Thursday, January 27, 2011

If there's any pollsters reading Liblogs right now...

Come hither, I have a request.

It's a very simple one: next time you do any party leader preference questions, it would be very interesting if you did it in the style that Ipsos MORI over in the UK does - whether people like the leader, or like the party.

If you want to see how its done, skip to page 33 of MORI's newest poll and read down. But if you don't want to hurt your brain with the huge amount of data in those 42 pages, here's a simple explanation.

Instead of asking simply whether voters approve of Harper, Iggy, and Layton, you ask voters whether they approve of the leaders plus how much they approve of their respective parties. It would go something like this:

Let's say the leadership "likability" numbers are as follows: Harper 28%, Iggy 14%, Layton 26%. These numbers are kind of like approval numbers, which would be said as "Harper and his party," as MORI phrases it.

Then add on to it the question of whether they like the leader but not the party, and vice versa. Just focusing on the Conservatives for this one, let's say 20% like Harper but not his party, and 15% like his party (Conservatives) but not Harper.

What you then get is a number that corresponds to 48% liking Harper (28+20), and 43% liking the Conservatives (28+15). Note that these are not real numbers I think would come true, but simply examples.

I think this would be an excellent indicator of who is leading what, exactly. Is the man more popular than the party, and is therefore an asset? Or is it vice versa, as I'm almost sure would be true for Ignatieff. An excellent way to dissect the political scene even further, no?

Think about it. No pressure.

No comments:

Post a Comment