Yesterday 9:17 am
Blackvegetable » 7 minutes ago » wrote: ↑ The establishment survey (Current Employment Statistics) has a much smaller margin of error for measuring month-to-month change than the household survey. For the total nonfarm employment change to be considered statistically significant, it typically must be at least 122,000 jobs. [1, 2, 3]
Key Margin of Error Statistics * Standard Error: The standard error for the estimated over-the-month change in total nonfarm employment is approximately 83,000.
* Confidence Interval: A 95% confidence interval for monthly job changes is roughly +/- 162,500.
* Statistical Significance Threshold: An over-the-month change of about 122,000 is required for the change to be statistically significant at a 90% confidence level. [1, 2, 4]
Annual Benchmark Revisions
The "total survey error" is gauged annually through benchmark revisions, which replace sample-based estimates with near-universe counts from unemployment insurance records. [5, 6]
* Historical Average: The absolute average benchmark revision for total nonfarm employment over the last 10 years is 0.2%.
* Recent Revisions:
* March 2025: Final revision was -898,000 (-0.6%), reflecting a significant overstatement in initial survey data.
* March 2024: Final revision was -589,000 (-0.4%). [6, 7, 8, 9] For more details on how these calculations are performed, you can consult the [CES Handbook of Methods](https://www.bls.gov/web/empsit/cesfaq.htm)
I just call it "Liberal Math"!