An opinion poll should not be accepted from its headline result alone. The first questions should be who paid for it, what outcome they sought, and how each prompt was phrased. Political sponsors can reproduce their own confirmation bias when they influence the questionnaire. In a later personal example, Dr. Clarke describes writing mirrored versions of a local-election question and receiving nearly identical but reversed results. The questionnaire itself must be criticised as a source before its percentages are interpreted.
Notable Point · 2:34:33-2:35:25
How poll wording can manufacture agreement
Polling should be read through its sponsor, purpose, and exact wording because mirrored questions can produce opposite-looking results from the same public.