知识卡片
按时间分组保留时间混杂
把早登录者放一组、晚登录者放另一组,并不等于独立随机分派;时间若同时影响结果,就产生政策引入的选择路径。能控制组别值,不代表两组可直接交换比较;分派机制必须作为DGP的一部分检查。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 干预与因果效应
原书出处 · 1 处
- · 查看原书依据
first 500 players who logged on to the group with high Side-Quest Engagement and the next 500 players to the group with low Side-Quest Engagement. This intervention would be sufficient to d-separate the path Side-Quest Engagement ← Guild Membership → In-Game Purchases. But what if our DAG was wrong, and there are other paths between Side-Quest Engagement and In-Game Purchases through unknown common causes?
参考来源
- 位置:《Causal AI》第07章《干预与因果效应》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-7.xhtml)
- 结论依据:摘录中关于按时间分组保留时间混杂的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:first 500 players who logged on to the group with high Side-Quest Engagement and the next 500 players to the group with low Side-Quest Engagement. This intervention would be sufficient to d-separate the path Side-Quest Engagement ← Guild Membership → In-Game Purchases. But what if our DAG was wrong, and there are other paths between Side-Quest Engagement and In-Game Purchases through unknown common causes?