What is your definition of seriously? Dvorak has 70% keypresses on the home row, with the best left/right hand alteration approach possible, with ability to switch to Programmer version of it if needed. I don't know what real-world tests can leave all of these features unnoticed. Dvorak layout is a magnum opus of the guy who was willing to arrange the letters in a best possible way for human, while qwerty is a clever trick to fool a customer for the sake of selling typewriters.
From what I've seen, differences in key arrangement has effects on:
- Different keyboard layouts have very little effect on speed. Querty vs Dvorak vs a completely random layout will have little difference in typing speed after learning it for a reasonable amount of time.
- Learning time is probably reduced by using a more intuitive keyboard layout. This was one of the key points in August Dvorak's studies while researching the layout, and found that Dvorak was significantly easier to learn for new typists.
- Hand and finger movement is reduced by placing more commonly used (mostly optimised for english) keys on the home row. More words can be typed using only the home row keys on Dvorak and especially Colemak than on Querty, which also likely reduces rate of misspelling. This can also be an improvement for typists with RSI or other injuries.
Different layouts have different goals, as you've said Dvorak has a great approach for altering between left and right hands for commonly used letter pairs. Colemak optimises for inward finger rolls, which I might learn as they're very satisfying to type in the rare occasions they appear on my default of Dvorak. Layouts like Engram aim to reduce use of the middle column between the left and right hands.
While I can't unequivocally recommend alternate keyboard layouts to every person that types regularly, one thing I can say is beneficial for any layout is simply to remap Caps lock to Backspace for reduced finger movement. Think about how often you use Caps lock and then realise it's wasting valuable space on the home row, while usually requiring the typist to move one hand off the keyboard every time they want to correct a mistake.
If its that much better, but no one uses it, who the hell cares. You can espouse all the features you want, but if testing shows that it doesn't really matter to anyone, meh.
What is your definition of anyone? I am teaching Dvorak to some really young students, whose habits are not malformed with Qwerty. They are extremely thankful to me, no one is willing to return to Qwerty, time to gain a touchtyping skill is 10 hours from zero to hero. So, my "anyone" reports that Qwerty does not matter to him. Probably your observation is based on the fact that an old dog can not learn a new trick.