Share
6,517 Posts.
lightbulb Created with Sketch. 0
clock Created with Sketch. 23/04/18
20:28
Share
Originally posted by eagle888
↑
Hi sparrasid. Just an observation that while using forced respondent allocated preferences gives a definitive view of the polled individuals second preference, it may not translate to the same outcome at the election. This is because a fair proportion of voters will follow the card of the main Party they support and not bother with a choice of their own. They go the Party line because they figure the Party would have framed their preference choices such as to give their candidate the best chance of winning. So imo the jury is out as to whether respondent allocated preferences in polling will prove more accurate than using the previous election data.
Expand
Yeah that's a good point eagle, and indeed Newspoll was equal best at predicting the 2016 federal election under its previous methodology. I'll be very interested to see which way works best at the next election.
