算机控制；基于精灵

，即人类演员根据预先定义的指令控制行为；以及基于模型，即使用计算模型或大型语言模型模拟学生的行为和表情。典型的以虚拟化身为中介的交互系统主要采用基于 WOz 的方法。而典型的以代理为中介的互动系统中的代理则是基于模型或基于情景的，这取决于系统的智能水平。有几种理论可以预测虚拟人不同程度的行为逼真度会如何影响人类的反应。最大化理论（Maximization Theory）指出，提高逼真度可以改善交互（Jafar 等人，2010 年），而 "不真实谷理论"（Uncanny Valley Theory）则警告说，过于逼真可能会导致不适（Johnston 等人，2020 年）。因此，需要进一步研究这对 XR 模拟中的 PTT 有何影响。

During HITLS, virtual humans engage in three types of activities: (1) responsive activity, such as answering questions in real-time, (2) goal-oriented activity, such as simulating student misbehavior in class, and (3) internal activity, such as retaining memory. The diversity and depth of activities in which virtual humans can engage also reflect their degree of autonomy, which in turn is related to their behavioral fidelity, impacting the PTT experience. Currently, in avatar-mediated interactive training systems, avatars primarily participate in responsive and goal-oriented activities facilitated by human actors who can interact with PTs in real-time and pre-plan behavior according to PT training requirements. In agent-mediated interactive training systems, agents engaged in scenario-based behavior primarily concentrate on goal-oriented activities and lack the ability to adapt behavior based on PT performance. While, agents equipped with advanced AI technology can engage in interactive, goal-oriented, and internal activities, closely resembling human-like behavior.
在 HITLS 期间，虚拟人从事三种类型的活动：(1) 反应活动，如实时回答问题；(2) 目标导向活动，如模拟学生在课堂上的不当行为；(3) 内部活动，如保持记忆。虚拟人可以参与的活动的多样性和深度也反映了他们的自主程度，这反过来又与他们的行为保真度有关，影响着 PTT 体验。目前，在以虚拟人为媒介的交互式培训系统中，虚拟人主要参与反应灵敏和以目标为导向的活动，这些活动由能与 PT 进行实时互动并根据 PT 培训要求预先计划行为的人类行为者提供便利。在以代理为中介的交互式培训系统中，参与基于场景的行为的代理主要集中于目标导向型活动，缺乏根据 PT 的表现调整行为的能力。而配备了先进人工智能技术的代理则可以参与互动、目标导向和内部活动，与人类行为非常相似。

In conclusion, the analysis of current HITLS systems within XD for PTT provides insights into into their functionality and effectiveness. These systems are utilized across diverse XR environments, typically featuring immersive 3D graphics. Virtual humans within these systems commonly replicate students, with variations in quantity between avatar-mediated and agent-mediated systems. Furthermore, agents exhibit higher behavioral fidelity with model-based behavior driven by advanced AI technologies,, allowing for a wider range of activities and training possibilities than avatars. Nonetheless, less sophisticated agent-mediated driven by scenario-based
总之，通过对当前用于 PTT 的 XD 中的 HITLS 系统进行分析，我们可以深入了解其功能和有效性。这些系统在不同的 XR 环境中使用，通常采用身临其境的 3D 图形。这些系统中的虚拟人通常是学生的复制品，但以虚拟人为媒介的系统和以代理为媒介的系统在数量上存在差异。此外，代理通过先进的人工智能技术驱动基于模型的行为，表现出更高的行为保真度，与虚拟化身相比，可以开展更广泛的活动和培训。尽管如此，由基于情景的人工智能技术驱动的、不太复杂的代理中介系统也能提供更多的活动和培训可能性。
behavior and avatar-mediated systems have also shown positive aspects of supporting PTT. Given the variations among these systems, it's essential to explore which kind of HITLS systems within XR supports better PTT outcomes, so to provide a reference for their application. Additionally, investigating how to design instructional strategies tailored to HITLS systems within XR for PTT is also a worthwhile endeavor.
行为和化身中介系统也显示出支持 PTT 的积极方面。鉴于这些系统之间的差异，有必要探索 XR 中哪种 HITLS 系统能支持更好的 PTT 效果，从而为它们的应用提供参考。此外，研究如何针对 XR 中的 HITLS 系统设计针对 PTT 的教学策略也是一项值得努力的工作。

2.2. Previous reviews and meta-analysis
2.2.以往的综述和荟萃分析

Researchers have increasingly explored the impact of HITLS within XR for PTT. A screening of the Web of Science database revealed that meta-analyses of HITLS within XR were frequently conducted in the health sciences, yet there were relatively few meta-analysis studies focusing on PTT. However, several related literature reviews provided valuable insights, as detailed in Appendix B.
研究人员越来越多地探讨了 XR 中的 HITLS 对 PTT 的影响。通过筛选 Web of Science 数据库，我们发现在健康科学领域经常会对 XR 中的 HITLS 进行荟萃分析，但针对 PTT 的荟萃分析研究相对较少。不过，一些相关的文献综述提供了有价值的见解，详见附录 B。

To our knowledge, few studies have specifically focused on XR simulations for PTT education, while a greater number have addressed teacher education more broadly, including both in-service and preservice teacher training (Bradley & Kendall, 2014; Nussli & Oh, 2014). The majority of review studies focused on MR simulations in teacher education (Dieker et al., 2023; Bondie et al., 2021), followed by XR simulations (Theelen et al., 2019), VR simulations (Huang et al., 2023), and AR simulations (Mena et al., 2023). Interestingly, only one review study found examined AR simulations in teacher education and revealed that the effectiveness was largely theoretical due to a lack of extensive research. This underscored the limited adoption of AR in teacher education; only two review studies considered HITLS (Bondie et al., 2021; Ade - Ojo., 2022), but were both within MR environments, focusing only on avatar-mediated interactive training. Regarding the effectiveness of XR simulations for teacher education, nearly all review studies reported positive findings. However, it is important to note that limitations were also highlighted in many review studies, including technical issues, inadequate reliability, and insufficient instructional support (Sánchez-Caballé et al., 2020; Lindberg & Jönsson, 2023). Concerning instructional designs, half of the current review studies did not address this aspect. Among those that did, most simply mapped isolated instructional design features without theoretical guidance and lacked empirical validation (Ade - Ojo et al., 2022; Sargent, 2020).
据我们所知，很少有研究专门关注 XR 模拟在 PTT 教育中的应用，而更多的研究则更广泛地涉及教师教育，包括在职和职前教师培训（Bradley & Kendall, 2014; Nussli & Oh, 2014）。大多数综述研究侧重于教师教育中的磁共振模拟（Dieker 等人，2023 年；Bondie 等人，2021 年），其次是 XR 模拟（Theelen 等人，2019 年）、VR 模拟（Huang 等人，2023 年）和 AR 模拟（Mena 等人，2023 年）。有趣的是，只有一项综述研究对教师教育中的 AR 模拟进行了研究，结果表明，由于缺乏广泛的研究，其有效性主要停留在理论层面。这凸显了 AR 在师范教育中的应用有限；只有两项综述研究考虑了 HITLS（Bondie 等人，2021 年；Ade - Ojo.，2022 年），但这两项研究都是在 MR 环境中进行的，仅侧重于以虚拟化身为中介的互动培训。关于 XR 模拟对教师教育的有效性，几乎所有的综述研究都报告了积极的结论。然而，值得注意的是，许多综述研究也强调了其局限性，包括技术问题、可靠性不足以及教学支持不足（Sánchez-Caballé 等人，2020；Lindberg & Jönsson，2023）。关于教学设计，目前的审查研究中有一半没有涉及这方面的问题。在有涉及的研究中，大多数只是简单地描绘了一些孤立的教学设计特点，没有理论指导，也缺乏经验验证（Ade - Ojo et al.）

Although previous review studies generally supported XR simulations for teacher training and concluded some instructional design features, their findings should be approached cautiously due to several reasons. First, some of these studies only considered the effect of a specific type of virtual simulation, lacking comprehensive comparisons of XR environments and without sufficient considerations on HITLS for PTT. Second, the focus was primarily directed toward teacher education, with relatively less attention to specific PT education. Third, the instructional design suggestions in these review studies merely mapped some features with no robust theoretical support and effectiveness validation. Therefore, there has been a notable absence of review studies specifically focused on PTT, which investigate the effectiveness of HITLS within XR environments and systematically identify effective instructional designs with statistical validation.
尽管以往的回顾性研究普遍支持将 XR 模拟用于教师培训，并总结了一些教学设计特点，但由于以下几个原因，应谨慎对待这些研究结果。首先，其中一些研究只考虑了特定类型虚拟仿真的效果，缺乏对 XR 环境的全面比较，也没有充分考虑 PTT 的 HITLS。其次，研究重点主要针对师范教育，对具体的 PT 教育关注相对较少。第三，这些综述研究中的教学设计建议只是映射了一些特征，没有强有力的理论支持和效果验证。因此，专门针对 PTT 的回顾性研究明显缺乏，这些研究调查了在 XR 环境中 HITLS 的有效性，并系统地确定了有效的教学设计，并进行了统计验证。

3. Method 3.方法

To address the research gaps identified above, a meta-analysis was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). PRISMA, a set of consensus-based reporting guidelines, ensures transparent and objective reporting of meta-analyses. While initially developed for healthcare interventions, the PRISMA checklist has found extensive use in educational research (Wilson et al., 2020; Chen et al., 2022; Chen et al., 2024; Wang et al., 2024). This section provides an outline of the search strategy, inclusion/exclusion criteria, coding, and analysis procedures.
为了填补上述研究空白，我们按照《系统综述和元分析首选报告项目》（Preferred Reporting Items for Systematic Reviews and Meta-Analyses，PRISMA）指南（Page et al.）PRISMA 是一套基于共识的报告指南，可确保荟萃分析报告的透明性和客观性。虽然 PRISMA 检查表最初是针对医疗保健干预措施开发的，但在教育研究中也得到了广泛应用（Wilson 等人，2020 年；Chen 等人，2022 年；Chen 等人，2024 年；Wang 等人，2024 年）。本节概述了检索策略、纳入/排除标准、编码和分析程序。

3.1. Literature search strategy
3.1.文献检索策略

Studies relevant to HITLS within XR for PTT were gathered from the Web of Science, Scopus, and Google Scholar. Integrating these three databases ensures sufficient stability, as concluded by some studies (Harzing & Alakangas, 2016; Martin-Martin et al., 2018). Search terms were identified through database thesauruses and keywords found in sample papers, and two keyword patterns were utilized: (1) keywords related to preservice teacher: preteacher, preservice teacher, teacher candidate, teacher trainee, novice teacher, student teacher; (2) keywords related to HILTS within XR environments: VR, AR, MR, XR, metaverse, simulation, virtual human, digital man, agent, avatar. The Boolean operator "AND" combined the two sets of keywords, while the "OR" operator connected terms within each set (Cooper, 2010). All articles relevant to HILTS within XR for PTT published from 2013 onwards were downloaded for analysis. This time frame was chosen because educational research and practice heightened interest in VR after Oculus Rift released the first consumer-priced developer versions of an HMD in 2013 (Desai et al., 2014). The search was carried out on February 1, 2024, which resulted in 9,478 articles.
我们从 Web of Science、Scopus 和 Google Scholar 收集了与 PTT XR 中 HITLS 相关的研究。正如一些研究得出的结论（Harzing & Alakangas, 2016; Martin-Martin et al.）通过数据库词库和样本论文中的关键词确定了检索词，并利用了两种关键词模式：（1）与职前教师相关的关键词：职前教师、职前教师、候选教师、受训教师、新手教师、学生教师；（2）与XR环境中的HILTS相关的关键词：VR, AR, MR, XR, metaverse, simulation, virtual human, digital man, agent, avatar。布尔运算符 "AND "组合了两组关键词，而 "OR "运算符则连接了每组关键词中的术语（Cooper，2010 年）。下载了 2013 年以来发表的所有与 PTT XR 中的 HILTS 相关的文章进行分析。之所以选择这个时间段，是因为在 Oculus Rift 于 2013 年发布首款消费者定价的 HMD 开发者版本后，教育研究和实践对 VR 的兴趣大增（Desai 等人，2014 年）。搜索于 2024 年 2 月 1 日进行，共搜索到 9,478 篇文章。

3.2. Inclusion and exclusion criteria
3.2.纳入和排除标准

To enhance the eligibility and effectiveness of this meta-analysis, two researchers manually searched for related articles, ensuring that the studies met the criteria: (1) Peer-reviewed publications (journals, full-text conference proceedings, book chapters, and dissertations) published from 2013 to 2024; (2) Studies utilizing HITLS within XR specifically for PTT; (3) Studies with experimental or quasi-experimental design; (4) Studies with sufficient data for calculating effect size, including sample size, mean, standard deviation,

-values and

-values; (5) Exclusion of the same study published in different formats.
为了提高这项荟萃分析的合格性和有效性，两名研究人员手动搜索了相关文章，确保研究符合标准：(1) 2013 年至 2024 年间发表的同行评议出版物（期刊、全文会议论文集、书籍章节和学位论文）；(2) 在 XR 中使用 HITLS 专门用于 PTT 的研究；(3) 采用实验或准实验设计的研究；(4) 有足够数据用于计算效应大小的研究，包括样本大小、平均值、标准差、

- 值和

- 值；(5) 排除以不同格式发表的同一研究。

The literature screening and inclusion process for this study was conducted by two researchers simultaneously. Figure 1 presents the PRISMA flowchart outlining the identification, screening, coding, and extraction processes (Moher et al., 2009; Page et al., 2021). First, non-English publications and duplicates were removed, resulting in 4277 articles. Subsequently, non-experimental or quasi-experimental articles and those unrelated to HITLS within XR for PTT were excluded, leading to the identification of a total of 54 articles. Further, 15 studies were excluded due to insufficient reporting of statistical information for effect size calculation. Finally, 142 studies from 39 articles, comprising 30 journals, 2 conference papers, 6 dissertations, and 1 book chapter, met all the criteria and were included in the final meta-analysis.
本研究的文献筛选和收录过程由两名研究人员同时进行。图 1 显示了概述识别、筛选、编码和提取过程的 PRISMA 流程图（Moher 等人，2009 年；Page 等人，2021 年）。首先，剔除了非英文出版物和重复的文章，得到 4277 篇文章。随后，排除了非实验性或准实验性文章，以及与 PTT XR 中的 HITLS 无关的文章，最终共识别出 54 篇文章。此外，还有 15 篇研究因未提供足够的统计信息以计算效应大小而被排除。最后，39 篇文章（包括 30 种期刊、2 篇会议论文、6 篇学位论文和 1 篇书籍章节）中的 142 项研究符合所有标准，被纳入最终的荟萃分析。

