| 陆伟,张文波,张帆.对话式搜索下学生群体的信息搜寻行为对搜索结果质量的影响研究—— 基于fsQCA 的组态分析[J].情报工程,2026,(1):086-096 |
| 对话式搜索下学生群体的信息搜寻行为对搜索结果质量的影响研究—— 基于fsQCA 的组态分析 |
| The Impact of Students’ Information-Seeking Behavior on Search Outcome Quality in Conversational Search: A Configuration Analysis Based on fsQCA |
| |
| DOI： |
| 中文关键词: 对话式搜索；信息搜寻行为模式；搜索结果质量；任务复杂度；模糊集定性分析 |
| 英文关键词: Conversational Search; Information-Seeking Behavior Patterns; Search Outcome Quality; Task Complexity; Fuzzy-Set Qualitative Comparative Analysis |
| 基金项目:国家自然科学基金青年科学基金项目“对话式搜索离线评价方法研究”（62202342）。 |
| 作者 | 单位 | | 陆伟 | 武汉大学信息管理学院 武汉 430072 | | 张文波 | 武汉大学信息管理学院 武汉 430073 | | 张帆 | 武汉大学信息管理学院 武汉 430074 |
|
| 摘要点击次数: 4 |
| 全文下载次数: 2 |
| 中文摘要: |
| [目的/意义]本文旨在揭示以学生为主的用户群体在对话式搜索环境下，不同信息搜寻行为特征如何实现高质量的搜索结果。[方法/过程]基于用户实验，通过设计搜索任务与开发对话式搜索交互系统，收集并量化了用户的查询、浏览、点击等信息搜寻行为特征。随后，采用模糊集定性比较分析（fsQCA）方法，探讨了在不同任务复杂度下，信息搜寻行为与搜索结果质量之间的组态关系。[局限]本文的样本规模有限，行为变量选取也存在一定约束，这可能影响研究发现的广泛适用性与覆盖范围，需在未来研究中进一步拓展。[结果/结论]在低任务复杂度条件下，识别出三种能够关联较高搜索结果质量的行为特征组态；而在高任务复杂度情境中，仅有两组组态表现出此类效果。综合来看，这些组态可归纳为快速浏览型、深度探索型、混合探索型和AIGC 主导型四类信息搜寻行为模式。研究表明，对话式搜索中信息搜寻行为与结果质量之间存在多重并发因果关系，AIGC 主导型的出现也反映出对话式搜索新范式对用户行为模式的影响与重塑。 |
| 英文摘要: |
| [Objective/Significance] This study aims to uncover how different information-seeking behavior patterns among student users in conversational search environments lead to high-quality search outcomes. [Methods/Processes] Based on user experiments, we designed search tasks and developed an interactive conversational search system to collect and quantify behavioral features such as querying, browsing, and clicking. We then employed fuzzy set Qualitative Comparative Analysis (fsQCA) to explore the configurational relationships between information-seeking behaviors and search outcome quality under tasks of varying complexity. [Limitations] The limited sample size and the constrained selection of behavioral variables may affect the generalizability and coverage of the findings, which calls for further expansion in future research. [Results/Conclusions] Under low task complexity, three configurations of behavioral features were found to be associated with high search outcome quality, while only two such configurations emerged in high task complexity contexts. These configurations can be categorized into four behavior patterns: rapid-browsing, deep-exploration, hybrid-exploration, and AIGC-dominated. The study reveals multiple concurrent causal relationships between behavioral features and search outcomes, and the emergence of the AIGCdominated pattern suggests that the new paradigm of conversational search is reshaping users’ information-seeking behavior to some extent. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
