Quote:
Originally Posted by Fuzz
|
I think post the Nenshi election that mainstreet changed their polling methods. They used to be a straight phone dialer which in that election oversampled the anti-Nenshi crowd.
If you read some of Janet Brown’s stuff on her current model and methods is that she is actively trying reach out to groups that are under-sampled or in the online panels. She does this by focusing on phone polling and multiple phone calls to phones to try to maintain the randomness of the samples.
From the Baired article on Janet browns polling
https://calgaryherald.com/news/local...94e44b504/amp/
Quote:
I do true random sampling using live telephone interviewers,” she says, “Most other polls are based on online panels which include people who have chosen to join the panel.
“Part of my method is to connect with hard-to-reach people. We’ll attempt to call a phone number five times at different times of the day. We try to get the right balance of landline and cell phone users.”
|
Quote:
|
Brown said: “I understand it’s shocking, but other polls have historically underestimated conservative voters, and my methodology does reach more conservative voters.”
|
When I read this quote from Brown I get flash backs to Mainstreets quote from the Smith /Nenshi election. It’s funny people still hack on mainstreet for that one without acknowledging he adjusted methods as a result of those results
None of this is saying she is wrong. As always a polling average made up of different methods of obtaining “random” samples is more likely to produce a better result than trying to pick the poll you think is right.