Yes, yes, I can hear the groans from here already, but when I see two polls one day apart from each other that give diametrically opposite results in the Florida Governor’s race, I get worried.
Republican Rick Scott had similarly erratic polling in his primary race which finished close with 3% final difference, so as his polling against Democrat Alex Sink is swingy, I worry about the need for another divisive recount.
The first poll comes from Mason Dixon and shows Sink ahead 44-40 (MoE 4), three points tighter than the firm had it last time.
Next we check in with Rasmussen Reports, whose latest poll shows Scott ahead 50-47 (MoE 4), two points tighter than the previous RR result.
If I average the two in my usual way I get a 53/47 split in Sink’s favor, practically indistinguishable from a coinflip. So yes, I am worried about a recount. With so many competitive races this year we’re practically guaranteed to have at least one, but it doesn’t mean I can’t root against having one.
House of Representatives Swingometer
Electoral College Swingometer
Lets see, my math shows (50 + 40)/2 = 45 for Scott, (44+47)/2 = 45.5 for Sink.
I average lead probabilities, which takes into account both the raw numbers and their gaps, but also the Margins of Error.
To me this is just another installment of your last Florida article “Terribly inconsistent polling in Florida.”
Do we have anything in the crosstabs that would point to one sample or another being a little bit off?