Supposedly 10% change in brightness is on the threshold of being able to visually detect the difference.
That being said, if I got another car that was 10% more fuel efficient, the difference might not jump out at me at the gas pump. But I would still want the more fuel efficient car. 10% improvement is 10% improvement, regardless of how detectable it is. Now how much is it worth to get that last 10%? I don’t know that I would bother recoating a primary for such a subtle difference. But diagonal generally isn’t that expensive, and is super easy, and can be used with multiple scopes. So even a 10% increase in transmission might be worth it.
Perhaps a reason to go with a dielectric coating.
But bear in mind there is a spectrum of transmission.
People studying carbon stars, red giants, reddish features on Mars or Jupiter, might prefer a silvered diagonal due to the higher transmission from 600nm up.
People studying CaK features on the sun might prefer a UV-enhanced aluminum coating on the mirror.
People doing infrared studies might want a gold coating, which is very flat well into the infrared.
For normal 450-600nm viewing with scotopic vision at night, though, it is hard to beat the dielectric coating for transmission.
The main caveat is that coatings don't influence sharpness of image, usually. It's the polish on the mirror that is the main determinant.
And that is often related to the price.
The second main issue for image quality is the collimation of the diagonal itself. This is something that is a pain to accomplish, but IS possible to adjust with a lot of trial and error.