How would a CRT scale the vertical height? Maybe you get a bit more space between lines, but the beam is only as tall as it is, right?

I would think the scaling would come horizontally ... you can put as many pixels horizontally as your RAMDAC can manage, but screen width per pixel depends on the pixel time vs the line time.

Yes, you do get more space between the lines, an effect is known as "scan lines". It is specially notorious when forcing a 240p mode like 8-bit and 16-bit systems used to do, which gives old games a distinctive look that some retrogamers crave.

https://en.wikipedia.org/wiki/Scan_line

Yes, you’re correct. I left out all of those details to try to simplify the explanation as much as possible. The reality is that the pixels we were drawing got narrower as our display hardware got better. In this VGA text mode the width of the pixels is 20⁄27ths of the line height. But you obviously don’t want to scale the 720 pixels of width down to display in just 533⅓ LCD pixels, so instead you have to scale the height up. It was just easier to start by saying that the height of the CRT’s pixels was 27⁄20ths of their width instead. :)