Setting Physical Activity Goals with a Virtual Coach: Data and Analysis Code

This is the data and analysis code underlying the paper "Setting Physical Activity Goals with a Virtual Coach: Vicarious Experiences, Personalization and Acceptance" by Nele Albers, Beyza Hizli, Bouke L. Scheltinga, Eline Meijer, and Willem-Paul Brinkman. The paper examines the use of personalized vicarious experiences in a goal-setting dialog for physical activity with a virtual coach. Study The paper is based on the study conducted in March 2022 for the publicly available Master's thesis by Beyza Hizli (http://resolver.tudelft.nl/uuid:b7225a91-6ae8-4a32-8441-38fb7ff74b4c). In this study, 39 fluently English-speaking adults set a running or walking goal with the text-based virtual coach Jody. Participants thereby received either two of the three overall most motivating vicarious experiences (i.e., "generic" experiences) or "personalized" vicarious experiences chosen based on a linear regression model. During the dialog, participants wrote what they could take away from the examples from other people that they saw. After the dialog, participants further provided information regarding their self-efficacy for the type of goal they set (i.e., running or walking), ratings of examples on perceived motivational impact, and a free-text response about what they found motivating about the examples they saw. The study was pre-registered in the Open Science Framework (OSF): https://osf.io/4duwh. Links to further resources: The implementation of the virtual coach Jody is available here: https://doi.org/10.5281/zenodo.6647381. The 72 examples from other people used in the study can be found in the dataset accompanying the Master's thesis by Beyza Hizli: https://doi.org/10.4121/20047328.v1. A video of part of the dialog with the virtual coach is available here: https://youtu.be/7WJy7-H3QEQ. Data We collected data in three separate studies: Study to collect the vicarious experiences (part A). We gathered the vicarious experiences together with data on individual characteristics of the participants (e.g., age, stage of change for becoming physically active). We collected vicarious experiences from 72 people. Study to rate the vicarious experiences on perceived motivational impact and similarity (part B). We asked 36 individuals to each rate 18 vicarious experiences on perceived motivational impact and similarity to the person in the experience. We further collected data on individual characteristics of these 36 participants. Study to evaluate the goal-setting dialog and the vicarious experiences therein (part C). 39 participants interacted with the text-based virtual coach Jody to set a running or walking goal. Participants saw either personalized or generic examples from other people. We also collected data on individual characteristics of these 39 participants. Please refer to the OSF pre-registration for more information on the data we collected. In addition, we describe our measures in detail in a separate sheet in each data file in the "Data"-folder. Note that we do not provide all data here: we describe in our README-files which data needs to be downloaded from the repository accompanying the Master's thesis by Beyza Hizli to reproduce some of our analyses (https://doi.org/10.4121/20047328.v1). If you have any questions, please contact Nele Albers ([e-mail]).

本数据集对应Nele Albers、Beyza Hizli、Bouke L. Scheltinga、Eline Meijer以及Willem-Paul Brinkman所撰写的论文《Setting Physical Activity Goals with a Virtual Coach: Vicarious Experiences, Personalization and Acceptance》的配套数据与分析代码。该论文探讨了虚拟教练辅助下的体育活动目标设定对话中，个性化替代性体验（vicarious experiences）的应用场景。

本研究依托Beyza Hizli于2022年3月开展、可公开获取的硕士论文（http://resolver.tudelft.nl/uuid:b7225a91-6ae8-4a32-8441-38fb7ff74b4c）中的相关研究进行。在本次研究中，39名英语熟练的成年受试者与基于文本的虚拟教练（virtual coach）Jody协作设定跑步或步行目标。受试者将被分为两组，分别接收基于线性回归模型筛选出的“个性化”替代性体验，或三类整体最具激励性体验中的两类“通用”替代性体验。在对话过程中，受试者需记录自身从他人示例中获得的启发。对话结束后，受试者还需提供其所设定目标类型（跑步或步行）相关的自我效能感信息、对示例感知激励效果的评分，以及关于其所感知到的示例激励因素的自由文本反馈。本研究已在开放科学框架（Open Science Framework, OSF）进行预注册：https://osf.io/4duwh。

相关补充资源如下：虚拟教练Jody的实现代码可通过以下链接获取：https://doi.org/10.5281/zenodo.6647381。本研究使用的72条他人示例数据集可在Beyza Hizli硕士论文的配套数据集中获取：https://doi.org/10.4121/20047328.v1。与虚拟教练对话的部分片段视频可通过以下链接观看：https://youtu.be/7WJy7-H3QEQ。

本次数据采集包含三项独立研究：
1. 替代性体验采集研究（A部分）：我们共收集了72名受试者的替代性体验数据，同时采集了受试者的个体特征信息，例如年龄、参与体育活动的行为改变阶段等。
2. 替代性体验评分研究（B部分）：我们邀请36名受试者分别对18条替代性体验的感知激励效果及与体验中人物的相似性进行评分，同时采集了这36名受试者的个体特征信息。
3. 目标设定对话与替代性体验评估研究（C部分）：39名受试者与基于文本的虚拟教练Jody交互，以设定跑步或步行目标。受试者将查看个性化或通用的他人示例。我们同时采集了这39名受试者的个体特征信息。

更多关于本次采集数据的详细信息，请参阅OSF预注册页面。此外，我们在“Data”文件夹内各数据文件的单独工作表中对各项测量指标进行了详细说明。请注意，本仓库并未包含全部数据：我们在README文件中说明了需要从Beyza Hizli硕士论文的配套存储库（https://doi.org/10.4121/20047328.v1）下载哪些数据以复现部分分析结果。如有任何疑问，请联系Nele Albers（邮箱：[e-mail]）。