I found two sources that show how the unemployment rate is calculated: http://www.bls.gov/cps/cps_htgm.pdf http://www.bls.gov/cps/rvcps03.pdf This is actually rather complicated and seems rather crude to me. The question I have is, if we used the same metric that we used in the thirties before all these changes that were made, what would the unemployment number be today? I estimate closer to 20% in 1930s style metric.
Shadowstats has a good compilation of alternate numbers, subtracting out some bls fudging. They show about 22%. http://www.shadowstats.com/alternate_data