I have to agree with Lawtonfogle on this. "Normal people" (which is a No True Scotsman fallacy, BTW) don't understand U3 unemployment because it's a wonky term with a precise (and possibly even counter-intuitive) definition and the people that "feel" something other than U3 are more than likely to not do basic research, the kind of people that don't bother to take personal responsibility and/or don't bother to find a way to train themselves for a better job (which likely makes them self-selecting for U6 unemployment). Isn't it more likely that the U3 number is simply a more useful number for the class of people that wonk-talk about employment statistics and set policy? It's more targeted than U6 (which is a catchall for "people who are not currently employed or underemployed and might or might not be looking for work"). It's the same reason the wonks tack on "seasonally adjusted non-farmworkers" when talking about these figures -- those terms mean relevant things and fundamentally change the population of workers you are talking about. Do the "normal" people concern themselves with the intricate details of Fed monetary policy, growth and yield details of equities, bond yeilds and their effects, foreign policy beyond the large military conflicts that the US is currently involved in? Nope. Why should they know the difference between "U3 unemployment" and "U6 unemployment" if they don't bother to learn? I agree with higher-raked comments that the U3 unemployment numbers are useful if (1) you consistently use U3 in all reporting (don't jump from U3 to U6 to other metrics), you talk about trends and not the absolute number, and you describe which one you are using and what the caveats are of using that number. For reference, the Wikipedia graph of U3 - U6: https://upload.wikimedia.org/wikipedia/commons/9/9b/US_Unemp... U3 and U6 track very well with each other. There are permanently structurally "unemployed"/"underemployed" people in the country. It doesn't seem to budge much, whereas the U3/U6 numbers do seem to change over time. This shows that U3 is a decent trend predictor of U6, suggesting that there is little difference between U3 and U6 when talking about trends. Also note that U3 data goes back MUCH further in time, which is another suggestion that it is a long-standing convention and not "sleight of hand".