Quote:
Originally Posted by chowhou
Yes, but it's the pollster's responsibility to normalize against that.
|
Which they did - but with a relatively small sample the reliability starts breaking down. (In 2018 the same polling company had nearly 900 in their mayoral polling).
On the 522 polled, 343 were aged over 65 (so their views were weighted at 110 because they were vastly over-represented compared to the general population). Only 73 respondents were aged 18-49, so they were weighted at 304 (as they were even more significantly under-represented).
They also didn't ask 'how likely are you to vote in the next election', and given the modest turnout we usually see, the responses to the survey may not match the responses of likely voters.