Isn’t 47-43 = 4?
Noticed the same...somebody needs to re-take kiddie math.
Could be a rounding thing, it could be 46.5 to 43.4. In which case it’s 47-43 when rounded but the actual lead is just 3.1%.
It’s 47-44.
“Obama’s previous lead over Romney of 45.3% to 44.4% has switched to a Romney lead of 46.8% to 43.7%. Although within the previous polls margin of error, the Republican convention appears to have given Romney a 2 point bump while taking 1 point from Obama. Switching to the Senate race, before the Republican convention, Democrat Sherrod Brown held a nar- row 43.6% to 41.4% lead over Republican Josh Mandel. Again, although within the margin of error, the post- convention poll gives Republican Mandel the advantage at 44.3% to 43.5%, representing a 3 point bump for Mandel, while Brown’s take staying basically constant.”
http://www.realclearpolitics.com/docs/2012/Gravis_OH_0904.pdf
Joe Biden math. J-O-B-S
Depends on the rounding... E.g. 46.6 (rounds to 47) minus 43.4 (rounds to 43) equals 3.2 (rounded to 3 in the final report). As you see, it's easy to be off a bit when you take this into account. It can get even closer... using rounding, the actual number could be 3.0000001 and yet still seem to be 4.
I don't know for CERTAIN, but this is my guess, heh.