I suspect tomorrow's going to be a bad day for tracking poll addicts here...

[regnaD kciN]

As I mentioned earlier, I have been following the Gallup Daily Tracking poll for some time. Thanks to a poster at DKos who actually had the daily data for several days in a row, I've been able to construct a spreadsheet that takes the current three-day average and figures out what the most recent day's results have been.

In the GDT, Obama currently holds a steady 49-42 lead (and actually increased it by one point over the day before). But looks can be decepetive. The result for today is due to a mediocre Sunday total (a 48-48 tie) dropping off the average and being replaced, I calculate, by results from yesterday putting Obama up 48-45.

However...the big reason for Obama's lead was a big day for him on Monday, when he was up 51-37. That number comes off tomorrow, and will be replaced by the first post-Palin speech results.

Already, taking a two-day average of results from Tuesday and Wednesday, which remove the Monday total, we find that Obama's lead will have shrunk to 48.0-44.5, a margin of only 3.5 points. Even if today's results come out another 48-48 tie, that will still drop the rolling average from today's 49-42 to 48-46. And there's good reason to believe that Obama will be lucky to merely get a tie today. Granted, it would take something like a 52-47 McCain lead today to tie the overall race in the GDT, but, with the media feeding frenzy over Palin's "grand slam" speech (and the fact that almost as many people watched her as Obama) that's not impossible. In any event, it's likely that the race will be within the MoE at best, which means we'll have to endure another onslaught of "McCain's comeback!" and "Sarah wows 'em!" in the media the next day.

I'll only point out what I did last week -- it's highly likely that the polls following McCain's "convention bounce" will show him in the lead. No reason to panic. As others have shown, convention bounces subside after one or two weeks, and I don't think any of us expect this to be anything other than a horse race from here on in.

In any event, don't shoot the messenger.

[Kdillard]

1. We'll see what happens. I am more concerned with his EV count and whether he has a

strong base support to put him over the top. End of September early October after the debates is when I will start worrying about the tracking polls if they don't show movement in Obama's favor.

[regnaD kciN]

5. That's the real key...and what the Obama campaign is focusing on...

n/t

[Gabi Hayes]

2. read this. daily tracking polls, especially gallup, are WORTHLESS, according to David Plouffe:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x6841885

[regnaD kciN]

4. That's the point I made in my first post on the subject last week...

...but tracking poll addicts are tracking poll addicts, and live or die by today's Gallup and Rassmussen numbers.

As I said last time, I hope we don't see too many "Chicken Little" posts over the next couple of days...but am fairly sure we will.

[the Sundance Kid]

3. thanks for the heads up, poindexter

I'll be sure to get all worried.

[writes3000]

6. Didn't you do this same post a few days ago? Weren't you wrong?

You may be right this time but I don't get the point.

Just to say you warned everyone?

[BzaDem]

9. The poster is simply breaking the polls down to day by day. That is helpful.

Tomorrow, Obama would have been likely to lose points even if Palin had not spoke, because a 14 point lead from Monday would have dropped off the average. That is good information to know.

[WeDidIt]

14. And again. I call bullshit

If this average is as erratic as you claim, the poll is compeltely irrelevent and has a 20 point margin of error.

[BzaDem]

18. Do you even know what margin of error means?

The margin of error they report is not the margin of error between the poll average and TODAY's support, it is the error between the poll average over the last 3 days and the average support for Obama/McCain averaged over the last 3 days. The poll would only have a 20 point margin of error if the actual 3 day average swung 20 points one way or another, which no one is claiming.

[WeDidIt]

7. I say you're full of shit.

It's a three day rolling average. Look at the past week, brainiac:



No way are you're numbers right.

[BzaDem]

11. You really don't know what you are talking about.

The only reason that 3 day average look stable is because some huge Obama days (14 point and 16 point days) are averaged in with lots of near ties. Tomrorow, one huge Obama day (14 points to be specific) is going to drop off the average, and a possible better-for-McCain day (due to Palin's speech) might replace it.

[WeDidIt]

13. Bullshit

Plain unadulterated bullshit.

There should be a Palin bounce, but it will take three days for the full effect.

No way is that 3 day averageas erratic as you say or else the margin of error would make the poll comp[letely irrelevent.

[BzaDem]

16. Why don't you actually compute the data on each of the last several days and then try again.

I thought about presenting it myself, but I figured that if you couldn't compute the data yourself you really have no credibility to talk about how tracking polls work.

[WeDidIt]

19. You really have no fucking clue what you're talking about. n/t

[GrizzlyMan]

8. Not gonna shoot you, but didn't you post this exact same thing last week

I distinctly remember a poster going through the three day averages and saying that Obama's "bounce" would be minimal and that the race would be tied by the start of the RNC convention. As we know, that didn't happen.

[JoshDem]

10. Agreed--and the next two days even worse

McCain probably ahead by Sunday but I still think its a dead even race.

[Rosa Luxemburg]

12. people will be eating cotton candy for a few days

and then the cotton candy machine will break down.

She

[Medium Baby Jesus]

15. Where are the "concern" police?

[ClayZ]

17. Track Polls here:

http://www.fivethirtyeight.com/

[book_worm]

20. It's inevitable that McCain will get a boost from his convention

and then it will begin to gravitate back to Obama as that wears off. People want change and McCain doesn't represent it.