When intuition conflicts with data

Copy this prompt

The data says {{data conclusion}} but my gut says {{intuition}}. Handle the conflict: (1) assume the data is right and ask where my intuition fails — did I see something outside the sample, or is this pure confirmation bias; (2) assume the intuition is right and ask where the data fails — definitions, sample bias, or a metric that does not proxy the real goal; (3) say which error is more common; (4) design the smallest experiment that would settle it; (5) tell me what to do until then — usually follow the data conditionally with an exit mechanism.

数据说 {{数据结论}}，但我的直觉说 {{直觉判断}}。请帮我处理这个冲突：1) 先假设数据是对的，我的直觉可能错在哪（我看到了样本外的什么，还是纯粹的确认偏误）；2) 再假设直觉是对的，数据可能错在哪（口径、样本偏差、指标不代理真实目标）；3) 指出哪种错误更常见；4) 设计一个能判定谁对的最小实验；5) 在实验出结果前，我该按哪个行动（通常是有条件地按数据走，同时保留退出机制）。