A Few Stats

By now you've heard about the August Jobs Report. And that it shit the bed. Surprise, surprise. It's even worse than the report that got Dr. McEntarfer fired from BLS. This report was worse. Way worser.

Ok. We all know it was Sleepy Joe's fault. All of it.

But, just so ya know, Here's the report. Some news organs give you more than the top-line results. Most don't give the number that really matters. As you likely do know: this report, and most government stats, are derived from sampling. No government is going to pay for monthly census data. Yikes.

The thing about sampling: it has errors. Census can, too, but that's another episode. Petty, Paranoid, Demented, Dictator Don and his merry band of fascists bitch and moan when the Damn Gummint data doesn't support their rhetoric. When you're a rabid right wingnut, that happens a lot.

Now, if you go through the report's tables, you'll see all the disaggregated numbers. What you won't see are confidence intervals for each number. What you can find is this

Statistics based on the household and establishment surveys are subject to both sampling and nonsampling error. When a sample, rather than the entire population, is surveyed, there is a chance that the sample estimates may differ from the true population values they represent. The component of this difference that occurs because samples differ by chance is known as sampling error, and its variability is measured by the standard error of the estimate. There is about a 90-percent chance, or level of confidence, that an estimate based on a sample will differ by no more than 1.6 standard errors from the true population value because of sampling error. BLS analyses are generally conducted at the 90-percent level of confidence.

Of some note: typical stat analysis uses 95% CI. Why use 90%? Because the CI is wider. Why is that a good thing? Because the purpose of the CI is to set the interval in which the sample estimate is 'guaranteed' to contain the true population value; for 90% of samples. Sort of. So, the wider the interval, the greater the chance the estimate and true population value will co-exist in that interval. But that also means that the sample estimate can vary over a wide range and still be 'signficant'. CI and stat sig are widely and wildly misused.

So, what do we have for the Jobs Report

For example, the confidence interval for the monthly change in total nonfarm employment from the establishment survey is on the order of plus or minus 136,000. Suppose the estimate of nonfarm employment increases by 50,000 from one month to the next. The 90-percent confidence interval on the monthly change would range from -86,000 to +186,000 (50,000 +/- 136,000). These figures do not mean that the sample results are off by these magnitudes, but rather that there is about a 90-percent chance that the true over-the-month change lies within this interval. Since this range includes values of less than zero, we could not say with confidence that nonfarm employment had, in fact, increased that month. If, however, the reported nonfarm employment rise was 250,000, then all of the values within the 90-percent confidence interval would be greater than zero. In this case, it is likely (at least a 90-percent chance) that nonfarm employment had, in fact, risen that month. At an unemployment rate of around 6.0 percent, the 90-percent confidence interval for the monthly change in unemployment as measured by the household survey is about +/- 300,000, and for the monthly change in the unemployment rate it is about +/- 0.2 percentage point.
[my emphasis]

As you see from the text, SE swamps reported change in employment. Frequently. And so it is for August.