3.3. Coding 3.3.编码

To ensure consistent analysis, data including mean scores, standard deviations, and sample sizes of both experimental and control groups were gathered for each paper. This study employed the NLN Jeffries simulation theory (Jeffries, 2021) as the foundational framework for exploring potentially effective instructional design features. This theory has demonstrated successful application in various medical health studies for instructional design purposes (Wilson & Hagler, 2012; Bowden et al., 2022). Comprising five key components-context, background, design, simulation experience, and outcomes-the NLN Jeffries simulation theory aims to create an experiential, interactive, collaborative, and learner-centered simulation experience. This theory can serve as robust theoretical support for investigating instructional design features across various phases of HITLS within XR for PTT. The detailed coding scheme, based on the framework, is elaborated below.
为确保分析的一致性，每篇论文都收集了实验组和对照组的平均分、标准差和样本量等数据。本研究采用了 NLN 杰弗里斯模拟理论（Jeffries，2021 年）作为基础框架，以探索潜在有效的教学设计特点。该理论已成功应用于各种医疗健康研究的教学设计中（Wilson & Hagler，2012；Bowden 等人，2022）。NLN 杰弗里斯模拟理论由五个关键部分组成--情境、背景、设计、模拟体验和结果--旨在创造一种体验式、互动式、协作式和以学习者为中心的模拟体验。该理论可作为研究 PTT XR 中 HITLS 各阶段教学设计特点的有力理论支持。下面将详细阐述基于该框架的编码方案。

Figure 1. PRISMA flowchart of the present study process
图 1.本研究过程的 PRISMA 流程图

Figure 2. The coding framework based on the NLN Jeffries simulation theory
图 2.基于 NLN 杰弗里斯模拟理论的编码框架

3.3.1. Coding scheme of context features
3.3.1.语境特征编码方案

(1) Setting: Establishing trust is crucial in virtual simulations (Jeffries, 2021), and the design of the simulation setting can influence learners' perception of realism (Hindmarsh et al., 2014). Therefore, studies were categorized into classroom, other offline places (e.g, lab, seminar room, home), online meeting room, blended (combination of offline places and online meeting rooms).
(1) 环境：在虚拟仿真中，建立信任至关重要（Jeffries，2021 年），仿真环境的设计会影响学习者对真实性的感知（Hindmarsh 等人，2014 年）。因此，研究被分为教室、其他线下场所（如实验室、研讨室、家庭）、在线会议室、混合（线下场所和在线会议室的组合）。

3.3.2. Coding scheme of background features
3.3.2.背景特征编码方案

(2) Curriculum integration: Some studies supported integrating HITLS sessions into PT education courses for comprehensive theoretical preparation and increased training enthusiasm (Walters et al., 2021), while others suggested that simulation performance should not impact PTs' course grades to reduce mental load and anxiety (Ely et al., 2018). Studies were categorized into yes or no based on this divergence.
(2) 课程整合：一些研究支持将HITLS课程整合到PT教育课程中，以进行全面的理论准备并提高培训热情（Walters等人，2021年），而另一些研究则建议模拟表演不应影响PT的课程成绩，以减轻心理负担和焦虑（Ely等人，2018年）。根据这一分歧，研究被分为 "是 "或 "否"。

3.3.3. Coding scheme of design features
3.3.3.设计特征的编码方案

Technology design: Regarding (3) XR type, studies were categorized into: MR, VR or AR. In terms of (4) immersion level, studies were categorized into non-immersive, semi-immersive, or fully immersive (Huang et al., 2021). With regard to (5) interactivity, inspired by Vatavu's (2023) study, three types were coded: natural interactions (intuitive and human-like, e.g., verbal and gesture interactions), non-natural interactions (conventional and artificial, e.g., menu choices); and both (a combination of both natural and non-natural interactions). Considering (6) HIRLS type, studies were categorized into avatar-mediated or agent-medicated. With respect to (7) autonomy, four types were coded: fully controlled (scenario-based, entirely user-directed), semi-autonomous (verbal behaviors typically controlled by real human actors and non-verbal behavior controlled by basic

), fully autonomous (completely driven by

, acting and making decisions independently).
技术设计：关于 (3) XR 类型，研究分为 MR、VR 或 AR：MR、VR 或 AR。关于（4）沉浸程度，研究被分为非沉浸、半沉浸和全沉浸（Huang 等人，2021 年）。关于（5）交互性，受 Vatavu（2023 年）研究的启发，对三种类型进行了编码：自然交互（直观和类人的，如语言和手势交互）、非自然交互（传统和人工的，如菜单选择）和两者（自然和非自然交互的组合）。考虑到（6）HIRLS 类型，研究被分为以虚拟化身为媒介的研究和以代理为媒介的研究。关于(7)自主性，我们对四种类型进行了编码：完全控制型（基于场景，完全由用户指导）、半自主型（语言行为通常由真人演员控制，非语言行为由基本

）、完全自主型（完全由

，独立行动和决策）。

Task design: In terms of (8) customized task, in line with the NLN Jeffries simulation theory advocating for a learner-centered approach, studies were coded as yes or no based on whether simulation tasks were tailored to individual training needs. Regarding (9) task type, studies were categorized into distributed practice or integrated practice. Distributed practice involves spaced-out learning sessions with intervals between them, allowing for better retention and skill mastery over time (Bahrick & Hall, 2005). Integrated practice, on the other hand, combines multiple skills or concepts into a single session, promoting a more comprehensive understanding and transferability of knowledge (Camburn et al., 2017). With respect to (10) task variation, studies were categorized into yes or no, considering that task variation can potentially enhance immediate task interest, increase willingness to learn, and improve task performance (Smith et al., 2009).
任务设计：在（8）定制任务方面，根据 NLN Jeffries 模拟理论提倡的以学习者为中心的方法，根据模拟任务是否根据个人培训需求量身定制，将研究编码为 "是 "或 "否"。关于（9）任务类型，研究被分为分布式练习和综合练习。分布式练习是指学习课程之间有间隔，以便更好地保持和掌握技能（Bahrick & Hall, 2005）。而综合实践则是将多种技能或概念结合到一个环节中，促进对知识更全面的理解和迁移（Camburn et al.，2017）。关于（10）任务变异，考虑到任务变异有可能增强直接任务兴趣、提高学习意愿并改善任务绩效，研究被分为 "是 "或 "否"（Smith 等人，2009 年）。

(11) Simulation length: Studies were categorized into shorter than 30 minutes and longer than 30 minutes (

et al., 2023).
(11) 模拟时长：研究时间分为短于 30 分钟和长于 30 分钟 (

et al., 2023)。

(12) Repeated training: Studies highlighted the significance of collaborative training to enhance the interactivity (Ke et al., 2016). However, individual training was also supported in many included studies (Delamarre et al., 2020). Hence, training modes were categorized as yes or no.
(12) 重复训练：研究强调了合作培训对增强互动性的重要意义（Ke 等人，2016 年）。然而，许多研究也支持个人培训（Delamarre 等人，2020 年）。因此，培训模式被分为 "是 "或 "否"。

(13) Training mode: To create a more interactive simulation experience for PTs, studies emphasized the importance of collaborative training (Ke et al., 2016). However, many of the included studies also utilize
(13) 培训模式：为了给护理人员创造更具互动性的模拟体验，研究强调了协作培训的重要性（Ke 等人，2016 年）。然而，许多纳入的研究还利用了
individual training (Delamarre et al., 2020). Therefore, two types of training mode were coded: individually or collaboratively.
个人培训（Delamarre et al.）因此，我们对两种培训模式进行了编码：个人培训或合作培训。

(14) Assessment approach: Studies were categorized into during training or after training based on when to assess HITLS outcomes within XD for PTT.
(14) 评估方法：根据 PTT 在 XD 内评估 HITLS 成果的时间，将研究分为培训期间和培训后两类。

3.3.4. Coding scheme of simulation experience features
3.3.4.模拟体验特征的编码方案

(15) Facilitator role: Facilitators play a crucial role in establishing HITLS trust, impacting PTs' perception of coercion (Richter et al., 2022). In the included studies, four types of facilitator roles were identified: faculty teacher only, researcher only, faculty teacher & researcher or faculty teacher & technical specialists.
(15) 调解人的作用：促进者在建立 HITLS 信任方面发挥着关键作用，影响着 PTs 对胁迫的感知（Richter 等人，2022 年）。在纳入的研究中，确定了四种类型的促进者角色：仅教师、仅研究人员、教师与研究人员或教师与技术专家。

(16) Participate role: The significance of vicarious observational learning has been highlighted in various studies (Hodges et al., 2007; Greer et al., 2020). Within the included studies, three participant roles were identified: participant only (actively engaged in simulation themselves), observer only (merely observing others training), and both (both engaged and observing).
(16) 参与角色：各种研究都强调了替代观察学习的重要性（Hodges 等人，2007 年；Greer 等人，2020 年）。在所纳入的研究中，确定了三种参与者角色：仅参与者（自己积极参与模拟）、仅观察者（仅观察他人训练）和两者（既参与又观察）。

Instructional support: Dai (2023) introduced a model-based learning support framework, offering instructional support at three stages: pre-training, in-situ training, and post-training. Inspired by this, in the pre-training stage three instructional support features were considered: introduction, theoretical preparation, and rehearsal, each coded as yes or no. (17) Introduction involves introducing the training process, objectives, technical features, etc., to help PTs reduce anxiety and cognitive load, thus facilitating PTs' engagement and learning in simulation sessions (Pellas & Boumpa, 2016). (18) Theoretical preparation involves PTs preparing lesson plans, attending lectures, and accessing theoretical materials before simulations. It aims to provide theoretical support, guide simulation sessions, and help PTs utilize HITLS effectively, facilitating theory-to-practice transfer (Brantlinger et al., 2022). (19) Rehearsal refers to PTs practicing before formal simulation sessions to familiarize themselves with training content and technical environments (Dai, 2023).
教学支持：Dai（2023）提出了一个基于模型的学习支持框架，在三个阶段提供教学支持：培训前、现场培训和培训后。受此启发，在培训前阶段，考虑了三个教学支持特征：介绍、理论准备和演练，每个特征都编码为 "是 "或 "否"。(17) 介绍包括介绍培训过程、目标、技术特点等，以帮助 PTs 减少焦虑和认知负荷，从而促进 PTs 在模拟课程中的参与和学习（Pellas & Boumpa，2016）。(18) 理论准备包括教师在模拟教学前准备教案、参加讲座和查阅理论资料。其目的是提供理论支持，指导模拟课程，帮助PT有效利用HITLS，促进理论到实践的转化（Brantlinger等人，2022）。(19) 排练是指康复治疗师在正式模拟训练前进行练习，以熟悉训练内容和技术环境（Dai，2023）。

In the in-situ training stage, (20) prompt was considered to reduce PTs' cognitive load and promote positive engagement (Derakhshan et al., 2023) and it was coded yes or no. In the post-training training stage, three instructional design features were considered: reflection, static feedback and other scaffolds, each coded yes or no. (21) Reflection, the process of creating and clarifying the meaning of experience, has been emphasized for its importance in numerous studies (Larson et al., 2020). (22) Static feedback, typically provided by peers and instructors in HITL systems, is crucial for the cognitive development and skill enhancement of PTs. (23) Other scaffolds, such as debriefing activities like group discussions, are considered essential steps in HITLS (Dufrene & Young, 2014).
在原位培训阶段，（20）提示被认为可以减轻 PTs 的认知负荷并促进积极的参与（Derakhshan et al.在培训后的培训阶段，有三个教学设计特征被考虑：反思、静态反馈和其他支架，每个都被编码为是或否。(21) 反思是创造和澄清经验意义的过程，在许多研究中都强调了其重要性（Larson 等人，2020 年）。(22) 静态反馈通常由 HITL 系统中的同伴和指导教师提供，对 PT 的认知发展和技能提高至关重要。(23) 其他支架，如小组讨论等汇报活动，被认为是 HITLS 的基本步骤（Dufrene 和 Young，2014 年）。

3.3.5. Coding scheme of outcome features
3.3.5.结果特征编码方案

(24) Outcome: Kirkpatrick's four-level training evaluation model categorizes training outcomes into four levels: level 1 evaluates trainees' satisfaction with the program; level 2 evaluates changes in attitudes, knowledge improvement, and/or skill enhancement; level 3 evaluates the application of acquired knowledge and skills in job performance; and level 4 evaluates the broader organizational impact, including reduced costs, improved quality, and increased productivity (Kirkpatrick & Kirkpatrick, 2016; Alsalamah & Callinan, 2021). Almost all the studies included in this meta-analysis focused on levels 1 and 2, so this study concentrated on these two levels of
(24) 成果：柯克帕特里克的四级培训评估模型将培训结果分为四个级别：第一级评估受训者对培训项目的满意度；第二级评估态度的转变、知识的提高和/或技能的增强；第三级评估所学知识和技能在工作绩效中的应用；第四级评估更广泛的组织影响，包括成本的降低、质量的改善和生产率的提高（Kirkpatrick & Kirkpatrick, 2016; Alsalamah & Callinan, 2021）。本荟萃分析中包含的几乎所有研究都侧重于第 1 和第 2 层次，因此本研究集中于这两个层次

Kirkpatrick's training evaluation model. The outcomes were coded as satisfaction, attitude, knowledge, and skills.
Kirkpatrick 的培训评估模型。结果被编码为满意度、态度、知识和技能。

Two researchers independently coded the included studies to ensure reliability. Cohen's Kappa showed agreement scores above 0.9 for all codes. And any discrepancies were resolved through discussion to reach a consensus. Appendix C provides details on the characteristics of publications included in this meta-analysis.
两名研究人员对纳入的研究进行了独立编码，以确保可靠性。科恩卡帕（Cohen's Kappa）显示，所有编码的一致性得分均高于 0.9。任何差异都通过讨论达成共识。附录 C 详细介绍了本荟萃分析所纳入的出版物的特点。

