Data and code underlying the bachelor thesis: Traits for a virtual coach to be a ”friend”

This dataset contains data and code snippets from the analysis performed during the thesis, "Traits for a virtual coach to be a friend". Here, an investigation has taken place to what characteristics the virtual coach must possess to establish this friendly relationship. Thus, the main research question is: <strong>What are the reasons for seeing the virtual coach as a stranger or friend?</strong> This research has been based on a retrospective study performed by Albers and Brinkman. Here, five hundred participants interacted with the text-based virtual coach Sam - developed for the Perfect Fit project to convince smokers to quit by performing small activities - in five separate sessions. Afterwards, each participant rated the relationship with the virtual coach, followed by an explanatory free-text response to which thematic analysis was applied. The quality of the thematic analysis has been ensured by researcher and method triangulation. Researcher triangulation, where multiple researchers were involved, determined the reliability of the coding scheme generated during thematic analysis. Method triangulation, which supported the findings, was executed with the free-text responses, literature, and quantitative data, containing demographic and smoke-related characteristics.

本数据集收录了题为《虚拟教练作为友人所需的特质》的学位论文研究分析所用的数据与代码片段。本研究围绕虚拟教练（virtual coach）需具备何种特质方可与用户建立友好伙伴关系展开，因此核心研究问题为：**将该虚拟教练视为陌生人或是友人的原因是什么？**本次研究基于Albers与Brinkman开展的一项回顾性研究，该研究中500名参与者在5个独立会话中与基于文本的虚拟教练Sam进行互动。Sam是为Perfect Fit项目开发的虚拟教练，旨在通过引导用户完成小型活动帮助吸烟者戒烟。互动结束后，每名参与者需对自身与虚拟教练的关系进行评分，并提交解释性自由文本回复，随后研究人员对该回复开展主题分析。主题分析的质量通过研究者三角验证与方法三角验证两种方式得以保障：其中研究者三角验证邀请多名研究者参与分析，以确定主题分析过程中生成的编码方案的信度；方法三角验证则结合自由文本回复、文献资料与包含人口统计学及吸烟相关特征的定量数据，为研究结论提供支撑。