Figure in the table was actually "Adjusted" Gross Income... that's actually entirely reasonable to use but might get you some queer results if you looked at, say, poor students, considering tuition and student loans do play a role in AGI vs GI.

Also, with the 'bottom 50%' <$32,396 was that the bottom 50% of people, adults, income tax payers, households, or what? I mean, if by individual does it include the 16 year old kid who earned $5000 at their summer and weekend job? That's not a small percentage of the population. Nor are people who are mostly home-makers, students, retirees, etc living with someone else who is the principle bread-winner but do some small business or part time job. If it's household income we're talking about, things get really shaky because all else being equal a family of 5 making 50k should be better off than a family of 4 making 40k. There are, also, a lot of households consisting of one person. Jane and Jon were two households at 25, making say 20k a year and doing well, they get married, become one household at 40k, weirding the data, which further gets confused when he gets promoted to 35k and she stay at home tending their two kids and saving them money by cooking from scratch and growing vegetables in the garden and babysitting her sisters kids for free too, who then repays the favor when she gets stir crazy and takes on a part time job paying 5k just to get a few hours out of the house... how's that data looking? We didn't even add in the effect of their child deductions. Or when they get a divorce and Jon deducts his alimony form his gross for his AGI while Jane goes and lives with her sister and brother in law which might not even register as a single household. And it really gets confused if some live in the city where stuff costs more, but some other stuff is free, versus rural, especially because those demographics don't average out. You do have more single people living in the cities by percentage then outside them, Jane and Jon might have lived in the city before they got married and found a nice suburban or rural place, not even factoring in lots of others things, some of which are too politically incorrect for anyone to want to mention in a study, about various demographic differences between urban and rural. Another one in there, what about people who have brand-name obsession or suck at math or chemistry compared to otherwise identical people who don't? Ten bucks says that significantly effects relative standard of living and that generally the latter also earn more too.

I remember doing a grocery run with a good friend before a BBQ we were doing where he was compares 2-liters to cans and asked me which was cheaper because he didn't have a calculator and my friends all use me as one, I tapped that little usually upper-left-corner panel where it listed price per ounce for each and he just stared at me as he realized he'd seen that a million times before and never knew the implications of it, smart guy too. You can't even try to factor crap like that into the analysis and it definitely does not average out by income level, even if one ignored that sort of knowledge is pretty heavily correlated to age too and income is very tied to age. If we're fundamentally talking about trying to raise standard of living, the actual goal, as opposed to income, the very loose metric we try to measure that through, that kind of stuff plays a huge role and you simply can't ignore it if you want actual results as opposed to good talking points.

So the data tends to be a mess and the factors don't generally 'average out', people hand-wave that in to save time or thought or the ruination of their pet theory in much the same way a young lady might say "Yeah he picked me up and bought me dinner and a movie, but I have to spend a lot on clothes and makeup" This is true, but the odds are, whichever way that balance sheet works out, it won't have been zero. It's a hand wave, which is fine in personal affairs but is a damned bad way to make national tax policy