No offense but this comparison video seems quite pointless 
I mean, it is supposed to show ... what exactly? In this test, it looks like the squished 1280 X 388 RAW videos would be better than regular 1920 X 1080 H.264 videos in terms of sharpness and detail. And that's just not the the case. Why did you downscale the H.264 footage? And why did you bulred it so hard? Why is the field of view so different in many shots? Did you crop the video, or did you actually changed the focal length?
I don't want to say that the comparison is horrible. But you should tell us exactly what we are seeing and what it is supposed to show.
That would make the video more interesting and helpfull.
By the way, if you film in the squished 720p mode, be carefull. It can give way worse aliasing than the unquished modes.
I would use it only if I needed a really shallow depth of field.
I downrezzed the 1080, mainly, to give it a bit of a chance. It looked horrendous when kept at 100% and, for web delivery, 720P is very common, so I decided that a downres would make the most sense.
I was filming in the squished 720P, there IS more aliasing/moire, but it is something I can definitely deal with for the extra detail in everything else. (Reolution baby!)
Some of the shots are not identical as I changed focal lengths to try and account for the crop that happens when filming RAW like this and, frankly, did my math wrong. (This is brought up in the youtube video description

) Oops?
The 720P mode needs much less data, so you can get more resolution out of it, compared to 1080P which limits you to near standard-def.