Some of us use purple... and some even understand it
I look terrible in purple. SEE!
Because the margin of error is 3.4% and the "difference" between the two results falls within that margin of error.
%32 vs %30 with a margin of error of +/- 3.4% is not a statistically significant difference. They are saying that for all they know, the Conservative support could actually be 28.6% and the Liberal support 33.4%, or any other possible combination within the margin of error.
Trying to predict likely voter turnout is something that is nice to try, but there is no evidence that these guys can do that with any degree of accuracy. Turnout for each party is mostly a function of each party's ability to get out the vote, and that varies from election to election.
Yep, look at that we agree.