2.4. Data analysis 2.4.数据分析

Meta-Analysis (CMA) Version 3.0 was used to analyze data extracted from primary studies, Data extracted from a primary study, including mean, standard deviation, and sample size, were used to calculate the effect size. If mean and standard deviation were not available in a study, other statistics like sample size, t-value and effect direction were used to estimate an effect size (Glass, 1976). Then, all the obtained effect sizes were converted to Hedges'

(Hedges, 1981) to minimize the impact of small sample studies. The overall effect size was estimated using the random-effects models for all of the studies considering the variability of the research design (Borenstein et al., 2021).
元分析（CMA）3.0 版用于分析从主要研究中提取的数据，从主要研究中提取的数据包括平均值、标准差和样本量，用于计算效应大小。如果研究中没有平均值和标准差，则使用样本量、t 值和效应方向等其他统计数据来估计效应大小（Glass，1976 年）。然后，将所有获得的效应大小转换为赫奇斯

（赫奇斯，1981 年），以尽量减少小样本研究的影响。考虑到研究设计的可变性（Borenstein et al.）

Cochran's

test is one of the non-parametric tests of variability used for statistical tests for analyzing the variability of multiple samples (with the same number of samples). I2 was used as a measure of heterogeneity due to non-accident as a percentage of total variation in the effect size. In view of the above explanation, calculation of Q-values and I2 test were used to find the presence of heterogeneity. According to Higgins et al. (2003), the thresholds for low, medium and high heterogeneity were 25,50 and 75 percent respectively. A significance level of 0.05 was chosen to describe between-group differences when calculating the Q-statistic.
Cochran's

检验是用于分析多个样本（样本数相同）变异性统计检验的非参数变异性检验之一。I2 用于衡量由于非偶然性导致的异质性占效应大小总变异的百分比。鉴于上述解释，计算 Q 值和 I2 检验用于发现异质性的存在。根据 Higgins 等人（2003 年）的研究，低、中、高异质性的临界值分别为 25%、50% 和 75%。在计算 Q 统计量时，选择 0.05 的显著性水平来描述组间差异。

Random effects models claim that the true effect sizes may be different due to certain moderating variables, such as age, grade level, etc. That is, the random effects model can be used to estimate and generalize effect sizes for larger populations if the studies are not functionally identical. Therefore, the moderator variable analysis was conducted using a random effects model to test the effect of the identified moderator variables on detecting and explaining heterogeneity in effect sizes.
随机效应模型认为，由于某些调节变量（如年龄、年级等）的存在，真实的效应大小可能不同。也就是说，如果研究在功能上不完全相同，随机效应模型可以用来估计和归纳更大人群的效应大小。因此，我们使用随机效应模型进行了调节变量分析，以检验已确定的调节变量对检测和解释效应量异质性的影响。

To avoid any possible publication bias in the meta-analysis, the authors used funnel plots, Egger's test, classical fail-safe N, and KendallTau rank test. First, the researchers drew and examined funnel plots based on individual study data, followed by Egger's regression tests (Egger et al., 1997) and a Kendall's Tau rank test (Begg & Mazumdar, 1994) of the random effects model to determine the presence of publication bias. For both tests, a significance level of 0.05 was used. Finally, a build and fill method was again implemented. Rosental (1995) stated that publication bias didn't affect the results of a meta-analysis if the 'number of fail-safes

' is greater than a tolerance of

(where

is the total number of effect sizes reported in the meta-analysis).
为了避免荟萃分析中可能出现的发表偏倚，作者使用了漏斗图、Egger 检验、经典失效安全 N 和 KendallTau 秩检验。首先，研究人员根据单个研究数据绘制并检查了漏斗图，然后对随机效应模型进行了 Egger 回归检验（Egger 等，1997 年）和 Kendall's Tau 秩检验（Begg & Mazumdar，1994 年），以确定是否存在发表偏倚。这两项检验的显著性水平均为 0.05。最后，再次采用建立和填充法。Rosental（1995）指出，如果 "失败保险箱的数量

"大于

的容忍度（其中

是荟萃分析中报告的效应大小的总数），则发表偏差不会影响荟萃分析的结果。

4. Results 4.成果

4.1. Descriptive statistics
4.1.描述性统计

The current meta-analytic data set included 39 studies yielding 142 effect sizes from a cumulative total 1564 participants. All the studies were published from 2013 to 2024, with the majority (80%) published after 2020 , and most of the studies were journal articles (80%). Appendix

describes the frequencies and percentages of studies for the subcategories of each code.
目前的荟萃分析数据集包括 39 项研究，从累计 1564 名参与者中得出了 142 个效应大小。所有研究均发表于 2013 年至 2024 年，其中大部分（80%）发表于 2020 年之后，且大部分研究为期刊论文（80%）。附录

描述了每个代码子类别的研究频率和百分比。

4.2. Overall effect on PTT
4.2.对 PTT 的总体影响

In this meta-analysis, a final

independent effect sizes were obtained, totaling

participants. According to the forest plot in Figure 3, effect size values ranged from -3.665 to 9.363, with only 39 studies producing negative effects. Table 1 shows the overall effect sizes. The random effects model was used to calculate the effect sizes of the 142 selected studies. The results showed that the overall effect size of HILTS within XR for PTT was 0.535 with a

confidence interval of

. According to Cohen's findings, the effect sizes of

, and 0.20 were regarded as larger, medium, and small effect sizes respectively. Therefore, HILTS within XR produced a medium effect on PTT outcomes. The test of heterogeneity revealed that the effect sizes were heterogeneous in the present study (Qtotal

). The

statistic value of

larger than

suggested a high heterogeneity which means the effects of moderator variables should be investigated.
在这项荟萃分析中，最终获得了

个独立的效应大小，共有

名参与者。根据图 3 中的森林图，效应大小值从-3.665 到 9.363 不等，只有 39 项研究产生了负效应。表 1 显示了总体效应大小。随机效应模型用于计算 142 项选定研究的效应大小。结果显示，XR 中 HILTS 对 PTT 的总体效应大小为 0.535，置信区间为

。根据 Cohen 的研究结果，

、0.20 和 0.20 的效应量分别被视为较大、中等和较小的效应量。因此，XR 中的 HILTS 对 PTT 结果产生了中等程度的影响。异质性检验显示，本研究中的效应大小具有异质性（Qtotal

）。

统计值

大于

表明异质性较高，这意味着应研究调节变量的影响。

Table 1. Overall effect sizes
表 1.总体效应大小

			95% CI		Test of mean 平均值测试	Test of heterogeneity 异质性检验
			Lower 较低	Upper 上层			I2

outcome 结果

Note.

. 注：.

The first research question of the study aimed to investigate whether HILTS within XR had a significant positive impact on PTT outcomes and what the effect size was. Based on the aforementioned meta-analysis results, it was evident that HILTS within XR exhibited a medium effect, with an effect size of 0.535 . To explore the varied impact of HILTS within XR for PTT, this study analyzed the effect size on satisfaction, attitude, knowledge, and skill, as shown in Appendix E. The results indicated that the effect sizes across different dimensions of training outcomes were not significantly different, with knowledge

satisfaction

skills

attitude (

). All effect sizes were positive, suggesting a positive influence of HITLS within XR on PTT training outcomes. Despite the inter-group effect test

) revealing no significant differences across dimensions, all dimensions exhibited noticeable promotion effects, reaching medium levels.
本研究的第一个研究问题旨在探讨 XR 中的 HILTS 对 PTT 结果是否有显著的积极影响，以及影响大小如何。根据上述荟萃分析结果，XR 中的 HILTS 显现出中等效应，效应大小为 0.535。为了探讨 XR 中的 HILTS 对 PTT 的不同影响，本研究分析了满意度、态度、知识和技能的效应大小，如附录 E 所示。结果表明，培训结果不同维度的效应大小差异不大，知识

满意度

技能

态度 (

)。所有效应量均为正值，表明 XR 中的 HITLS 对 PTT 培训结果有积极影响。尽管组间效应检验

) 表明不同维度之间没有显著差异，但所有维度都表现出明显的促进效应，达到中等水平。

4.3. Moderator effects on HILTS within XR
4.3.XR 中对 HILTS 的调节效应

4.3.1. Moderator effects during the planning phase
4.3.1.规划阶段的主持人效应

APPENDIX E indicated that among the context & background features, setting

0.05) moderated the HILTS within XR effects. HILTS within XR in a blended experimental setting produced the largest amount of effects, followed by other offline places, online meeting room and classroom. Though curriculum integration did not moderated the HILTS within XR effects, the result showed that HILTS within XR embedded in a PT education course achieved more significant outcomes.
附录 E 表明，在情境和背景特征中，环境

0.05）调节了 XR 中的 HILTS 效果。混合实验环境中的 XR 中的 HILTS 效果最大，其次是其他线下场所、在线会议室和教室。虽然课程整合没有调节 XR 中的 HILTS 效果，但结果表明，嵌入 PT 教育课程的 XR 中的 HILTS 取得了更显著的效果。

Among the design features, graphical realism

), autonomy (

, task variation (

, and simulation length

moderated the HILTS within XR effects. Among all the type of autonomy, semi-autonomous virtual humans had the highest effect size, followed by fully controlled, and fully autonomous virtual humans. For task variation, the effective size was larger when HILTS within XR has task variation. Additionally, simulation length was found to be a significant moderator. Longer simulation sessions yielded better PTT outcomes compared to sessions lasting less than 30 minutes.
在设计特征中，图形逼真度

）、自主性 (

, 任务变化 (

, 和模拟长度

调节了 XR 内的 HILTS 效应。在所有自主类型中，半自主虚拟人的效应大小最高，其次是完全受控虚拟人和完全自主虚拟人。就任务变异而言，当 XR 中的 HILTS 具有任务变异时，有效大小更大。此外，模拟时间的长短也是一个重要的调节因素。与少于 30 分钟的模拟课程相比，更长的模拟课程能产生更好的 PTT 效果。

Other moderators did not show statistical significance, but notable findings emerged from variations in effect sizes. All types of XR exhibited similar and significant effects, indicating XR environments support PTT, with each type offering unique advantages. Non-immersive HITLS within XR showed the largest effect size, followed by fully immersive and semi-immersive HITLS. While there was little variation in effect sizes among the three types of system interactivity, avatar-mediated interactive training stood out with a larger effect size, often employing natural interactions. Moreover, HITLS within XR, incorporating customized tasks, distributed practice, demonstrated superior outcomes. Notably, HITLS conducted collaboratively, with repeated training yielded significantly larger effect sizes. And training outcomes assessed during sessions surpassed those evaluated post-training.
其他调节因素没有显示出统计学意义，但从效应大小的变化中得出了值得注意的结论。所有类型的 XR 都表现出类似的显著效果，这表明 XR 环境支持 PTT，每种类型都具有独特的优势。XR 中的非沉浸式 HITLS 显示出最大的效应大小，其次是完全沉浸式和半沉浸式 HITLS。虽然三种类型的系统交互性在效果大小上差异不大，但以虚拟化身为中介的交互式培训效果较大，通常采用自然的交互方式。此外，XR 中的 HITLS 结合了定制任务、分布式练习，显示出更优越的效果。值得注意的是，合作进行的 HITLS 和重复培训的效果显著更大。而且，在培训过程中评估的培训效果超过了培训后评估的效果。

4.3.2. Moderator effects during the implementation phase
4.3.2.实施阶段的主持人效应

While the role of facilitators and participants did not significantly moderate the effectiveness of PTT, some insights emerged from the variations in effect sizes. The impact of different facilitator roles on PTT outcomes did not show significant differences, and participants who both engaged in HITLS sessions and observed others training achieved the largest effect.
虽然主持人和参与者的角色对 PTT 的效果没有明显的调节作用，但从效果大小的变化中可以得出一些启示。不同主持人的角色对 PTT 效果的影响没有显示出显著差异，既参与 HITLS 课程又观察他人培训的参与者取得的效果最大。

Among instructional support features, in the pre-training and in-situ stages, introduction (

, theoretical input

, rehearsal

, and prompt (

) were significant moderators. In the post-training stage, static feedback

and other scaffolds

moderated the PTT outcomes, with HILTS within XR providing static feedback and other scaffolds producing significantly larger effect sizes. Interestingly, the effect-size difference of reflection was found to be non-significant

).
在教学支持特征中，在培训前和现场阶段，导入（

、理论输入

、演练

和提示（

）是显著的调节因子。在培训后阶段，静态反馈

和其他支架

调节了 PTT 的结果，XR 中的 HILTS 提供的静态反馈和其他支架产生了明显更大的效应大小。有趣的是，反思的效应大小差异并不显著

）。

4.3.3. Moderator effects during the evaluation phase
4.3.3.评估阶段的主持人效应

Training outcome was found not a significant moderator. Across the four dimensions of training outcomes satisfaction, attitude, knowledge, and skills, HILTS within XR demonstrated similarly positive effects.
培训结果并不是一个重要的调节因素。在培训结果的满意度、态度、知识和技能这四个方面，XR 中的 HILTS 也显示出类似的积极效果。

The second research question of this study aimed to explore potentially effective instructional design features of HILTS within XR in the planning, implementation, and evaluation phases, respectively. Through the moderating effects analysis described above, the study identified five potentially effective instructional design features of HILTS within XR during the planning phase: setting, graphical realism, autonomy, task variation, and simulation length. In the implementation phase, eight potentially effective instructional design features were identified: introduction, theoretical preparation, rehearsal, prompt, static feedback, and other scaffolds. However, no significant features were found in the evaluation phase of HILTS within XR for PTT.
本研究的第二个研究问题旨在探索 XR 中的 HILTS 在计划、实施和评估阶段分别具有的潜在有效教学设计特征。通过上述调节效应分析，本研究确定了在计划阶段，XR 中的 HILTS 的五个潜在有效的教学设计特征：设置、图形真实性、自主性、任务变化和模拟长度。在实施阶段，发现了八个潜在有效的教学设计特征：导入、理论准备、演练、提示、静态反馈和其他支架。然而，在针对 PTT 的 XR 中的 HILTS 评估阶段，没有发现任何重要特征。

4.4. Publication bias 4.4.出版偏差

The funnel plot produced by the meta-analysis (see Figure 4) shows a predominantly symmetrical distribution of effect values, with only a few points showing bias, suggesting that publication bias may not be significant. To further check the symmetry, Egger's test

was performed, which were statistically non-significant. The Classic fail-safe

test showed that 4375 more studies would be required to nullify the overall effect size found in this meta-analysis, which is more than

