@Dbeasy I've tried to explain this stuff to him a couple of times before, but he completely ignores the explanation. I think he just prefers to live in a conspiratorial fantasy land.
There are three things going on with employment numbers and job figures that get published by the BLS. This was my first job out of grad school for a year, so this is all from first-hand experience.
Current Employment SURVEY (CES): As you can guess, it's a monthly survey of employers asking how many employees they have at a given time. It's simply an ever-shifting estimate by geography (states and metro areas, not counties) and by industry sector. The first guestimate (that's really what it is because response rates can be low) and the second revised estimate to fill in the gaps later in the same month are done at the STATE level, not done by the feds but reported to them. It can get tricky because it does not account for business openings or closures. Also, the survey is entirely voluntary. So, some employers only intermittently will respond.
Quarterly CENSUS of Employment and Wages (QCEW): This is your job count. These data are real, hard figures that are required to be reported on a quarterly basis for each month: headcount and wages by location (states and counties). It's an arduous process that often requires STATE workforce commission staff to call up the HR departments of individual businesses when numbers look funky and don’t match up well with previous reports. Because of all that, there's a six-month lag before they are publicly published.
As the actual real job counts come in, the estimates go through a revision process to bring them in line with the hard count, culminating at the end of each fiscal year with the final "benchmarked" numbers, which is a big to-do. Now, both of those above are concerned with JOBS, an important thing to keep in mind.
Finally, we have the monthly employment/unemployment figures through a program called Local Area Unemployment Statistics (LAUS). I was not directly involved in this, so I can't say as much. That said, I worked in a giant cubicle farm in a converted warehouse on Rutherford, east of I-35) with other analysts who did this stuff. It's a household survey of people, basically asking them one question: Are you currently employed? I think they get the contact information from folks who've lost a job and have filed for unemployment compensation.
At the local level, my impression was that these data are a lot more squishy and are only "benchmarked" once a decade when the US Census results are published. That's because it's the only time analysts can get a real handle on the Civilian Labor Force (persons 16+ years of age) living in a county.
So, basically, you have the universe of people over that age (CLF) taken from the decennial Census and whether they have jobs (Employment). You take the difference (Unemployed) and calculate the Unemployment Rate. From there, they just apply household survey results to keep track of the number of unemployed people in each county. The problem is that it often doesn't capture significant population changes in a county, and so by the end of the decade, those figures can be pretty off and aren’t fixed until the next Census comes out.
So, the first two (CES and QCEW) programs are all about jobs, while LAUS is all about people. For labor market and economic data geeks, that's an important distinction because a person can hold more than one job at a time.
I hope this doesn't seem like I'm lecturing, especially to you Dbeasy, but I just wanted to lay this all out there to provide transparency about how the foundational figures from which just about all employment data are calculated, so folks don't fall back on paranoid (and lazy) conspiracies.