86% of all Statistics are Made Up

The Toronto Star has joined the Globe and Mail in commissioning a poll of Liberal members. The star is not disclosing how they got their hands on the membership lists. All I can say is the Liberal lists seem to be easier to get than the common cold. I have a couple of questions about the accuracy of the polls:

1. The Star notes that a majority of riding delegates will come from Ontario and Quebec. This is true. It is also true that 20% of delegates will come from Alberta and BC where the membership rolls are considerably smaller. I wonder if Alberta and BC liberals were properly represented in this poll. If not, numbers for candidates with strength out West (like Kennedy) will be significantly lower than their delegate count will be next weekend.

2. There are going to be a lot of youth going to the convention. Nearly a third of all delegates by reckoning will be youth. Now, I know that youth do not decide which youth delegates go to the convention but I guarantee you that youth will go out and vote this weekend. Furthermore, there will be about 200 delegates from student clubs which will be chosen only by youth. Pollsters have a tendency to ignore young people when they do polls. This makes sense in a normal election because young people don't have that much of a say in who wins (their voter turnout is abysmal), but in this case the youth have real power and will shape this race. If the pollsters discounted youth members because they're hard to find or because they don't think that they matter, strength for certain candidates with strong youth support, (like Kennedy) would be underestimated.

Now it's possible that both of these things were considered by the pollsters. In which case their numbers should be accurate. I guess we'll find out this weekend.