(i.e. 145). Accordingly, the overall findings suggested a low chance of publication bias.
荟萃分析得出的漏斗图（见图 4）显示，效应值的分布主要是对称的，只有少数几个点出现偏倚，这表明发表偏倚可能并不严重。为了进一步检验对称性，我们进行了 Egger 检验

，结果显示在统计学上并不显著。经典失效安全

检验表明，要使本次荟萃分析中发现的总体效应大小失效，还需要 4375 项研究，而这一数字超过了

（即 145 项）。因此，总体研究结果表明发表偏倚的可能性较低。

Figure 4. Funnel plot of standard error by effect size
图 4.按效应大小划分的标准误差漏斗图

5. Recommendations on instructional design guidelines for HILTS within XR
5.关于 XR 系统内 HILTS 教学设计指南的建议

The findings of this study suggested that HITLS within XR generally had a moderately positive impact on PTT
这项研究的结果表明，XR 中的 HITLS 对 PTT 一般具有适度的积极影响。
outcomes. Furthermore, the study revealed that PTT outcomes seemed to have homogeneous effect sizes among subcategories, including satisfaction, attitude, knowledge, and skill. The findings revealed that HITLS within XR can effectively address various training objectives in PTT, enhancing both cognitive development and fostering a positive attitude towards their future careers. Through immersive and responsive simulation experiences, PTs can refine their teaching skills, particularly in scenarios requiring quick decision-making and adaptability in the classroom. These results extended previous arguments that human-in-the-loop simulations offered real-time, interactive training environments, which not only enhanced trainees' cognitive performance but also evoked their emotional responses, as supported by previous research (Davidson et al., 2021; Kucek & Leitner, 2020).
此外，研究还显示，PTT 结果在满意度、态度、知识和技能等子类别中的效应大小似乎是相同的。此外，研究还显示，在满意度、态度、知识和技能等子类别中，PTT 成果似乎具有相同的效应大小。研究结果表明，XR 中的 HITLS 可有效实现 PTT 的各种培训目标，既能促进认知发展，又能培养他们对未来职业的积极态度。通过身临其境、反应灵敏的模拟体验，PTT 可以提高他们的教学技能，特别是在需要快速决策和课堂适应能力的场景中。这些结果扩展了之前的论点，即人在环模拟提供了实时、互动的培训环境，不仅提高了学员的认知表现，还唤起了他们的情感反应，这一点得到了之前研究的支持（Davidson 等人，2021 年；Kucek & Leitner，2020 年）。

The moderator analysis of instructional features during the planning, implementation, and evaluation phases of HITLS within XR for PTT revealed potentially effective instructional design guidelines for its successful application. These guidelines are outlined as follows:
通过对 PTT XR 中 HITLS 的规划、实施和评估阶段的教学特点进行主持人分析，发现了成功应用 HITLS 的潜在有效教学设计准则。这些指导原则概述如下：

5.1. Instructional design guidelines during the planning phase
5.1.规划阶段的教学设计指南

HITLSs within XR for PTT integrated into a curriculum or training program had a larger mean effect size than those that were not integrated. Therefore, when applying HITLSs within XR for PTT, it's recommended to integrate it into PT education courses or programs. This aligns with Cox and King's (2006) viewpoint that simulations offered learners the chance to tackle problem-solving tasks within a "real-world" context, and when training sessions were embedded within courses, it can effectively bridge the gap between theory and practice, thereby enhancing learners' engagement in practice sessions.
与未整合的课程或培训项目相比，整合到 PTT XR 中的 HITLS 平均效应大小更大。因此，在 PTT XR 中应用 HITLS 时，建议将其整合到 PT 教育课程或项目中。这与Cox和King（2006）的观点不谋而合，即模拟训练为学习者提供了在 "真实世界 "背景下解决问题的机会，当训练课程嵌入课程中时，它能有效地弥合理论与实践之间的差距，从而提高学习者在实践课程中的参与度。

When considering choosing HITLS type for PTT, a larger mean effect size was observed with avatar-mediated interactive training. Therefore, it's recommended to opt for this type of HITLS when designing HITLS sessions within XR. Such virtual environments design an interface that combines virtual and real elements, with avatars controlled by human actors, which facilitates a more stable and controlled training environment. This setup enables interactive textual conversations without forced choices or branching, allowing PTs to respond freely, thereby refining their teaching skills in authentic and complex teaching scenarios (Ade-Ojo et al., 2022). While agent-mediated interactive training also proved effective for PTT outcomes, the effect size was notably smaller.
在考虑为 PTT 选择 HITLS 类型时，观察到以化身为媒介的互动式培训的平均效应大小更大。因此，在 XR 中设计 HITLS 课程时，建议选择这种类型的 HITLS。这种虚拟环境设计了一个结合虚拟和现实元素的界面，虚拟化身由人类演员控制，这有利于营造一个更加稳定和可控的培训环境。这种设置可实现交互式文本对话，而不强制选择或分支，使教学人员能够自由回应，从而在真实、复杂的教学场景中提高教学技能（Ade-Ojo 等人，2022 年）。虽然以代理为中介的互动培训也被证明对 PTT 成果有效，但其效果明显较小。

Among current agent-mediated interactive training systems, those situated in non-immersive environments with scenario-based agent behaviors, featuring limited branches, showed better training outcomes than fully immersive environments with autonomous agents driven by advanced AI (e.g., large language models) did. This could be attributed to the relatively limited presence of autonomous agents in current HITLS systems, requiring further research for confirmation, but also indirectly indicating that the application effectiveness of autonomous agents may need enhancement due to technological limitations. Additionally, the study found no statistically significant difference between XR type, immersion level, and interactivity, suggesting that the fidelity of avatars or agents within the system is more crucial for the consideration of HITLS systems.
在目前的以代理为媒介的互动培训系统中，那些在非沉浸式环境中使用基于场景的代理行为、具有有限分支的系统，比在完全沉浸式环境中使用由高级人工智能（如大型语言模型）驱动的自主代理的系统显示出更好的培训效果。这可能是由于目前的 HITLS 系统中自主代理的存在相对有限，需要进一步研究证实，但也间接表明，由于技术限制，自主代理的应用效果可能需要提高。此外，研究还发现，XR 类型、沉浸程度和交互性之间在统计学上没有显著差异，这表明系统中虚拟化身或代理的保真度对于 HITLS 系统的考量更为关键。

It is worth noting that simulation setting and training mode weren't significant moderators. Specifically, HITLS carried out in a blended manner collaboratively supported the best PTT outcomes. This echoes the Ade-Ojo et al.'s (2022) conclusions that combining online video connectivity tools such as Zoom© or Skype© in offline MR environments can not only prevent any emergency situations but also provide more opportunities to promote human connectivity and increase the flexibility of simulations for PTT. Furthermore, though no significance found, HITLS within XR with repeated training opportunities showed a much larger mean effect size. This is consistent
值得注意的是，模拟环境和培训模式并不是重要的调节因素。具体来说，以混合方式开展的 HITLS 协同支持了最佳 PTT 成果。这与 Ade-Ojo 等人（2022 年）的结论不谋而合，即在离线 MR 环境中结合在线视频连接工具（如 Zoom© 或 Skype©）不仅可以防止任何紧急情况的发生，还能提供更多机会促进人与人之间的连接，提高 PTT 模拟的灵活性。此外，虽然没有发现显著性，但在 XR 中进行重复训练的 HITLS 显示出更大的平均效应大小。这与
with Taylor et al.'s (2015) findings that repeated training is essential as it promotes reflection and enhances the retention and recall of information. In terms of the assessment approach of HITLS, better PTT outcomes were achieved when PTT performance was assessed during training sessions. This may be attributed to the fact that if assessments were solely conducted after training sessions, relying on teachers' recall of their performance, they may lack the detailed information necessary for informed decision-making.
泰勒等人（2015 年）的研究结果表明，重复培训至关重要，因为它能促进反思，增强对信息的记忆和回想。就 HITLS 的评估方法而言，如果在培训课程期间对 PTT 表现进行评估，则 PTT 成果会更好。这可能是因为，如果仅在培训课后进行评估，依靠教师对自己表现的回忆，可能会缺乏做出明智决策所需的详细信息。

Considering specific task design of HITLS within XR for PTT, while customized tasks were not significant moderators, it's noteworthy that tasks tailored for PTs resulted in better outcomes. This may be attributed to the effectiveness of tasks that accommodate diverse learning styles and training needs, enhancing training outcomes through personalized, and flexible learning experiences, thus boosting PTs' motivation and promoting their training effectiveness (Niebaum & Munakata, 2023). This aligns with the NLN simulation theory's emphasis on learner-centered simulation experiences for optimal training outcomes (Jeffries, 2021). Additionally, though task type did not significantly affect the effectiveness of HITLS, it is advised to consider both integrated practice for distributed practice tasks. Distributed practice allows PTs to practice specific teaching skills, deepening their understanding and mastery of handling certain teaching scenarios. Simultaneously, integrated practice provides a more realistic and complex teaching context, enabling PTs to apply and integrate the skills they have learned in simulated classroom environments (Scheutz et al., 2019). Task variation and simulation length were significant moderators. Including a variety of tasks and extending the simulation duration in HITLS is advisable to enable teachers to broaden their knowledge and skills (Sullivan et al., 2020).
考虑到针对 PTT 的 XR 中 HITLS 的具体任务设计，虽然定制任务不是显著的调节因素，但值得注意的是，为 PT 量身定制的任务取得了更好的结果。这可能归因于适应不同学习风格和培训需求的任务的有效性，通过个性化和灵活的学习体验提高培训效果，从而提高 PTs 的积极性并促进其培训效果（Niebaum & Munakata，2023 年）。这与 NLN 模拟理论强调的以学习者为中心的模拟体验以获得最佳培训效果的理念相一致（Jeffries，2021 年）。此外，虽然任务类型对 HITLS 的效果没有显著影响，但建议同时考虑综合练习和分布式练习任务。分布式练习允许康复治疗师练习特定的教学技能，加深他们对处理特定教学场景的理解和掌握。与此同时，综合练习提供了更真实、更复杂的教学情境，使教师能够在模拟课堂环境中应用和整合所学技能（Scheutz 等人，2019 年）。任务变化和模拟长度是重要的调节因素。在 HITLS 中加入各种任务并延长模拟时间是明智之举，可使教师拓宽知识和技能（Sullivan 等人，2020 年）。

5.2. Instructional design guidelines during the implementation phase
5.2.实施阶段的教学设计指南

While the NLN Jeffries simulation theory emphasizes the importance of simulation facilitator roles in creating an environment of trust, no significant differences in effect size were found (Jeffries, 2021). This contrasts with the expectations of many studies suggesting that the presence of faculty teachers or training experts would increase PTs' trust in HITLS and enhance their engagement (Putman, 2021; Spencer et al., 2019). One possible explanation for this could be that PTs, as adult learners, are inherently motivated and self-directed as implied by adult learning theory (Merriam, 2018). Even though the results in the present study suggested that the impact of participate role might not be significant, PTs acted as observers achieved positive training outcomes. Therefore, it is advisable to consider the role of observers teachers played during HITLS. By observation, learners can pick up a range of fundamental motor skills, and the mechanisms that underlie this learning appear to be automatic, enduring, and unaffected by distraction (Burke et al., 2010).
虽然 NLN 杰弗里斯模拟理论强调了模拟促进者角色在创造信任环境方面的重要性，但在效果大小方面并未发现显著差异（杰弗里斯，2021 年）。这与许多研究的预期形成了鲜明对比，这些研究表明，教师或培训专家的存在将增加 PTs 对 HITLS 的信任并提高他们的参与度（Putman，2021；Spencer 等人，2019）。对此的一种可能解释是，根据成人学习理论（Merriam，2018），作为成人学习者的 PTs 本身具有积极性和自我导向性。尽管本研究的结果表明，参与角色的影响可能并不显著，但作为观察员的 PT 取得了积极的培训成果。因此，不妨考虑教师在 HITLS 培训中扮演的观察者角色。通过观察，学习者可以掌握一系列基本的运动技能，这种学习的基础机制似乎是自动、持久和不受分心影响的（Burke 等人，2010 年）。

Considering the dynamic nature of HITLS, where PTs need to constantly adapt and make decisions, providing comprehensive instructional support becomes imperative (Dai, 2023). In the pre-training stage, three instructional support features - introduction, theoretical preparation, and rehearsal - all showed significant effect sizes. Introduction sessions help PTs establish a cognitive framework for the task, clarifying learning objectives and enhancing motivation. Theoretical preparation sessions provide essential knowledge, aiding understanding of the task's background. Rehearsal sessions can boost preparedness and confidence, reducing anxiety. These findings align with previous studies suggesting that such support aids in overcoming preconceptions, facilitating theoretical input, and increasing familiarity before virtual simulations (Walters et al., 2021; Ely et al., 2018; Dueck, 2021). In the in-situ training stage, prompt plays a significant moderating role, emphasizing its importance in assisting PTs to quickly respond and adjust strategies in HITL, thereby enhancing learning outcomes and training effectiveness (Tilson et al., 2017). In the post training stage, the static feedback from peers or instructors into HITLS is strongly recommended. This aligns with the findings of a study by Ellis and
考虑到 HITLS 的动态性质，即 PT 需要不断适应并做出决策，提供全面的教学支持势在必行（Dai，2023 年）。在培训前阶段，三种教学支持特征--导入、理论准备和演练--都显示出显著的效应大小。导入环节帮助学员建立任务认知框架，明确学习目标，增强学习动力。理论准备环节提供了必要的知识，有助于理解任务的背景。排练环节可以增强准备和信心，减少焦虑。这些研究结果与之前的研究结果一致，即这种支持有助于克服先入为主的观念，促进理论输入，并在虚拟模拟之前增加熟悉度（Walters 等人，2021 年；Ely 等人，2018 年；Dueck，2021 年）。在原位训练阶段，提示起着重要的调节作用，强调其在协助 PT 在 HITL 中快速反应和调整策略方面的重要性，从而提高学习成果和训练效果（Tilson 等人，2017 年）。在培训后阶段，强烈建议将同行或导师的静态反馈纳入 HITLS。这与 Ellis 和 Ellis 等人的研究结果一致。

