I did place the level on the sides of the head and seat tubes, eliminating the possibility of touching any welds with a ruler as a spacer, and the readings confirmed the ones that I got with the level on the ends.

I spent two whole days figuring this out, checking, double checking, triple and quadruple checking everything making sure that it wasn't all just a big mind-eff, even going so far as to swap in a fork and a handlepost from another bike that's years newer, and got the same result. I put the bike in question's handlepost and fork on the other, newer, bike and nothing changed on that bike.

Besides, I think the deviation is too great to be caused by a poorly made handlepost. I mean, just look at it:











It's obviously not something that most people would notice or even look for, but it is clearly visible. To answer an earlier question, I started having doubts some years ago after one of our cats tried to take the bike for a ride while it was leaned against a chair and knocked it over on its left side. At the time, I had a very short, one-piece handlepost on the bike, along with a high rise Joseph Kuosac handlebar:




The bike hit the ground at the bar end, with the handlebar taking the brunt of the impact. Not long after, I noticed that the left side of the handlebar was higher than the right, with the damage being at the circled bend (I found this out by physically comparing it to another, identical handlebar that I know to be good.) Now, I have no way of knowing whether this handlebar actually did get damaged because of that fall or if it was already defective when delivered, but I have no reason to think that the latter was the case.

So here we are. Whether the frame got twisted because of that fall or it was defective when it left the factory is anyone's guess.