October 3, 2008

Polling 101

If you are conducting a poll, an online poll no less, in the MIDDLE of a debate, it proves your respondents, are

1) Not paying attention,
2) May already have had their mind made up,
3) Assuming everybody waits to answer online polls while trying to decipher what the 5 leaders are saying.

Absolutely ludicrous and all media outlets reporting on its results should be ashamed of themselves.

I watched the whole thing, but I am not going to tell you who I think won, although I do think the winner was glaringly obvious.

Labels:

4 Commentaires:

Blogger Emerys a dit...

It is true, Layton was head and shoulders above the crowd.

10/03/2008 8:17 a.m.  
Blogger Antonio a dit...

oh god

Layton and the NDP just dont understand that in a time of economic turmoil, taxing businesses really is the best way to make it worse

10/03/2008 11:37 a.m.  
Blogger Emerys a dit...

Yeah, totally, those crazy lefties just don´t realize that FDR saved capitalism through neoliberal corporate tax cuts and deregulation(rolls eyes)... You have to love how with the economy rapidly going down the toilet the neoliberals are trying to hand their buddies as much cash as they can through bailouts and tax cuts before they get kicked out of the drivers seat again.

10/03/2008 12:15 p.m.  
Blogger KC a dit...

The whole way the media reports on polling is a disgrace. It drives me to drinking.

1) The poll you commissioned is only one poll. Don't use it to draw conclusions like "X is closing the gap on Y" or "X's support is slipping away" when other polls are contradicting that assertion. Nothing pisses me off more than going to one newspaper and seeing a headline saying "Party X is surging", and going to another saying "Party X's support is slipping away".

2) Be mindful of your margin of error, particularly on regional polls. If changes of party support are within the margin of error dont conclude that there has actually be a substantial change of support. This is particularly true with regional polls that generally have high margins of error.

3) Quit assuming that short term trends are going to continue over the long run. Just because Party X has gone up by a significant amount over the past four weeks dont assume that they are poised to run away with it. Same with the converse. Just because a party is down substantially doesn't mean they are going to fade away and (in the case of the LPC be replaced as the official opposition).

Polls are useful when you keep the margin of error in mind, and focus on them in the aggregate. Following the fluctuations of a single poll from day to day and drawing strong conclusions from that data is just stupid.

10/04/2008 2:58 p.m.  

Post a Comment

<< Home