Dataset of analysis of search engine’s question answering system responses to queries concerning first aid in common adult health emergencies

This dataset contains results of evaluation of the Google search engine's question answering system advice on first aid generated in response to queries concerning provision of help in common adult health emergencies.

In December 2022-January 2023, an English-language search was carried out in Google with ten queries based on the following keyword combinations: (what to do OR how to help) AND (bleeding OR chest pain OR choking OR not breathing OR seizure).

The search engine’s question answering system responses (Featured snippets and People also ask items) were collected (up to 11 per search query). Unique eligible question answering system responses were evaluated for adherence to the International Federation of Red Cross First Aid Guidelines 2020 using the predeveloped checklists. Adherence to the guidelines was rated for each item of respective checklist as complete (checklist item wording satisfied in full), incomplete (checklist item wording satisfied in part) or absent (where the question answering system response omitted corresponding first aid instruction).

Additionally, source webpages of eligible question answering system responses were categorised as Academic, Charity/Non-profit entity, Commercial, Government or Media outlet, and content of the source webpages was evaluated for quality of first aid instructions using the same checklists with a focus on source article text.

本数据集收录了针对成人常见健康急症施救相关查询，谷歌（Google）搜索引擎问答系统生成的急救建议的评估结果。

2022年12月至2023年1月期间，研究人员以英文在谷歌搜索引擎中开展10组查询搜索，所用关键词组合为：(what to do OR how to help) AND (bleeding OR chest pain OR choking OR not breathing OR seizure)，可译为「采取何种措施 OR 如何施救」与「出血 OR 胸痛 OR 气道异物梗阻 OR 呼吸停止 OR 惊厥发作」的逻辑组合。

研究人员收集了搜索引擎问答系统的回复结果（包括精选摘要（Featured snippets）与相关问题（People also ask）条目），单条查询的结果收集上限为11条。对经去重后的符合筛选条件的问答系统回复，采用预先开发的核查清单，评估其是否符合《红十字会与红新月会国际联合会2020版急救指南》。依据核查清单的各条目，将指南依从性分为三类：完全符合（核查清单条目表述被完整满足）、部分符合（核查清单条目表述被部分满足）以及缺失（问答系统回复未提及对应急救指导）。

此外，将符合筛选条件的问答系统回复对应的源网页划分为学术机构、慈善/非营利组织、商业机构、政府部门或媒体平台五类，并采用相同的核查清单，重点围绕源网页正文内容，评估其急救指导的质量。