Loughland (2017) which underscored the importance of teacher candidates receiving feedback to make periodic adjustments and achieve desired objectives. This offers PTs the chance to exchange experience and knowledge with peers, facilitating a deeper grasp of HITLS. Furthermore, it fosters an interactive and collaborative learning atmosphere, enhancing PTs' interest and motivation towards HITLS (Minott, 2021). Regarding reflection, the present study found that while reflection is vital for teachers' development (Zeichner & Liu, 2019), surprisingly, studies without explicit reflection achieved better results. This could be because teachers naturally reflect during training, especially when supported by customized tasks and rehearsals. Therefore, excessive post-training reflection might burden PTs if they already engage in reflection-in-actions (Cattaneo & Motta, 2021).
Loughland (2017)强调了师范生接受反馈以进行定期调整和实现预期目标的重要性。这为师范生提供了与同伴交流经验和知识的机会，有助于他们更深入地掌握 HITLS。此外，它还营造了一种互动和协作的学习氛围，提高了师范生对 HITLS 的兴趣和动力（Minott，2021 年）。关于反思，本研究发现，虽然反思对教师的发展至关重要（Zeichner & Liu, 2019），但令人惊讶的是，没有明确反思的研究取得了更好的结果。这可能是因为教师在培训过程中会自然而然地进行反思，特别是在定制任务和演练的支持下。因此，如果教师已经在行动中进行了反思，过多的培训后反思可能会给他们带来负担（Cattaneo & Motta, 2021）。

5.3. Instructional design guidelines during the evaluation phase
5.3.评估阶段的教学设计指南

Kirkpatrick's training evaluation model outlines four levels of evaluating training programs: satisfaction, learning, job performance, and organizational impact. Advancing through these levels enhances the depth of evaluation but also requires more time and effort. However, evaluations of HITLS within XR for PTT outcomes in the included studies have primarily focused on the first two levels. Assessments of job performance (level 3) and organizational impact (level 4) are more intricate and time-consuming yet essential for determining PTs' ability to apply HITLS learning in real classrooms and for gauging overall program effectiveness. Thus, future research should incorporate evaluations of job performance and organizational impact to provide a more comprehensive assessment of HITLS within XR for PTT.
柯克帕特里克的培训评估模型概述了培训项目评估的四个层次：满意度、学习、工作绩效和组织影响。通过这些层次的提升可以增强评估的深度，但也需要更多的时间和精力。然而，在所纳入的研究中，针对 PTT 成果的 XR 中的 HITLS 评估主要集中在前两个层次。对工作绩效（第 3 层次）和组织影响（第 4 层次）的评估更为复杂和耗时，但对于确定 PTs 在实际课堂中应用 HITLS 学习的能力以及衡量项目的整体效果至关重要。因此，未来的研究应纳入工作绩效和组织影响的评估，以便对 PTT XR 中的 HITLS 进行更全面的评估。

6. Conclusions and future research
6.结论和未来研究

This meta-analysis presented compelling evidence in favor of the effectiveness of HILTS within XR for PTT, and systematically explored potentially effective instructional design features throughout the stages of planning, implementation, and evaluation phase, respectively. Below is a summary of the key findings:
这项荟萃分析提供了令人信服的证据，证明 XR 中的 HILTS 对 PTT 的有效性，并分别在计划、实施和评估阶段系统地探讨了可能有效的教学设计特点。以下是主要研究结果摘要：

A total of 142 separate effect sizes from 39 studies, encompassing 1564 participants, were collected and analyzed. The findings indicated that HILTS within XR yielded a significant effect size of 0.535 for PTT outcomes.
共收集并分析了 39 项研究的 142 个独立效应大小，涉及 1564 名参与者。研究结果表明，XR 中的 HILTS 对 PTT 结果的显著效应大小为 0.535。
During the planning phase of HILTS within XR, potentially effective instructional design features include setting, autonomy, task variation and simulation length.
在 XR 中的 HILTS 计划阶段，可能有效的教学设计特征包括设置、自主性、任务变化和模拟长度。
During the implementation phase of HILTS within XR, among instructional support features, introduction, theoretical preparation, rehearsal, prompt, static feedback and other scaffolds significantly moderated the effect sizes.
在 XR 中实施 HILTS 阶段，在教学支持功能中，导入、理论准备、演练、提示、静态反馈和其他支架显著调节了效果大小。
During the evaluation phase of HILTS within XR, no moderators were identified, indicating a relatively consistent effect size across different outcome dimensions
在 XR 内部的 HILTS 评估阶段，没有发现调节因素，这表明不同结果维度的效应大小相对一致

Drawing from the findings mentioned above, it appears that that HILTS within XR yielded a medium impact on PTT outcomes. Particular attention should be paid to the 11 moderating effects mentioned above as potentially effective instructional features to consider when implementing HILTS within XR. In addition, while other instructional features did not reach statistical significance, the differences in effect sizes provided valuable insights into how to apply HILTS within XR. Therefore, this study proposed instructional design guidelines for the planning, implementation, and evaluation phases of applying HILTS within XR, respectively.
根据上述研究结果，在 XR 中开展的 HILTS 似乎对 PTT 的结果产生了中等程度的影响。在 XR 中实施 HILTS 时，应特别注意上文提到的 11 种调节效应，这些效应可能是有效的教学特点。此外，虽然其他教学特点没有达到统计学意义，但效果大小的差异为如何在 XR 中应用 HILTS 提供了宝贵的启示。因此，本研究为在 XR 中应用 HILTS 的计划、实施和评估阶段分别提出了教学设计指南。

Nevertheless, this study has some limitations. Firstly, the number of empirical studies on HILTS within XR is
不过，本研究也有一些局限性。首先，在 XR 中对 HILTS 的实证研究数量有限。
limited, with only 39 empirical articles meeting the inclusion criteria. Secondly, small or unbalanced samples of moderating variables for some subcategories may bias the results. Additionally, the variety of measures evaluating PTT outcomes may bias the final results. Finally, some instructional support features coded in the study were simply categorized as yes or no, and a more detailed categorization of these features may be needed for more accurate analysis. Although research on the use of HILTS within XR for PTT is growing, the results of this study reveal that the full potential of this approach is yet to be fully explored. Additionally, when applying HILTS within XR, serious consideration beyond system selection is needed to establish a solid theoretical foundation to guide the scientific implementation of PTT within such virtual simulations.
其次，一些子类别的调节变量样本较小或不平衡，可能会使结果出现偏差。其次，某些子类别的调节变量样本较小或不平衡，可能会使结果产生偏差。此外，评价 PTT 结果的措施多种多样，可能会使最终结果产生偏差。最后，本研究中的一些教学支持特征被简单地归类为 "是 "或 "否"，为了进行更准确的分析，可能需要对这些特征进行更详细的归类。尽管有关在 XR 中使用 HILTS 进行 PTT 的研究越来越多，但本研究的结果表明，这种方法的全部潜力还有待充分挖掘。此外，在 XR 中应用 HILTS 时，除了系统选择之外，还需要认真考虑建立坚实的理论基础，以指导在此类虚拟仿真中科学实施 PTT。

Reference citations 参考引文

Ade-Ojo, G. O., Markowski, M., Essex, R., Stiell, M., & Jameson, J. (2022). A systematic scoping review and textual narrative synthesis of physical and mixed-reality simulation in pre-service teacher training. Journal of Computer Assisted Learning, 38(3), 861-874. https://doi.org/10.1111/jcal . 12653
Ade-Ojo, G. O., Markowski, M., Essex, R., Stiell, M., & Jameson, J. (2022).职前教师培训中的物理和混合现实模拟的系统性范围研究和文字叙述综述。计算机辅助学习期刊》，38（3），861-874。https://doi.org/10.1111/jcal。12653

Al-Sada, M., Yang, P.-C., Chieh Chiu, C., Pradhono Tomo, T., Yamen Saraiji, M., Ogata, T., & Nakajima, T. (2021). From Anime To Reality: Embodying An Anime Character As A Humanoid Robot. Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems, 1-5. https://doi.org/10.1145/3411763.3451543
Al-Sada, M., Yang, P.-C., Chieh Chiu, C., Pradhono Tomo, T., Yamen Saraiji, M., Ogata, T., & Nakajima, T. (2021).从动漫到现实：将动漫人物化身仿人机器人》。2021 年计算机系统中的人为因素 CHI 会议扩展摘要，1-5。https://doi.org/10.1145/3411763.3451543。

*Alban Delamarre, Lisetti, C., & Buche, C. (2020). Modeling Emotions for Training in Immersive Simulations (METIS): A Cross-Platform Virtual Classroom Study. IEEE. https://doi.org/10.1109/ismar-adjunct51615.2020.00036
*Alban Delamarre, Lisetti, C., & Buche, C. (2020)。身临其境模拟培训情绪建模（METIS）：跨平台虚拟课堂研究》。IEEE. https://doi.org/10.1109/ismar-adjunct51615.2020.00036

Alsalamah, A., & Callinan, C. (2021). Adaptation of kirkpatrick's four-level model of training criteria to evaluate training programmes for head teachers. Education Sciences, 11(3), 116. https://doi.org/10.3390/educsci11030116
Alsalamah, A., & Callinan, C. (2021).Adaptation of Kirkpatrick's four-level model of training criteria to evaluate training programmes for head teachers.教育科学》，11（3），116。 https://doi.org/10.3390/educsci11030116

*Anton, S., Piro, J. S., Delcourt, M. A. B., & Gundel, E. (2023). Pre-Service Teachers' Coping and Anxiety within Mixed-Reality Simulations. Social Sciences, 12(3), 146. https://doi.org/10.3390/socsci12030146
*Anton, S., Piro, J. S., Delcourt, M. A. B., & Gundel, E. (2023)。职前教师在混合现实模拟中的应对与焦虑》。https://doi.org/10.3390/socsci12030146

Arjun Nagendran, Remo Pillat, Kavanaugh, A., Welch, G., & Hughes, C. E. (2013). AMITIES. VRST '13: Proceedings of the 19th ACM Symposium on Virtual Reality Software and Technology, 143-152. https://doi.org/10.1145/2503713.2503731
Arjun Nagendran, Remo Pillat, Kavanaugh, A., Welch, G., & Hughes, C. E. (2013)。AMITIES。VRST '13: Proceedings of the 19th ACM Symposium on Virtual Reality Software and Technology, 143-152. https://doi.org/10.1145/2503713.2503731

Bahrick, H. P., & Hall, L. K. (2005). The importance of retrieval failures to long-term retention: A metacognitive explanation of the spacing effect. Journal of Memory and Language, 52(4), 566-577. https://doi.org/10.1016/j.jml.2005.01.012
Bahrick, H. P., & Hall, L. K. (2005).The importance of retrieval failures to long-term retention：A metacognitive explanation of the spacing effect.记忆与语言杂志》，52（4），566-577。https://doi.org/10.1016/j.jml.2005.01.012。

Begg, C. B., & Mazumdar, M. (1994). Operating Characteristics of a Rank Correlation Test for Publication Bias. Biometrics, 50(4), 1088. https://doi.org/10.2307/2533446
Begg, C. B., & Mazumdar, M. (1994).出版偏差等级相关检验的操作特征》。https://doi.org/10.2307/2533446

Billingsley, G. M., & Scheuermann, B. K. (2014). Using virtual technology to enhance field experiences for pre-service special education teachers. Teacher Education and Special Education: The Journal of the Teacher Education Division of the Council for Exceptional Children, 37(3), 255-272. https://doi.org/10.1177/0888406414530413

Bosch, C., & Ellis, T. (2021). Using avatars to address teacher self-efficacy. Journal of Global Education and Research, 5(1), 15-35. https://doi.org/10.5038/2577-509x.5.1.1069

Bossard, C., Kermarrec, G., Buche, C., & Tisseau, J. (2008). Transfer of learning in virtual environments: A new challenge? Virtual Reality, 12(3), 151-161. https://doi.org/10.1007/s10055-008-0093-y

Bowden, A., Traynor, V., Chang, H.-C. R., & Wilson, V. (2022). Beyond the technology: Applying the NLN jeffries simulation theory in the context of aging simulation. Nursing Forum, 57(3), 473-479. https://doi.org/10.1111/nuf.12687

Bradley, E. G., & Kendall, B. (2014). A Review of Computer Simulations in Teacher Education. Journal of Educational Technology Systems, 43(1), 3-12. https://doi.org/10.2190/et.43.1.b

Brantlinger, A., Grant, A. A., Miller, J., Viviani, W., Cooley, L., & Griffin, M. (2020). Maintaining gaps in teacher diversity, preparedness, effectiveness, and retention? A program theory evaluation of mathematics teacher training in the new york city teaching fellows. Educational Policy, 36(6), 089590482095111. https://doi.org/10.1177/0895904820951117

*Brooksbank, A., Piro, J., Straub, C., & Gundel, E. (2023). The Effect of Emotional Awareness Exercises on Adaptiveness in Coping and Empathy with Pre-Service Teachers in Mixed Reality Simulations. International Journal of Integrating Technology in Education, 12(1), 25-43. https://doi.org/10.5121/ijite.2023.12103

Burke, C. J., Tobler, P. N., Baddeley, M., & Schultz, W. (2010). Neural mechanisms of observational learning. Proceedings of the National Academy of Sciences, 107(32), 14431-14436.https://doi.org/10.1073/pnas.100311110

Cambourne, B. L., Kiggins, J., & Ferry, B. (2003). Replacing traditional lectures, tutorials and exams with a Knowledge Building Community (KBC): a constructivist, problem-based approach to pre-service primary teacher education. English Teaching: Practice and Critique, 2(3), 7-21.

