Lies, Damned Lies, Statistics & 2004 Campaign Polls

Polls are a refined breed of statistics. Both genres have in common the manipulation of data to support an objective. Polls, however, do have the added functionality of affecting voter turn-out and providing disinformation to opposing camps; all in real time video game style. In a close race like the one at hand, it is reasonably possible to see the pivotal points massaged by Democratic and Republican projections to achieve the desired results.

The Republican Spin on the Polls:

On the EV front, using the Rasmussen Report data only, it looks like we still have 131 within 5%: Colorado, Florida, Iowa, Michigan, Minnesota, Nevada, New Mexico, Ohio, Pennsylvania & Wisconsin. On the basis of spreads over 5% we have Bush at 213 and Kerry at 194. As good a way as any to conisder the data is to ask whether the undecided vote in a given state is sufficient to overcome the spread between the candidates.

On that basis (U=Undecided Vote & S*=Spread Bush / Kerry leading) here is how the Republican Pollsters get Bush over the line:

Colorado U(3%) < SB(4%) = Bush +9
Florida U(3%) = SB(3%) = In Play (+B) +27
Iowa U(3%) < SK(4%) = Kerry +7
Michigan U(3%) = SK(3%) = In Play (+K) +17
Minnesota U(5%) > S?(0%) = In Play +0 (10?)
Nevada U(4%) > SB(2%) = In Play (+B) +5
New Mexico U(3%) > S?(0%) = In Play +0 (5?)
Ohio U(3%) > SB (2%) = In Play (+B) +20
Pennsylvania U(4%) > SK(1%) = In Play (+K) +21
Wisconsin U(1%) = SK (1%) = In Play (+K) +10

Bush= 213 + 9 + 27 + 5 + 20 = 274
Kerry= 194 + 7 + 17 + 21 + 10 = 249

If you spott Kerry the 15 from Minnesota and New Mexico dead heats that would take him to 264. That seems to be how the Republican pollsters make the prediction of 274-264 Bush (see www.electionprojection.com... below)(both cited by The Big O in the thread shown below.)

The slight hitch: basically, must ignore the disadvantage of the incument among the undecided voters and assume Bush wins in all the states that are In Play in which he has a lead and those where the spread is greater than the undecided vote.

The Democratic Spin

To get to the Democratic prediction of 253-247 Kerry & counting, the Democratic Pollsters account for the fact that New Hamshire has moved into the

Originally posted by G_Scard
The slight hitch: basically, must ignore the disadvantage of the incument among the undecided voters and assume Bush wins in all the states that are In Play in which he has a lead and those where the spread is greater than the undecided vote.

I do not see where this is a hitch with their use of the numbers. If this disavantage cannot be quantified it is useless in the examination of statistics.

As a note, both sites to which you linked make this same assumption. So, I don't see where this is spin at all. Rather I see it just as I said, an inability to quantify.

Originally posted by Raphael_UO
If this disavantage cannot be quantified it is useless in the examination of statistics.

As a note, both sites to which you linked make this same assumption. So, I don't see where this is spin at all. Rather I see it just as I said, an inability to quantify.

Good points. Thank you.

Since I am a voter in search for the truth, I thought I would share where the republican spin could be found.

www.electionprojection.com...

I am certain you can find several things in that which would qualify as spin.



[edit on 17-10-2004 by Raphael_UO]

Quite frankly, I had seen that formula and thought it well beyond mere spin. In the absence of any quantitative demonstration of the use of the formula, and given the result I got by just taking the SS route to Republican projection, I did not think I could give any credance to the lengthy explanation.

The difference of position on Florida is alarming. One side calls it a 3% Bush lead and the other calls it a dead heat. ummm...

[edit on 18-10-2004 by G_Scard]