Online information problem solving (OIPS) is essential for 21st century information literacy which often requires information selection competencies. However, students usually have problems in discriminating information from complex web sources. This study, utilizing eye-tracking technology, aims to examine the relationships among learners' visual search patterns, information anxiety and OIPS task performance in web search contexts. In this study, 46 university students volunteered to participate in a web search task for solving a landslide problem. Student’ visual behaviors were recorded by eye-trackers during the task and information anxiety was self-reported immediately after the task. Reaction time and task performance were also recorded and scored. Pearson's correlation analyses, multiple regression analyses, cluster analyses and lag sequential analyses (LSA) were conducted. The results show that learners' eye-tracking measures, information anxiety, and task performances are significantly correlated in OIPS. Students' visual attention paid onto irrelevant web information can significantly and positively predicts their information anxiety, but negatively predicts their task performance. Additionally, based on eye-tracking measures and reaction time, three visual search patterns are identified: Confused, Slow-thinking and Fast-thinking. The LSA results further show that different information selection and attentional control strategies are utilized by different groups, especially when discriminating the relevancy of web information. This study bridges the associations among eye-tracking measures, information anxiety and task performances in OIPS contexts. Future studies are suggested to analyze eye-tracking data using cluster analyses plus LSAs to profile online learners' characteristics in terms of visual search or information processing patterns.

在线信息问题解决 (OIPS) 对于 21 世纪的信息素养至关重要，这通常需要信息选择能力。然而，学生通常在从复杂的网络资源中区分信息方面存在问题。本研究利用眼动追踪技术，旨在研究网络搜索环境中学习者的视觉搜索模式、信息焦虑和 OIPS 任务表现之间的关系。在这项研究中，46 名大学生自愿参与解决滑坡问题的网络搜索任务。在任务期间，眼动仪记录了学生的视觉行为，并在任务完成后立即自我报告了信息焦虑。反应时间和任务表现也被记录和评分。Pearson 相关分析、多元回归分析、进行了聚类分析和滞后序列分析（LSA）。结果表明，学习者的眼球追踪测量、信息焦虑和任务表现在 OIPS 中显着相关。学生的视觉注意力不相关的网络信息可以显着地积极预测他们的信息焦虑，但消极地预测他们的任务表现。此外，基于眼动追踪测量和反应时间，确定了三种视觉搜索模式：困惑、慢速思考和快速思考. LSA 结果进一步表明，不同的群体使用不同的信息选择和注意力控制策略，尤其是在区分网络信息的相关性时。这项研究将 OIPS 环境中的眼动追踪测量、信息焦虑和任务表现之间的关联联系起来。建议未来的研究使用聚类分析和 LSA 来分析眼动追踪数据，以在视觉搜索或信息处理模式方面描述在线学习者的特征。