Camburn, B. A., Auernhammer, J. M., Sng, K. H. E., Mignone, P. J., Arlitt, R. M., Perez, K. B., Huang, Z., Basnet, S., Blessing, L. T., & Wood, K. L. (2017). Design innovation: A study of integrated practice. Asmedigitalcollection.asme.org; American Society of Mechanical Engineers Digital Collection. https://doi.org/10.1115/DETC2017-68382

Cattaneo, A. A., & Motta, E. (2021). "I reflect, therefore I am... a good professional". On the relationship between reflection-on-action, reflection-in-action and professional performance in vocational education. Vocations and Learning, 14(2), 185-204.https://doi.org/10.1007/s12186-020-09259-9
Cattaneo, A. A., & Motta, E. (2021)。"我反思，所以我是......优秀的专业人员"。职业教育中的行动反思、行动反思与专业表现之间的关系。Vocations and Learning, 14(2), 185-204.https://doi.org/10.1007/s12186-020-09259-9

Chang, H.-Y. (2016). How to augment the learning impact of computer simulations? The designs and effects of interactivity and scaffolding. Interactive Learning Environments, 25(8), 1083-1097. https://doi.org/10.1080/10494820.2016.1250222

*Chen, C. (2022). Immersive virtual reality to train preservice teachers in managing students' challenging behaviours: A pilot study. British Journal of Educational Technology, 53(4). https://doi.org/10.1111/bjet.13181

Chen, X., Hu, Z., & Wang, C. (2024). Empowering education development through AIGC: A systematic literature review. Education and Information Technologies. https://doi.org/10.1007/s10639-024-12549-7

Clark, R. E. (1994). Media will never influence learning. Educational Technology Research and Development, 42(2), 21-29. https://doi.org/10.1007/bf02299088

Cooper, H. (2015). Research synthesis and meta-analysis: A step-by-step approach (5th ed.). Sage publications.

Cohen, J., Wong, V., Krishnamachari, A., & Berlin, R. (2020). Teacher coaching in a simulated environment. Educational Evaluation and Policy Analysis, 42(2), 208-231. https://doi.org/10.3102/0162373720906217

Cox, S., & King, D. (2006). Skill sets: an approach to embed employability in course design. Education + Training, 48(4), 262-274. https://doi.org/10.1108/00400910610671933

Dai, C.-P., Ke, F., Pan, Y., & Liu, Y. (2023). Exploring students' learning support use in digital game-based math learning: A mixed-methods approach using machine learning and multi-cases study. Computers & Education, 194, 104698. https://doi.org/10.1016/j.compedu.2022.104698

*Dai, C. (2023). Enhancing Learning Achievements and Self-Efficacy for Preservice Teachers Using Model-Based Support in Simulation-Based Learning With Artificial Intelligence-Powered Virtual Agents. ProQuest Dissertations & Theses Global. (2868557948).

Dalgarno, B., Gregory, S., Knox, V., & Reiners, T. (2016). Practising teaching using virtual classroom role plays. Australian Journal of Teacher Education, 41(1), 126-154. https://doi.org/10.14221/ajte.2016v41n1.8

Davidson, J. B., Graham, R. B., Beck, S., Marler, R. T., & Fischer, S. L. (2021). Improving human-in-the-loop simulation to optimize soldier-systems integration. Applied Ergonomics, 90, 103267. https://doi.org/10.1016/j.apergo.2020.103267

Delamarre, A., Shernoff, E., Buche, C., Frazier, S., Gabbard, J., & Lisetti, C. (2021). The interactive virtual training for teachers (IVT-T) to practice classroom behavior management. International Journal of Human-Computer Studies, 152. https://doi.org/10.1016/j.ijhcs.2021.102646

Derakhshan, A., Sedigheh Karimpour, & Nazari, M. (2023). Making sense of emotion and identity construction through metaphors: A prompt-based study in an english as a foreign language context. Journal of Multilingual and Multicultural Development, 1-15. https://doi.org/10.1080/01434632.2023.2270951

*DeSantis, W. J., Delcourt, M. A. B., Shore, B. M., & Greenwood, J. C. (2023). Impact of data-driven feedback and coaching on preservice teachers' questioning skills for higher-order thinking within a mixed-reality simulation environment. Education Sciences, 13(6), 596. https://doi.org/10.3390/educsci13060596

Desai, P. R., Desai, P. N., Ajmera, K. D., & Mehta, K. (2014). A review paper on oculus rift-a virtual reality headset. arXiv preprint arXiv:1408.1173.

Dieker, L., Hughes, C., & Hynes, M. (2023). The Past, the Present, and the Future of the Evolution of Mixed Reality in Teacher Education. Education Sciences, 13(11), 1070. https://doi.org/10.3390/educsci13111070

*Dueck, B. S., Pelletier, G. N., Goegan, L. D., & Daniels, L. M. (2021). A new avenue for teacher education: Virtual reality intervention as a path towards reducing public speaking anxiety and supporting growth mindsets.

https://doi.org/10.7939/r3-pgmz-kq24

Dufrene, C., & Young, A. (2014). Successful debriefing - Best methods to achieve positive learning outcomes: A literature review. Nurse Education Today, 34(3), 372-376. https://doi.org/10.1016/j.nedt.2013.06.026

Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315(7109), 629-634. https://doi.org/10.1136/bmj.315.7109.629

Ellis, N. J., & Loughland, T. (2017). 'Where to next?': Examining feedback received by teacher education students. Issues in Educational Research, 27(1), 51-63.

*Ely, E., Alves, K. D., Dolenc, N. R., Sebolt, S., & Walton, E. A. (2018). Classroom Simulation to Prepare Teachers to Use Evidence-Based Comprehension Practices. Journal of Digital Learning in Teacher Education, 34(2), 71-87. https://doi.org/10.1080/21532974.2017.1399487

Folds, D. J. (2015). Human in the loop simulation. Simulation Foundations, Methods and Applications, 175-183. https://doi.org/10.1007/978-1-4471-5634-5_15

*Ghazarian, P. G., Kormos, E., & Wisdom, K. (2022). Simulated field experiences and culturally responsive pedagogy among teacher candidates. Intercultural Education, 1-17. https://doi.org/10.1080/14675986.2022.2144073

Glass, G. V. (1976). Primary, Secondary, and Meta-Analysis of Research. Educational Researcher, 5(10), 3.

*Goff, W. (2023). VR Simulations to Develop Teaching Practice with Pre-service Teachers. 13-24.

https://doi.org/10.1007/978-981-99-2785-2_2

Goldstein, L. S. (2005). Becoming a teacher as a hero's journey: Using metaphor in preservice teacher education. Teacher Education Quarterly, 32(1), 7-24.http://www.jstor.org/stable/23478686

Goodman, J. (1986). Making Early Field Experience Meaningful: a critical approach. Journal of Education for Teaching, 12(2), 109-125. https://doi.org/10.1080/0260747860120201

Greer, R. D., Dudek-Singer, J., & Gautreaux, G. (2020). Observational learning. In Behavior Analysis Around the World (pp. 486-499). Psychology Press.

Groom, V., Nass, C., Chen, T., Nielsen, A., Scarborough, J. K., & Robles, E. (2009). Evaluating the effects of behavioral realism in embodied agents. International Journal of Human-Computer Studies, 67(10), 842-849.

https://doi.org/10.1016/j.ijhcs.2009.07.001

*Gundel, E. M. (2018). Understanding the Effects of Mixed Reality Simulation on Pre-Service Teacher Self-Efficacy (Order No. 10840405). Available from ProQuest Dissertations & Theses Global. (2075116364).

https://www.proquest.com/dissertations-theses/understanding-effects-mixed-reality-simulation-on/docview/2075116364/se-2

Harzing, A.-W., & Alakangas, S. (2015). Google Scholar, Scopus and the Web of Science: a longitudinal and cross-disciplinary comparison. Scientometrics, 106(2), 787-804. https://doi.org/10.1007/s11192-015-1798-9

Haverly, C., & Davis, E. A. (2023). Unpacking readiness for elementary science teaching: what preservice teachers bring and how that can be shaped through teacher education. Studies in Science Education, 60(1), 75-119.

https://doi.org/10.1080/03057267.2023.2188703

Hedges, L. V., & Olkin, I. (1981). The asymptotic distribution of commonality components. Psychometrika, 46(3), 331-336. https://doi.org/10.1007/bf02293739

Higgins, J. P. T., Thompson, S. G., Deeks, J. J., & Altman, D. G. (2003). Measuring inconsistency in meta-analyses. BMJ, 327(7414), 557-560. https://doi.org/10.1136/bmj.327.7414.557

Hindmarsh, J., Hyland, L., & Banerjee, A. (2014). Work to make simulation work: "Realism", instructional correction and the body in training. Discourse Studies, 16(2), 247-269. https://doi.org/10.1177/1461445613514670

*Hirsch, S. E., Griffith, C. A., Kelley, M. H., Carlson, A., & McKown, G. (2023). Increasing Preservice Teachers' Knowledge and Skills Using a Mixed-Reality Simulator: The Tale of Two Studies. Teacher Education and Special Education, 46(3), 204-222. https://doi.org/10.1177/08884064231154028

Hodges, N. J., Williams, A. M., Hayes, S. J., & Breslin, G. (2007). What is modelled during observational learning? Journal of Sports Sciences, 25(5), 531-545. https://doi.org/10.1080/02640410600946860

*Huang, Y., Richter, E., Thilo Kleickmann, & Richter, D. (2021). Virtual reality in teacher education from 2010 to 2020: A review of program implementation, intended outcomes, and effectiveness measures. EdArXiv (OSF Preprints). https://doi.org/10.35542/osf.io/ye6uw

*Hudson, M. E., Voytecki, K. S., Owens, T. L., & Zhang, G. (2019). Preservice teacher experiences implementing classroom management practices through mixed-reality simulations. Rural Special Education Quarterly, 38(2), 79-94.

https://doi.org/10.1177/8756870519841421

Hudson, S., Matson-Barkat, S., Pallamin, N., & Jegou, G. (2019). With or without you? Interaction and immersion in a virtual reality experience. Journal of Business Research, 100, 459-468. https://doi.org/10.1016/j.jbusres.2018.10.062

Hughes, C. E., Arjun Nagendran, Dieker, L., Hynes, M. J., & Welch, G. C. (2015). Applications of Avatar Mediated Interaction to Teaching, Training, Job Skills and Wellness. Virtual Realities, 8844, 133-146.

https://doi.org/10.1007/978-3-319-17043-5_8

Jafar, H., Muda, I., Zainal, A., & Yasin, W . (2010). Profit maximization theory, survival-based theory, and contingency theory: A review on several underlying research theories of corporate turnaround. Jurnal Ekonom, 13(4), 136-143.

Jeffries, P. R., & National League For Nursing. (2016). The NLN jeffries simulation theory. Wolters Kluwer. Kaufman, D., & Ireland, A. (2016). Enhancing Teacher Education with Simulations. TechTrends, 60(3), 260-267. https://doi.org/10.1007/s11528-016-0049-0

Johnston, J. L., Kearney, G. P., Gormley, G. J., & Reid, H. (2020). Into the uncanny valley: Simulation versus simulacrum? Medical Education, 54(10), 903-907. https://doi.org/10.1111/medu. 14184

*Ke, F., Lee, S., & Xu, X. (2016). Teaching training in a mixed-reality integrated learning environment. Computers in Human Behavior, 62, 212-220. https://doi.org/10.1016/j.chb.2016.03.094

Khan, R., Plahouras, J., Johnston, B. C., Scaffidi, M. A., Grover, S. C., & Walsh, C. M. (2018). Virtual reality simulation training for health professions trainees in gastrointestinal endoscopy. Cochrane Database of Systematic Reviews. https://doi.org/10.1002/14651858.cd008237.pub3

*KIM, K., PARK, S., RYU, J., & LIM, T. (2023). Scenario Usefulness and Avatar Realism in an Augmented Reality-based Classroom Simulation for PTT. Educational Technology International, 24(1), 1-27.

https://doi.org/10.23095/ETI.2023.24.1.001

*King, S., Boyer, J., Bell, T., & Estapa, A. (2022). An automated virtual reality training system for teacher-student interaction: A randomized controlled trial. JMIR Serious Games, 10(4), e41097. https://doi.org/10.2196/41097

Kirkpatrick, J. D., & Kirkpatrick, W. K. (2016). Kirkpatrick's four levels of training evaluation. Association for Talent Development.

Kraft, M. A., & Blazar, D. (2017). Individualized coaching to improve teacher practice across grades and subjects: New experimental evidence. Educational Policy, 31(7), 1033-1068. https://doi.org/10.1177/0895904816631099

Kucek, S., & Leitner, M. (2020). Training the Human-in-the-Loop in Industrial Cyber Ranges. Lecture Notes in Electrical Engineering, 107-118. https://doi.org/10.1007/978-3-030-48602-0_10

Lamb, R., & Etopio, E. A. (2020). Virtual Reality: a Tool for Preservice Science Teachers to Put Theory into Practice. Journal of Science Education and Technology, 29(4), 573-585. https://doi.org/10.1007/s10956-020-09837-5

*Landon-Hays, M., Peterson-Ahmad, M. B., & Frazier, A. D. (2020). Learning to teach: How a simulated learning environment can connect theory to practice in general and special education educator preparation programs. Education Sciences, 10(7), 184. https://doi.org/10.3390/educsci10070184

*Larson, K. E., Hirsch, S. E., McGraw, J. P., & Bradshaw, C. P. (2019). Preparing preservice teachers to manage behavior problems in the classroom: The feasibility and acceptability of using a mixed-reality simulator. Journal of Special Education Technology, 35(2), 63-75. https://doi.org/10.1177/0162643419836415

*Lee, H., & Hwang, Y. (2022). Technology-Enhanced Education through VR-Making and Metaverse-Linking to Foster Teacher Readiness and Sustainable Learning. Sustainability, 14(8), 4786. https://doi.org/10.3390/su14084786

Levin, O., & Flavian, H. (2020). Simulation-based learning in the context of peer learning from the perspective of preservice teachers: a case study. European Journal of Teacher Education, 45(3), 373-394.

https://doi.org/10.1080/02619768.2020.1827391

Lindberg, S., & Jönsson, A. (2023). Preservice Teachers Training with Avatars: A Systematic Literature Review of "Human-in-the-Loop" Simulations in Teacher Education and Special Education. Education Sciences, 13(8), 817. https://doi.org/10.3390/educsci13080817

Lu, J.-J., Zhou, M.-M., Jiang, B.-H., Zhang, M.-M., & Pan, Z.-G. (2023). The impact of MR on preservice teachers' creativity and performance in instructive task designs. International Journal of Human-Computer Interaction, 1-13. https://doi.org/10.1080/10447318.2023.2283643

*Luke, S. E., Ford, D., Vaughn, M., & Fulchini-Scruggs, A. (2021). Using Mixed Reality Simulation and Roleplay to Develop Preservice Teachers' Metacognitive Awareness. Journal of Technology and Teacher Education, 29(3), 389-413. https://www.learntechlib.org/p/219543/

*Luqrirr Jean-Luc, Oberdorter, S., Marc Erich Latoschik, Wittmann, A., Seufert, C., & Grafe, S. (2018). VR-Assisted vs Video-Assisted Teacher Training. https://doi.org/10.1109/vr.2018.8446312

Matthias Scheutz, Williams, T. N., Krause, E., Oosterveld, B., Sarathy, V., & Frasca, T. M. (2019). An Overview of the Distributed Integrated Cognition Affect and Reflection DIARC Architecture. 165-193.

https://doi.org/10.1007/978-3-319-97550-4_11

Martín-Martín, A., Orduna-Malea, E., Thelwall, M., & Delgado López-Cózar, E. (2018). Google Scholar, Web of Science, and Scopus: A systematic comparison of citations in 252 subject categories. Journal of Informetrics, 12(4), 1160-1177. https://doi.org/10.1016/j.joi.2018.09.002

Merriam, S. B. (2018). Adult learning theory: Evolution and future directions. In Contemporary theories of learning (pp. 83-96). Routledge.

