[3Fingas]

There’s no way the Communist director of the BLS will allow the unemployment number to be any higher than 7.8% before the election. If anything, it will magically decrease a couple of percentage points.

[null and void]

OK, I read this to mean the new method will make Obama’s numbers look even worse.

I’m actually not happy with that, I would prefer that the methodology not be changed right before the election, at worst, I’d like to see a side-by-side comparison of the old and new methods.

Can’t say I’d be disappointed in the slightest if this results in Obama losing. (Ecstatic is the word that comes to mind) But I will say that a consistent baseline is preferable.

[AU72]

He who lives by fudged numbers, dies by fudged numbers.

[ozzymandus]

The numbers will be bogus to help Obama, like they always are.

[FredZarguna]

http://confoundedinterest.files.wordpress.com/2012/10/civempnfp.png?w=604&h=362

The ADP numbers are actually the ones the BLS will (eventually) use to "correct" their own household survey, as this chart from the Federal Reserve shows. Take a look at the month that has one of the largest discrepancies between the two: September 2012. ADP doesn't have anywhere near the fluctuations (can't be manipulated by bad Census surveys, either) that the Household Survey has.

The ADP numbers track very reliably to true growth. We are creating around 100-120,000 jobs or fewer per month and that's what you expect from 2% GDP or lower.

[griswold3]

My prediction for Friday is +72,000