Unemployment, Again

The Washington Post has an editorial pointing out that in spite of best intentions, the BLS method of collecting unemployment statistics can, and does, produce conflicting data, depending on who is being surveyed.

The government doesn't really know the exact number of people with or without jobs. The number reported each month is based on surveys, and surveys often can have methodology issues. As it turns out, the surveys estimating the number of people with jobs reported over the past couple of years suffered from some really big problems. That's where government falsely claiming 824,000 more jobs than actually existed comes into play. Unfortunately, those adjustments have so far been made only through March 2009, and there are strong reasons to believe the survey data since then also need to be adjusted downward.

I bet one could get a perfectly good estimate of employment by surveying traffic on a major artery between a bedroom suburb, and the downtown area. You'd need to adjust your numbers if the road was widened, and use a factor to account for increasing or decreasing population, but this should be much easier to figure out than the relationship between sunspots and grant money for climate studies.