26 Nov 2009

I'd love to read the methodology used to produce this poll

Via Brad DeLong


2 comments:

OBloodyHell said...

I'd assume it's non-exclusionary.

"Which of the following would you support in 2012 (check all that apply)"

The pie chart is sort of odd for display, sure, but there are ways to make that semi-sensible, too -- what percentage of the totals support each individual, or something.

Joao Pereira said...

Well, maybe you're right, but if one were given the possibility of multiple choices then they shouldn't use percentages on a pie chart, it just looks silly.
On the other hand, it would be more informative to make a poll in which people may only choose one candidate. That would give us really an idea of which candidate was really preferred.