Pollster.com

Articles and Analysis

 

Lunchtime Status Update for 9-18

Topics: Status Update

Another 24 hours, another 17 new statewide polls on the presidential race. These include two new polls each three battleground states: Florida, Wisconsin and (surprising that it may be a battleground) Indiana.

The two new polls from Indiana shift Indiana on or map from lean McCain to toss-up. Both polls are first measurements for Indiana on the Obama-McCain race by their respective pollsters. CNN/ORC has McCain leading by six points (51% to 45%), while the Indianpolis Star/Selzer and Company have Obama ahead by three (47% to 44%). These two results, combined another close result from Howey-Gauge in late August, move our current estimate to a not quite 3 point lead for McCain (48.0% to 45.6), close enough for toss-up status given the typical sample sizes in Indiana.

The change cuts McCain's electoral vote total on our map from 219 to 208, while the toss-up total is now 87.

The afternoon should be a busy one, as we are expecting imminent releases of two sets of battleground polls from our colleagues: The first All State/National Journal Battleground surveys and the first Big Ten battleground surveys in six midwestern states, codirected by our own Charles Franklin.

 

Comments
JFactor:

That Big Ten battleground survey seems very interesting. By the way, I hope someone would survey Arkansas and Lousiana because it bothers me they're still light red. Oh well, it's just cosmetics.
____________________________________
http://www.internationalpoliticstoday.com

____________________

hyh:

On another topic...

Given the short amount of time between this year's two conventions, measuring how big the bounce was from the second convention might be more complicated than usual.

One could argue that the second bounce should be measured from when the convention started. By this method, McCain had a bigger bounce, and it seems many media types have used this type of argument.

But is this really appropriate given the ephemeral bounce from the first convention in at near full force at the start of the second convention?

Perhaps in cases where there was not enough time for the polling numbers to settle down to disregard the first convention bounce, a better method for measuring the second convention bounce is to use the polling numbers before the first convention as baseline?

Democratic convention started on Aug 25. Taking Gallup Daily aggregate from Aug 20-25, Obama and McCain both have polling average of 45% (average of Aug 26 and Aug 23 numbers). At the height of Obama DNC bounce, Obama has 50% of votes or a 5% increase (Aug 30-Sep 1). Republican convention started on Sep 1. At the height of McCain RNC bounce, McCain has 49% of votes or 4% increase (Sep 5-7). By this reckoning, one cannot argue that McCain had a bigger bounce.

For Rasmussen, Aug 20-25 numbers are Obama 47% and McCain 46% - with Obama 5 point (9/2) increase at his peak and with McCain 4 point increase at his peak (9/12-14).

Yet another method would be to use the polling data from 2 weeks or maybe ~3 weeks after the end of second convention as the baseline to measure how big the second bounce was. If the first and second bounces were essentially of same size, there would be no change between before-conventions and after-conventions polling numbers. The numbers we’re seeing now and over next few days should settle this matter with caveat that it does not account for other meaningful factors (e.g. major gaffes, disasters, etc) that genuinely moved the polling numbers one way or the other during that time.

____________________

Allen:

Nationally, McCain's bounce peaked around Sept 8. In the state polls, he peaked somewhere around Sept 14. The shift back toward Obama is now starting to show in the electoral vote projections. See in particular
http://election-projection.net

____________________

Vicente Duque:

Mr Blumenthal :

I have seen an Awful Disparity between the red and blue maps for U. S. states on several websites and bets on candidates. Maps are based on the latest polls, I assume.

Betting money on presidential candidates is a very risky operation ( toss up states )

I assume that bettors have inside information, they look constantly at state polls, they live in the State for which they bet, they work for campaigns or know campaigners and the mentality of folks in rural areas, they know the towns and the weaknesses of pollsters.

I would love if someone addresses these disparities and contradictions.

Vicente Duque

____________________

hyh:

National polls that are internally consistent are Gallup Daily and Rasmussen Daily only. With any other polls (including other dailies, which started too late for this type of analysis), it's somewhat like comparing apples to oranges.

According to Gallup, McCain peaked on Sept 8 (for calls made between 9/5-9/7). Plot for Gallup could not be better for showing convention bounces even if someone made up the numbers.

According to Rasmussen, McCain peaked on Sept 12 (for calls made between 9/9-9/11) and stayed there until Sept 14 (for calls made between 9/11-9/13). Rasmussen numbers for reports published between Sep 8 and Sep 11 are a bit bumpy.

I just noticed another thing about Rasmussen - was their numbers for calls made on 9/11 a big "outlier"? Interesting coincidence...

____________________

RaleighNC:

Very quick to include the poll that made IN go yellow, but slow to add the polls that will make it go back pink and make PA go yellow. Disappointing.

____________________

My fellow Americans,

I'm here in the Crawford taking a break from doing analyzations of our economical miasma. Before I go out for a round of golf, I just want to fulminate about my extreme disappointedness in Pollster's lack of interest in my job approval.

I'm the decider and I have decided that it's time for an update to my Presidential Approval graph!

Now watch this drive...

____________________

Mark Lindeman:

@RaleighNC: Pshaw. It isn't as if Pollster.com has a staff of ten.

____________________

thoughtful:

@Mark

If Florida is considered battleground why is it still pink as a number of repected recent polls have it a tie. it should be yellow surely.

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR