It doesn't matter. All that matters is that you compare the same level of unemployment historically. Including more groups in the umemployment number doesn't d

It doesn't matter. All that matters is that you compare the same level of unemployment historically. Including more groups in the umemployment number doesn't do anything other than scale & stretch the value. The trends largely stay the same, but the values are shifted up or down, and the "waves" move at slightly different rates The reason for using the U3 number is because it is a good snapshot into what is happening now and it's where the bulk of the movement is. It doesn't really matter if someone reports unemployment going from 9.2% to 6% vs 6% to 3%. That extra 0.2% comes from "long tail" of the unemployed. You can plot the different unemployment levels on the BLS website to build an intuition for why it doesn't really matter which figure is reported on, just that figures are compared against their same level.