Margins -- Perry vs. Obama, Romney vs. Obama

[Wonkish1]

You know I thought this forum was filled with people that actually knew something about election analysis.

Ask any professional poster, this far out head to head match ups are near useless, but what is a little more telling is:

A's approval/disapproval vs. B's approval/disapproval...the polls always converge on that anyway.

Yet, it seems like everybody would rather just use the head to heads.

[Wonkish1]

Favorable/unfavorable
and
Approval/disapproval

Are vastly more consistent and less volatile than head to heads. Its not even close. It takes a very long time for a person to get positive movement in them.

2 examples:
1) Throughout 2007 Clinton had an absolutely horrible favorables. It took an entire year of strong debate performances and casting herself as a centrist for her favorables to even move to even. And that's where they remained for remainder of the primary season.
2) Newt's debate performances over the last few months have been among the best among the GOP candidates, yet his favorable/unfavorable ratings are still in the toilet and have only nudged a little off their lows.

In 2007 I predicted Obama would take the nomination almost purely on the favorable/unfavorable metric when every head to head had him getting crushed and every pollster writing him off.

Also, a pollster can just as easily press an undecided to pick a candidate as they can push someone to pick a favorable/unfavorable rating.