My memory is long for things I’ve read over and over again, especially numbers, so I want to know what the pollsters are trying to put over on us.
You see I remember the polls from the ’80s and even possibly the ’70s and it was always a survey of over 1000 people that they would say was correct to 4%, 19 times out of 20.
This means that the polls, once every 20 times are total bull$#!+ . That’s my first point.
But recently I paid attention to the characteristics of a modern poll, in this, Canada’s 41st federal election. They had the results of less than 1000 people and they claimed their result was accurate to 3.1%. I don’t remember hearing the 19 times out of 20, but I believe statistical analysis hasn’t progressed any further, so it should still be there.
Now how can a lesser poll have more accurate estimates? Then it hit me what they had done. They had simply subtracted the undecided vote that used to be ever present in polls and that’s why the polling numbers were less than with the polls of decades earlier.
It seems to me they don’t want to remind the sheeple that usually there are 20% of the voters that are undecided. That means, in the present Canadian race, the NDP (New Democrat Party) which is currently in third at 17% of decided voters, could form a majority government with just the pull of the undecided voters. That’s the 2nd major reason the polls may be relative bull$#!+ . It might be a greater than 1 in 20 chance that such mayhem could ensue.
And I still haven’t worked on the third thing that seems suspiciously out of whack. That decrease in poll inexactitude from 4% decades ago to 3.1%, now. It’s even worse than that. Here I have a poll table that “shows” some of the errors are allegedly less than 1.7 %. I simply don’t believe these figures. Not unless the pollsters have gone the expensive route of polling 10 000 people at a time. I seriously doubt they’ve gone this route.
The only other thing I can think for inexactitude being lessened would be if the pollsters did other polls, like asking “Do you use polls to strategically vote?” In this case the pollsters might have a case for inexactitude being lessened but the response would have to be overwhelmingly “Yes.” But notice that this “correction” relies on another poll.
All I can say is don’t be a sheeple. If you ever strategically vote, do it based on the last election, not a random survey.
A corrupt survey system might result if we give their polls too much power. In these times of less oversight and a much smaller watch dog press, I don’t think we can afford to let any part of our democracy get any more powerful. Let’s go back to those ’70s and ’80s qualifiers for our polls. I don’t believe there is anything that would make me agree that statistics have miraculously gotten better. I think the pollsters are getting deeper in the bull$#!+ .

