Do you know how the job numbers are run each month? I do, because it was my job for a year right out of grad school in the late 1990s.
Each stateβs labor department and without federal influence (e.g., Labor Market Information division of the Texas Workforce Commission in this state) has a section devoted to:
Independently calculating the number of jobs by sector in its own state, based on a sample of data from individual employers who are required to report their staff on the payroll. These calculations are done twice monthly, with the second revised estimate simply having more data, allowing for greater accuracy as the month passes. This is a "job count" by state, county, and industry, and is continuously verified by calling HR departments around the state, cleaned up, and revised with a final "benchmark" at the end of each year.
Surveying households to calculate employment and unemployment estimates on whether or not people are working. This survey is then applied to Census data, estimating the number of people over the age of 16 years (working age civilian labor force) by state and county. This is a "people count" by state and county, with major revisions only conducted each decade after the decennial census.
These figures come from the individual analysts at each stateβs employment commission. The only federal oversight occurs during the "benchmarking process," where analysts from the US Bureau of Labor Statistics fly in and pore over the numbers with the party responsible for each dataset. For example, I personally was responsible for defending a year of my own numbers in a two-on-one meeting with federal analysts.
None of this is to say that they are precise, ball-on-accurate data. Still, I can affirmatively attest that all the calculations are performed with zero federal influence and only a modicum of federal oversight.