The SST has a story titled “Public demands ‘obnoxious’ Holmes be shown the door“.
And that would be a fine headline except in fact the public do no such thing.
Later down they report “The national poll shows one in five believe he should resign or be fired from NewstalkZB” which means four in five do not. Good God one in five people probably think Holmes should be sacked regardless of his “darkie” comments.
Also the poll had only 214 people take part which is a margin of error of 6.8% at the 95% confidence level. Worse they report a breakdown for the 45 men who took part which has a 14% margin of error!
And with 79% of the respondents being female, one can only conclude that it was not a random sample but some sort of self selecting sample which is useless as a serious indicator. No newspaper should report a self selecting sample as a seientific random poll.
The company they appear to have used is 18 Tracker which does focus groups not polls.
All in all one of the most incorrect and ignorant stories I have seen for a while.No tag for this post.