Reporting on polls during elections may be journalistically valid (I can argue both sides of that), but one of the problems is not so much media reliance on them as it is media reliance on a single poll, most often one that they are a sponsor of, and using that single set of numbers to drive the storyline.
Why does that matter? Isn’t a poll a poll?
Not according to this wrap-up from Taegan Goddard’s Political Wire:
In Canada, the latest Ipsos Reid survey indicates that Stephen Harper and the Conservatives “are well on their way to forming a strong minority government come January 23rd, and are within close-striking distance of winning a majority government.”
Other polls:
CTV shows the Conservatives with an 18% lead over the incumbent Liberal Party.
Ekos has the Conservatives up by 11 points.
SES Research gives the Conservatives an eight point lead.
Strategic Counsel puts the Conservatives ahead by 13 points.
All the polls may point to the same trend, but there’s a big difference between an 18 per cent lead and and eight-point lead.
TAGS: POLLING, JOURNALISM, ELECTION
Poll is poll if sampling is right