The New York Post is nobody's idea of "the liberal media", so when they come out with something like this it tends to catch my attention:

New York Post Online Edition: business: " ON Friday the government will report the latest employment figures for the nation. And while I won't - because I can't - predict whether the numbers will continue to show a miserable job market, you can count on one thing: The data will be misleading.
Here are some reasons why, good or bad, the U.S. Labor Department's 8:30 a.m. announcement will be as credible as the fortune you get in those Chinese cookies.

• The government missed a whopping 440,000 jobs that were lost last year. Why should this year's figures be any more accurate?
• Last January the unemployment rate was 'adjusted' downward by 0.2 percent for changes in surveying methods. Right now the closely watched rate should actually be 6.6 percent, not 6.4 percent. If the rate does drop - as some on Wall Street are hoping/predicting - so what?
• The government recently started seasonally adjusting its employment figures each and every month. Washington may as well let the numbers be picked by a Lotto machine.
• Back in the 1990s, the government changed the questions it asked in its household unemployment survey; more recently, it lowered the number of people it canvassed in chronically underemployed inner cities.
  The result, not surprisingly: an unemployment rate that is lower than it was in the last recession more than 10 years ago.
• In a less widely watched section of its report, the government is reporting that the unemployment rate in June was 10.6 percent, when you include people who are too discouraged to look for jobs and/or not fully employed.
  The figure would be worse if the government hadn't booted millions of people from the discouraged worker category into a no-man's-land where they aren't counted at all.
• The two surveys the government conducts aren't even close in their picture of the job market.

The survey of households, from which the unemployment rate is calculated, reports that there are 138 million jobs in this country. The survey of employers counts 129 million.
Even when you adjust for things like one person having more than one job, the figures can't be reconciled to within a million jobs of each other.
That error could be due to the fact that people have a tendency to lie to surveyors, especially those poking around for the government. Or it could be the result of hundreds of thousands of phantom jobs that Washington counts because it thinks the employment situation can't possibly be as bad as it is. "

I'm now curious whether there's a source for the last couple of years' worth of unadjusted numbers, or even better, a set of numbers that were seasonally adjusted following the old scheme so they could be more readily compared with older data.