I don't know how many times I have imagined pulling what's left of my hair out after reading something like this:
OMG Bush is at 45% according to Gallup, yesterday Time had him at 37% oh NOOOOOOOOOOOOO!!!!!!!!!!!!!!!!!
Or it might go the other direction and instead of deep despair the "trend" throws thousands of kos-people into fits of religious extacy.
If you're NOT one of those people who are hypersensitive to these types of shifts then you probably know what I'm going to say ... or you might not. If you ARE one of those people, keep reading ... or I will track you down individually (virtually) and I will not be so nice about educating you next time.
I apologize for coming across a little brusque but this really, really needs to end.
Did you know that if Gallup and Time polled on the same day, they could easily show that Bush's approval is simultaneously 37% and 45%....??? What kind of weird quantum mechanics could possibly cause that? PAGING DARKSYDE TO THE THREAD!
Wait -- no, we don't really need him. What we really need is to look at all the data, and soon we will realize there's no fancy physics involved at all. Bush's approval isn't a "point" -- it's a "cloud".
Check this out:
Every time you see a new poll result, click on Professor Pollkatz's Pool of Polls. Then you will temporarily be un-blind and the elephant will look like an elephant. (It's updated periodically.)
If you study this graph for more than 15 seconds you will notice a few things: ABC/WaPo (blue diamonds) registered the highest approval ever for Bush, and they tend to give him higher numbers than the rest of the pack. American Research Group (yellow diamonds) not only tends to report lower numbers but based on my amateur analysis they look like a leading indicator to me. Gallup (pink diamonds) clearly is not polling a reality-based community.
Note also that this year's molehill looks a lot like 2004 ... from a low point around April/May the downward slide reverses, peaks around November, then returns to its natural Bushian (downward) progression. This also corresponds with heavy-handed PR campaigns emanating from the White House. If 2006 is like 2004 then we can expect Bush's approval "cloud" to increase by two or three more precentage points before the elections.
Now that I have cured you of the day-to-day apples-to-oranges comparison jitters, please take a look at one more graph that the good Professor makes available: Bush Approval: In Fact It's A Gas?
If you missed Jerome a Paris's recent debunking of gas-price manipulations conspiracy theories, I strongly encourage you to find it and read it before jumping to conclusions.
Moral of the story: look at all the data. If you don't have all the data, try to assess how much you're missing, and how much it might change your hypothesis. In other words, don't freak out until you're certain that you need to freak out.