Jobs added and unemployment rate are not directly related in their calculation. Jobs added is from a survey of 410,000 employers. The Unemployment rate is calculated from the BLS survey of 60,000 households out of 150,000,000. So the sampling rate is pretty low and potentially has much more room for error even if they did it right. In this case they have certainly produced an outlier as all statistical methods are capable of. Especially if the same size is such a small percentage (.04%). There didn't seem to be any mention of the people who stopped looking nor any distinction of what kind of jobs these people got.