I don’t think the bureuacrats at the BLS cook the data. But I don’t think it’s implausible to worry that census workers in the field might somehow–spontaneously, subconsciously, and subtly–inflate the “household” employment numbers when it looks like those numbers might decide a presidential election. Call me Jack. … If the new "establishment" numbers–reported directly by employers–confirm the relatively good news in last month’s household survey, though, paranoia will be difficult to sustain. …

If you can't actually understand that passage, do not adjust your computer.