One thing I did was add uncertainty to the commonly used Gallup numbers about the change in public opinion. Folks often get into heated discussions about why support for a certain policy dropped by 3%, but such small differences are often not meaningful: