you have to understand that the nanos polling is a rolling average, not a snap poll. it's not nik's fault. he's clear about what he's doing. but, it's consistently reported inaccurately.

this report does not suggest that the parties are running at that level right now. it suggests that it is the average level of support that they ran at last month. to present it as a snap poll is dishonest - perhaps accidentally, as i know that journalists are often not so good at math.

the value of this kind of polling is to track long term trends. and, because nik's samples are so small and consequently vary a lot, it works together to smooth itself out.

it's a measure of party viability. and, because this election is so long, it may be useful in understanding where people's brains really are in the long run, rather than where their votes are in the short-run.

it's just important that you understand what you're actually reading. and, it would help if the media were clearer about it.

nik's polls have a lag built into them - it's by design. it's just measuring something else. they will align with the current consensus over the next month, as they catch up. but, by then, the consensus may have changed, and they'll be behind again.

right now, they're useful because the race is young and there's a lot of undecideds. but, as we get closer and closer to the election, this method will be unable to predict last minute changes.

for example, something you can pull out of nik's polling that you won't see elsewhere is that the ndp support base in quebec is relatively weak - some of that support is torn between the bloc and the ndp, whereas some of it is torn between the ndp and the liberals. it's not that snap polling would be over-estimating current support so much as it's pointing out that the current support level is not strong, not decided and still open to switching.

it also suggests that the ndp support in ontario is very weak, and possibly driven by perception about who has a better chance of winning. again: that's not to suggest that the snap polling is wrong. it's to suggest that the support is not strong.

and, so the takeaway is that the race is still wide open, and that the snap polling at this early stage should be taken with some caution.

http://www.theglobeandmail.com/news/politics/three-main-federal-parties-tangled-in-three-way-tie-nanos-poll/article26090470/