McGarr, O. (2020). The use of virtual simulations in teacher education to develop pre-service teachers' behaviour and classroom management skills: implications for reflective practice. Journal of Education for Teaching, 47(2), 274-286. https://doi.org/10.1080/02607476.2020.1733398

*McKown, G., Hirsch, S. E., Carlson, A., Allen, A. A., & Walters, S. (2021). Preservice special education teachers' perceptions of mixed-reality simulation experiences. Journal of Digital Learning in Teacher Education, 1-16. https://doi.org/10.1080/21532974.2021.1995796

Mulvaney, P., Rooney, B., Friehs, M. A., & John Francis Leader. (2024). Social VR design features and experiential outcomes: narrative review and relationship map for dyadic agent conversations. Virtual Reality, 28(1).

https://doi.org/10.1007/s10055-024-00941-0

Mena, J., Odiel Estrada Molina, & Pérez-Calvo, E. (2023). Teachers' Professional Training through Augmented Reality: A Literature Review. Education Sciences, 13(5), 517-517. https://doi.org/10.3390/educsci13050517

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2010). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. International Journal of Surgery, 8(5), 336-341. sciencedirect.

https://doi.org/10.1016/j.ijsu.2010.02.007

Minott, M. (2021). A case for a reflective approach to teaching 'practicum debriefing' with implications for teacher educators and education. Teacher Education Advancement Network Journal, 13(1), 73-85. https://ueaeprints.uea.ac.uk/id/eprint/81935

Niebaum, J. C., & Munakata, Y. (2022). Why Doesn't Executive Function Training Improve Academic Achievement? Rethinking Individual Differences, Relevance, and Engagement from a Contextual Framework. Journal of Cognition and Development, 1-19. https://doi.org/10.1080/15248372.2022.2160723

Nussli, N., & Oh, K. (2014). The components of effective teacher training in the use of three-dimensional immersive virtual worlds for learning and instruction purposes: A literature review. The Journal of Technology and Teacher Education, 22(2), 213-241.

*Nussli, N., Oh, K., & McCandless, K. (2014). Collaborative science learning in three-dimensional immersive virtual worlds: Pre-service teachers' experiences in second life. Journal of Educational Multimedia and Hypermedia, 23(3), 253-284.

Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., & Moher, D. (2021). Updating guidance for reporting systematic reviews: development of the PRISMA 2020 statement. Journal of Clinical Epidemiology, 134(134), 103-112.

https://doi.org/10.1016/j.jclinepi.2021.02.003

*Pellas, N., & Boumpa, A. (2015). Open Sim and Sloodle Integration for Preservice Foreign Language Teachers' Continuing Professional Development. Journal of Educational Computing Research, 54(3), 407-440.

https://doi.org/10.1177/0735633115615589

*Peterson-Ahmad, M. B., Keeley, R., & Frazier, A. (2023). Using Mixed Reality to Support Inclusive Teaching Strategies in General and Special Education Preparation Programs. Social Sciences, 12(11), 596-596.

https://doi.org/10.3390/socsci12110596

*Putman, S. (2021). Immersive virtual reality and education: a study into the effectiveness of using this technology with preservice teachers. https://doi.org/10.18297/etd/3728

Putman, S. M. (2012). Investigating teacher efficacy: Comparing preservice and inservice teachers with different levels of experience. Action in Teacher Education, 34(1), 26-40. https://doi.org/10.1080/01626620.2012.642285

Radu-Daniel Vatavu. (2023). From natural to non-natural interaction: Embracing interaction design beyond the accepted convention of natural. INTERNATIONAL CONFERENCE on MULTIMODAL INTERACTION.

https://doi.org/10.1145/3577190.3616122

Rayner, C., & Fluck, A. (2014). Pre-service teachers' perceptions of simSchool as preparation for inclusive education: a pilot study. Asia-Pacific Journal of Teacher Education, 42(3), 212-227. https://doi.org/10.1080/1359866x.2014.927825

Reiser, R. A. (2001). A history of instructional design and technology: Part II: A history of instructional design. Educational Technology Research and Development, 49(2), 57-67. https://doi.org/10.1007/bf02504928

*Remacle, A., Bouchard, S., & Morsomme, D. (2023). Can teaching simulations in a virtual classroom help trainee teachers to develop oral communication skills and self-efficacy? A randomized controlled trial. Computers & Education, 200, 104808-104808. https://doi.org/10.1016/j.compedu.2023.104808

*Richter, E., Hußner, I., Huang, Y., Richter, D., & Lazarides, R. (2022). Video-based reflection in teacher education: Comparing virtual reality and real classroom videos. Computers & Education, 190, 104601. https://doi.org/10.1016/j.compedu.2022.104601

*Robbins, S. H., Gilbert, K. A., Chumney, F. L., & Green, K. B. (2019). The effects of immersive simulation on targeted collaboration skills among undergraduates in special education. Teaching and Learning Inquiry, 7(2), 168-185. https://doi.org/10.20343/teachlearninqu.7.2.11

*Rosati-Peterson, G. L., Piro, J. S., Straub, C., & O’Callaghan, C. (2021). A Nonverbal Immediacy Treatment with Pre-Service Teachers Using Mixed Reality Simulations. Cogent Education, 8(1), 1882114. https://doi.org/10.1080/2331186x.2021.1882114

Rosenthal, R. (1995). Writing meta-analytic reviews. Psychological Bulletin, 118(2), 183-192.

https://doi.org/10.1037/0033-2909.118.2.183

Sánchez-Caballé, A., Esteve-Mon, F. M., & González-Martínez, J. (2020). What to expect when you are simulating?: About digital simulation potentialities in teacher training. International Journal of Online Pedagogy and Course Design (IJOPCD), 10(1), 34-47.

Sargent, D. L. (2020). The evolution of teacher education through emerging technologies of mixed reality simulation. In Next generation digital tools and applications for teaching and learning enhancement (pp. 16-30). IGI Global.

Stanney, K. M., Skinner, A., & Hughes, C. (2023). Exercisable learning-theory and evidence-based andragogy for training effectiveness using XR (ELEVATE-XR): Elevating the ROI of immersive technologies. International Journal of HumanComputer Interaction, 39(11), 2177-2198. https://doi.org/10.1080/10447318.2023.2188529

Satrio Pradono Suryodiningrat, Arief Ramadhan, Harjanto Prabowo, Harry Budi Santoso, & Hirashima, T. (2023). Mixed reality systems in education: a systematic literature review. Journal of Computers in Education.

https://doi.org/10.1007/s40692-023-00281-z

*Seufert, C., Oberdörfer, S., Roth, A., Grafe, S., Lugrin, J.-L., & Latoschik, M. E. (2022). Classroom management competency enhancement for student teachers using a fully immersive virtual classroom. Computers & Education, 179, 104410. https://doi.org/10.1016/j.compedu.2021.104410

Smith, J. L., Wagaman, J., & Handley, I. M. (2009). Keeping it dull or making it fun: Task variation as a function of promotion versus prevention focus. Motivation and Emotion, 33(2), 150-160. https://doi.org/10.1007/s11031-008-9118-9

*Stavroulia, K. E., & Lanitis, A. (2023). The role of perspective-taking on empowering the empathetic behavior of educators in VR-based training sessions: An experimental evaluation. Computers & Education, 197, 104739.

https://doi.org/10.1016/j.compedu.2023.104739

Spencer, S., Drescher, T., Sears, J., Scruggs, A. F., & Schreffler, J. (2019). Comparing the Efficacy of Virtual Simulation to Traditional Classroom Role-Play. Journal of Educational Computing Research, 57(7), 1772-1785. https://doi.org/10.1177/0735633119855613

Tam, M. (2000). Constructivism, instructional design, and technology: Implications for transforming distance learning. Journal of Educational Technology & Society, 3(2), 50-60.

Taylor, J., Macpherson, T., Spears, I., & Weston, M. (2015). The Effects of Repeated-Sprint Training on Field-Based Fitness Measures: A Meta-Analysis of Controlled and Non-Controlled Trials. Sports Medicine, 45(6), 881-891. https://doi.org/10.1007/s40279-015-0324-9

Theelen, H., van den Beemt, A., & Brok, P. den. (2019). Classroom simulations in teacher education to support preservice teachers' interpersonal competence: A systematic literature review. Computers & Education, 129, 14-26.

https://doi.org/10.1016/j.compedu.2018.10.015

Tilson, J., Sandretto, S., & Pratt, K. (2017). Connecting theory to practice: Using preservice teachers' beliefs, theories and video-recorded teaching to prompt a cycle of praxis. Teaching and Teacher Education, 67, 454-463.

https://doi.org/10.1016/j.tate.2017.07.012

Wages, R., Grünvogel, S. M., & Benno Grützmacher. (2004). How realistic is realism? Considerations on the aesthetics of computer games. Lecture Notes in Computer Science, 3166, 216-225. https://doi.org/10.1007/978-3-540-28643-1_28

*Walters, S. M., Hirsch, S. E., McKown, G., Carlson, A., & Allen, A. A. (2021). Mixed-Reality Simulation With Preservice Teacher Candidates: A Conceptual Replication. Teacher Education and Special Education: The Journal of the Teacher Education Division of the Council for Exceptional Children, 088840642110014. https://doi.org/10.1177/08884064211001453

*Walters, S. M. (2021). Clarifying Vocabulary in a Mixed-Reality Classroom: Coaching Preservice Teachers (Order No. 28648946). Available from ProQuest Dissertations & Theses Global. (2827285716).

https://www.proquest.com/dissertations-theses/clarifying-vocabulary-mixed-reality-classroom/docview/2827285716/se-2

Wang, C., Chen, X., Yu, T., Liu, Y., & Jing, Y. (2024). Education reform and change driven by digital technology: a bibliometric study from a global perspective. Humanities and Social Sciences Communications, 11(1), 1-17.

https://doi.org/10.1057/s41599-024-02717-y

Wilson, M. L., Ritzhaupt, A. D., & Cheng, L. (2020). The impact of teacher education courses for technology integration on pre-service teacher knowledge: A meta-analysis study. Computers & Education, 156, 103941.

https://doi.org/10.1016/j.compedu.2020.103941

Wilson, R. D., & Hagler, D. (2012). Through the lens of instructional design: Appraisal of the jeffries/national league for nursing simulation framework for use in acute care. The Journal of Continuing Education in Nursing, 43(9), 428-432. https://doi.org/10.3928/00220124-20120615-27

Yeo, J., & Jang, M. S. (2022). Nursing students' Self-directed Learning Experiences in Web-based Virtual simulation: a Qualitative Study. Japan Journal of Nursing Science, 20(2), e12514. https://doi.org/10.1111/jjns . 12514

Yilmaz, O., & Hebebci, M. T. (2022). The use of virtual environments and simulation in teacher training. International Journal on Social and Education Sciences (IJonSES), 4(3), 446-457. https://doi.org/10.46328/ijonses . 376

Zeichner, K., & Liu, K. Y. (2010). A critical analysis of reflection as a goal for teacher education. Handbook of reflection and reflective inquiry: Mapping a way of knowing for professional reflective inquiry, 67-84.

https://doi.org/10.1007/978-0-387-85744-2_4

*Zhang, J., Pan, Q., Zhang, D., Meng, B., & Hwang, G.-J. (2024). Effects of Virtual Reality Based Microteaching Training on Pre-Service Teachers' Teaching Skills from a Multi-Dimensional Perspective. Journal of Educational Computing Research. https://doi.org/10.1177/07356331231226179

APPENDIX A

Avatar-mediated interactive training system

environment

Immersion level

Semi-immersive

Non-immersive

Fully-immersive

Non-immersive

System

interactivity

Technological

requirements

Large screen

Mobile

Desktop, HMD,

mouse

Desktop, HMD,

motion tracking

device

Desktop, HMD,

controller

Desktop, mouse

Environment

fidelity

Avatar/agent

graphics

3D lifelike

Avatar/agent status

Middle school

students

Early adolescence

Primary school

student

Middle school

students

Avatar/agent

More than 25

Behavioral

fidelity

Behavior realism

WOz-based

behaviors

None

Scenario-based

behaviors

Model-based

behavior

Model-based

behavior

Scenario-based

behaviors

Affect realism

WOz-based facial

expression and

gesture expression

None

Scenario-based

facial expression

and gesture

expression

None

Scenario-based

facial expression

and gesture

expression

Avatar/agent

activity

Interaction activity,

goal-oriented

activity

Goal-oriented

activity

Goal-oriented

activity

Interaction

activity,

goal-oriented

activity,

internal activity

Interaction activity,

goal-oriented

activity

Goal-oriented

activity

Similar

systems

TeachLivE

(Putman, 2021)

3BVR (Chen,

2022); Virtual

Classroom

(Remacle et al.,

2023)

Not found.

Etopio, 2020)

Open Sim (Pellas

Boumpa, 2016);

Second life (Nussli

et al., 2014);

simSchool (Rayner

Fluck, 2014)

APPENDIX B

Review studies of teacher education within XR simulations

Studies

Years covered

Focus

Results of the study

Instructional

design suggestions

Bradley and

Kendall (2014)

Current teacher

education

simulation systems.

The characteristics of

different simulation

systems were analyzed

Not mentioned.

Ade-Ojo et al

2020 年 3 月至 2020 年 12 月

March,

2020-Decembe

r, 2020

Pre-service teacher

training with

simulations

including physical

and mixed-reality

types.

