Can a poll published after a debate tell us how people reacted to that debate?
PoliticsI'm putting together a reading-group handout and have invented a simple example to test a claim: a televised debate happens on Wednesday, a poll is published on Friday, but its interviews all took place on Monday and Tuesday. Someone wants the caption to say, 'Friday's poll shows the debate changed minds.'
The publication date makes that sound plausible at a glance. But the people were interviewed before the event. AAPOR's disclosure standards include the dates of data collection among the information to disclose.
In this illustrative example, I would say the poll describes views gathered before the debate and cannot establish a response to it. Would you remove the debate claim completely, or keep the poll beside the debate report with a clear date label? I don't want layout to suggest a causal link the wording has withdrawn.