You're simplifying too much. The tobacco company example illustrates very well how you can distract and lead astray by asking biased questions. Leading (smoking) statisticians like Fisher continued to point out there may be confounding causes. He did not falsify data. Even a spurious discussion about a "smoking gene" was created. Asking questions may seem innocent but it can be done strategically and can hinder drawing the right conclusions. Eventually the scientific method would correct those views, but it took more than 40 years even though the statistical data was available from the start.

Judea Pearl dedicates a whole chapter (Ch. 5) of his The Book of Why to this example.