Lee says his data is from powder companies, which are a better source than bullet company data because the powder companies have nominal characteristic reference lots to test with and SAAMI standard pressure/velocity guns to test in, in most instances. Info from other sources often is developed in production guns then the maximum load is sent for independent testing afterward to make sure it doesn't exceed the SAAMI MAP, but I don't think they usually adjust it upward if it is below SAAMI MAP. Also, bullet companies have a habit of listing one set of data for several bullets of the same weight, and you know from the different shapes and seating depths that they won't all behave exactly the same. So you can guess that the bullet that caused the most pressure probably governed the load limit for them all.
However, the above factors and allowances usually (but not always) mean the bullet company data is wimpier than the powder company data. So I think Lee's low load levels must indicate that they took the powder company data and reduced it by a fudge factor. Possibly this was done because a lot of Lee's measures use powder VLD numbers (cc's/grain) to get loads by volume. That would mean they had to allow for lot-to-lot bulk density variation, which can exceed 5%. I think they may have erred low in some cases where a charge fell between their available powder measure bushing or powder scoop sizes, too.
I don't really know the above without asking Lee to confirm it. I usually do look at bullet company loads for their bullets, then see if I can find something like it in a powder company's data. I usually use the powder company's starting load and see, after allowing for any difference in barrel length, how my expected velocity from their data compares to what I actually get. If my velocity is lower, due to difference in the gun, case, and primer, then I'm pretty confident I am operating below their starting load's pressure, and vice versa.