Simulation, including

physical and

mixed-reality types,

could be used as a tool

to increase confidence,

self-efficacy, classroom

management skills and

communication.

Reflection;

technical support;

collaboration.

Theelen et al.

Preservice teacher

education with

computer-based

classroom

simulations (within

MR, VR

environments).

Simulations are

promising tools to

contribute to improving

PSTs' teaching skills.

Co-assessment,

receiving teacher

feedback, available

resources, thinking

spaces, active

learning, the ability

to take the role of a

teacher, peer

observation,

discussions, and

collaboration

Dieker et al.

Teacher education

with MR

simulation.

Current simulations are

mostly made by teacher

educators, with

feedback playing a key

role in training

outcomes and ensuring a

safe, reflective space for

teacher training. Future

developments in MR

aim to incorporate AI

agents and advanced

data collection.

Reflection,

collaboration,

co-teaching,

repeated training,

preparedness.

Sánchez-Caballé

et al. (2020)

Teacher education

with digital

simulations.

Digital simulations are

effective in teacher

training and there are

technological limitations

and shortcomings.

Not mentioned.

Bondie and Dede

Teacher education

Suggested promising

Concluded three

(2021)

with digital

puppeteering

(avatars) within

MR environments.

results of teacher

learning through mixed

reality simulations and

concluded three

learning design models

and five design

principles

modes of

instructional

designs, and the

features including

orientation,

coaching or course

activities, debrief

activities, repeated

training.

Mena et al. (2023)

Teacher education

with AR

simulation.

The benefits of

integrating (AR) into

teacher education are

primarily at a theoretical

level.

Not mentioned.

Sargent

Teacher education

with MR

simulation.

MR simulation is an

effective tool for teacher

education. When

coupled with modeling,

feedback, and

self-reflection, it

becomes a more potent

method for preparing

teachers.

Rehearsal,

feedback,

self-reflection,

repeated practice,

refinement.

Nussli and Oh

Teacher education

with 3D immersive

virtual worlds

(mainly within VR

environments).

3D immersive virtual

worlds are significant

educational platforms

for teacher education,

yet they still face certain

challenges, such as

technical issues, being

perceived as games, and

lacking structure and

support.

Synchronicity,

scaffolding, prior

experience,

stimulating spaces,

technical support.

Billingsley and

Scheuermann

(2014)

Teacher education

with virtual

technologies

(mainly within VR

environments).

Virtual technologies can

greatly improve

preservice teacher

education.

Not mentioned.

Lindberg and

Jönsson (2023)

virtual

"human-in-the-loop

" simulations

(avatar-mediated

interactive training)

for preservice

Preservice teachers

generally have a

positive experience with

virtual simulations, but

technical limitations

pose significant

Not mentioned.

teachers (mainly

within MR

disadvantages.

environments).

APPENDIX C

The characteristics of publications included

Author (year)

Publication characteristics

Publication

type

XR type

HITLS type

Training outcome

Huang et al.

(2023)

Journal

Agent-mediated

Attitude

Lee & Hwang

(2022)

Journal

Agent-mediated

Knowledge, satisfaction

Seufert et al.

(2022)

Journal

Avatar-mediated

Skill, satisfaction

King et al.

(2022)

Journal

Avatar-mediated

Skill, knowledge,

attitude, satisfaction

Remacle et al.

(2023)

Journal

Avatar-mediated

Attitude

Chen (2022)

Journal

Avatar-mediated

Skill

Dueck (2021)

Dissertation

Agent-mediated

Attitude

Dai (2023)

Dissertation

Agent-mediated

Skill, attitude

Putman (2021)

Dissertation

Agent-mediated

Skill, knowledge

Boumpa (2016)

135

Journal

Agent-mediated

Attitude 态度

Oh & Nussli

(2014)

Journal 期刊

Agent-mediated 代理中介

Satisfaction 满意度

哈德逊（2019）

Hudson

(2019)

Journal 期刊

Avatar-mediated 以 "阿凡达 "为媒介

Satisfaction, skill 满意度、技能

Emily Ely

(2018)

Journal

Avatar-mediated

Knowledge

Sally Spencer

(2019)

Journal

Avatar-mediated

Skill, satisfaction

Sharon (2021)

Journal

Avatar-mediated

Skill, attitude,

satisfaction

Larson et al.

(2019)

Journal

Avatar-mediated

Attitude, satisfaction,

attitude

Hirsch et al.

(2023)

Journal

Avatar-mediated

Knowledge, attitude,

satisfaction, skill

Ke et al. (2016)

Journal

Avatar-mediated

Satisfaction, attitude

DeSantis et al.

Journal

Agent-mediated

Attitude

Landon-Hays

et al. (2020)

Journal

Avatar-mediated

Attitude

Kukhyeon et

al. (2023)

Journal

Agent-mediated

Satisfaction

Maria B et al.

(2023)

Journal

Avatar-mediated

Attitude

Ashley

Brooksbank

(2023)

Journal

Avatar-mediated

Skill

Robbins (2019)

Journal

Avatar-mediated

Skill

Delamarre

Conference

Agent-mediated

Satisfaction, attitude

Lugrin (2018)

Conference

Avatar-mediated

Satisfaction

Bosch & Ellis

(2021)

Journal

Avatar-mediated

Attitude

Wendy Goff

(2023)

238

Book

chapter

Avatar-mediated

Attitude

Gundel (2018)

Dissertation

Avatar-mediated 以 "阿凡达 "为媒介

Attitude 态度

Lanitis ( 2023) 拉尼蒂斯 ( 2023)

Journal 期刊

Agent-mediated 代理中介

Skill 技能

安东等人（2023 年）

Anton et

al.(2023)

Journal 期刊

Avatar-mediated 以 "阿凡达 "为媒介

Skill, attitude 技能、态度

加扎里安等人（2022 年）

Ghazarian et al.

(2022)

Journal

Avatar-mediated

Skill

Richter et

al.(2022)

Journal

Agent-mediated

Attitude

McKown et

al.(2021)

Journal

Avatar-mediated

Attitude, satisfaction

Rosati-Peterson

et al. (2021)

Journal

Avatar-mediated

Skill

DeSantis

(2018)

Dissertation

Avatar-mediated

Attitude

Luke et

al.(2021)

Journal

Avatar-mediated

Skill

Walters (2021)

Dissertation

Avatar-mediated

Skill

Zhang et

al.(2024)

Journal

Agent-mediated

Attitude, skill

Note. Abbreviations:

, control group;

, experimental group;

, no data reported.

APPENDIX D

Frequencies and percentages for the coded variables

Phase 阶段

Dimension 尺寸

Code

Subcategory

经常

Frequen

百分比

Percenta

Design 设计

Context 背景

Setting

Classroom

Other offline

places

Online meeting

room

Blended

Background

Curriculum integrati

Yes

Design

Technology design

XR type

Immersion

level

Non-immersive

Semi-immersiv

Fully

immersive

Interactivit

Non-natural

Natural

Both

24.65%

Avatar-mediate

108

HIRLS type

Agent-mediate

Autonomy

Fully

preparation

Rehearsal

Yes

At the

in-situ

stage

Prompt

Yes

At the

post

trainin

Reflection

Yes

stage

Static

feedback

Yes

Other

scaffolds

Yes

47.18%

Evaluatio

Outcome

Satisfaction

Yes

103

Attitude

Yes

Knowledge

Yes

134

Skill

Yes

APPENDIX E

The analysis results for moderator variables
调节变量的分析结果

Code 代码
(1) Setting (1) 设置				9.087	3	0.028
Classroom 教室	21	0.175	0.986
Other offline places 其他线下场所	64	0.568	5.669
Online meeting room 在线会议室	23	0.307	1.318
Blended 混合型	34	0.836	5.498

(2)Curriculum integration (2) 课程整合
Yes	94	0.604	7.622
No	48	0.378	2.151
(3) XR type (3) XR 型				0.106	2	0.948

MR
VR
AR
(4) Immersion level
Fully immersive
Non-immersive
Semi-immersive
(5) Interactivity
Non-natural interaction
Natural interaction
Both

(6)HIRLS type

Avatar-mediated interactive

training

Agent-medicated interactive training

(7) Autonomy

Fully controlled

Semi-autonomous

Fully autonomous

(8) Customized task

Yes

(9) Task type

Integrated practice

Distributed practice

(10)Task variation

Yes

(11) Simulation length

Shorter than 30 minutes

Longer than 30 minutes

(12) Repeated training

Yes

(13) Training mode

Individually

Collaboratively

(14)Assessment approach

During training

After training

(15) Facilitator role

Faculty teacher only

67	0.458	3.965
75	0.600	6.815

66	0.338	4.440
76	0.741	5.194

2.633

0.105

Researcher only

0.643

5.561

Faculty teacher

researcher

0.432

-1213

Faculty teacher & technical

specialists

0.510

3.337

(16) Participant role

1.472

0.479

Participant only

0.462

4.195

Observer only

0.300

0.772

Both

0.613

6.469

(17)Introduction

7.044

0.008

Yes

119

0.596

7.135

0.296

3.897

(18)Theoretical preparation

3.934

0.047

Yes

123

0.602

7.894

0.086

0.344

(19) Rehearsal

19.746

0.000

Yes

1.109

5.815

0.211

3.145

(20) Prompt

16.889

0.000

Yes

0.818

8.467

0.219

2.014

(21)Reflection

0.251

0.616

Yes

0.489

7.323

0.584

3.279

(22) Static feedback

8.414

0.004

Yes

0.749

5.816

0.305

3.686

(23)Other scaffolds

7.502

0.006

Yes

0.723

6.842

0.330

3.409

(24) Outcome (24) 成果

0.052

0.997

Satisfaction 满意度

0.524

5.712

Attitude 态度

0.510

3.713

Knowledge 知识

0.635

1.171

Skill

0.521

4.408

*Bondie, R., Mancenido, Z., & Dede, C. (2021). Interaction principles for digital puppeteering to promote teacher learning. Journal of Research on Technology in Education, 53(1), 107-123. https://doi.org/10.1080/15391523.2020.1823284

Human-in-the-Loop Simulations Within Extended Reality for Preservice 用于职前教育的扩展现实中的人环模拟

Teacher Training: A Meta-Analysis and Instructional Design Guidelines教师培训：元分析与教学设计指南

Abstract 摘要

1. Introduction 1.导言

2. Literature review 2.文献综述

2.1. Current HITLS systems within XD for PTT2.1.目前在 XD 内用于 PTT 的 HITLS 系统

2.2. Previous reviews and meta-analysis2.2.以往的综述和荟萃分析

3. Method 3.方法

3.1. Literature search strategy3.1.文献检索策略

3.2. Inclusion and exclusion criteria3.2.纳入和排除标准

3.3. Coding 3.3.编码

3.3.1. Coding scheme of context features3.3.1.语境特征编码方案

3.3.2. Coding scheme of background features3.3.2.背景特征编码方案

3.3.3. Coding scheme of design features3.3.3.设计特征的编码方案

3.3.4. Coding scheme of simulation experience features3.3.4.模拟体验特征的编码方案

3.3.5. Coding scheme of outcome features3.3.5.结果特征编码方案

2.4. Data analysis 2.4.数据分析

4. Results 4.成果

4.1. Descriptive statistics4.1.描述性统计

4.2. Overall effect on PTT4.2.对 PTT 的总体影响

4.3.1. Moderator effects during the planning phase4.3.1.规划阶段的主持人效应

4.3.2. Moderator effects during the implementation phase4.3.2.实施阶段的主持人效应

4.3.3. Moderator effects during the evaluation phase4.3.3.评估阶段的主持人效应

4.4. Publication bias 4.4.出版偏差

5. Recommendations on instructional design guidelines for HILTS within XR5.关于 XR 系统内 HILTS 教学设计指南的建议

5.1. Instructional design guidelines during the planning phase5.1.规划阶段的教学设计指南

5.2. Instructional design guidelines during the implementation phase5.2.实施阶段的教学设计指南

5.3. Instructional design guidelines during the evaluation phase5.3.评估阶段的教学设计指南

6. Conclusions and future research6.结论和未来研究

Reference citations 参考引文

APPENDIX A

Avatar-mediated interactive training system

APPENDIX B

APPENDIX C

APPENDIX D

APPENDIX E

MR VR AR (4) Immersion level Fully immersive Non-immersive Semi-immersive (5) Interactivity Non-natural interaction Natural interaction Both

Human-in-the-Loop Simulations Within Extended Reality for Preservice
用于职前教育的扩展现实中的人环模拟

Teacher Training: A Meta-Analysis and Instructional Design Guidelines
教师培训：元分析与教学设计指南

2.1. Current HITLS systems within XD for PTT
2.1.目前在 XD 内用于 PTT 的 HITLS 系统

2.2. Previous reviews and meta-analysis
2.2.以往的综述和荟萃分析

3.1. Literature search strategy
3.1.文献检索策略

3.2. Inclusion and exclusion criteria
3.2.纳入和排除标准

3.3.1. Coding scheme of context features
3.3.1.语境特征编码方案

3.3.2. Coding scheme of background features
3.3.2.背景特征编码方案

3.3.3. Coding scheme of design features
3.3.3.设计特征的编码方案

3.3.4. Coding scheme of simulation experience features
3.3.4.模拟体验特征的编码方案

3.3.5. Coding scheme of outcome features
3.3.5.结果特征编码方案

4.1. Descriptive statistics
4.1.描述性统计

4.2. Overall effect on PTT
4.2.对 PTT 的总体影响

4.3.1. Moderator effects during the planning phase
4.3.1.规划阶段的主持人效应

4.3.2. Moderator effects during the implementation phase
4.3.2.实施阶段的主持人效应

4.3.3. Moderator effects during the evaluation phase
4.3.3.评估阶段的主持人效应

5. Recommendations on instructional design guidelines for HILTS within XR
5.关于 XR 系统内 HILTS 教学设计指南的建议

5.1. Instructional design guidelines during the planning phase
5.1.规划阶段的教学设计指南

5.2. Instructional design guidelines during the implementation phase
5.2.实施阶段的教学设计指南

5.3. Instructional design guidelines during the evaluation phase
5.3.评估阶段的教学设计指南

6. Conclusions and future research
6.结论和未来研究

MR
VR
AR
(4) Immersion level
Fully immersive
Non-immersive
Semi-immersive
(5) Interactivity
Non-natural interaction
Natural interaction
Both