Do they adjust for the bias introduced by using a phone survey? I would hypothesize that job holders are more likely to have a stable phone connection than the unemployed.
Having been on the business end of a BLS survey for a year, they do more than call. They visit in person for an in-depth interview at the start, then follow up by phone for a very detailed interview every month for a year. If you're not available by phone (or don't call back, or they can't get in touch), they will send someone out to find you.
The questions are very detailed and probing, and they have additional questions each month related to some special surveys that they conduct (e.g., food spending and availability). We'd typically have to spend about 20 minutes giving answers.
The methodology seemed pretty sound to me, certainly not anything that would be skewed by phone access. (It might be skewed by people being completely off the grid, for instance, people not measured in the US Census, but I would guess their methods can correct for that.)
Do they adjust for the bias introduced by using a phone survey? I would hypothesize that job holders are more likely to have a stable phone connection than the unemployed.