Well, if you really want to get persnickety about it, it’s not “Kelvins” either. It’s “kelvin”. The word is not capitalized (although the symbol K is), and there is no -s on the end.
For many decades, the Kelvin temperature scale coexisted with the unit of thermodynamic temperature called the “kelvin”. Many well-edited science publications were still using “degrees Kelvin” in favor of “kelvin” well into the 1970s. So if it takes a few extra years for the rule to trickle down to a non-technical writer like Cecil, I don’t see a big problem. (Note that the cited column is from 1981).
Also note that just because the SI fails to approve of something doesn’t mean it ceases to exist. There are many other units and scales in current use in technical publications that are not sanctioned by the Système International d’Unités, including for instance, the Richter scale, the day, the electron-volt and the liter/litre.
IMO, the biggest advantage of an absolute temperature scale is that you don’t have to make that distinction. Is saying that my body temperature is 310 kelvins any less correct than saying that my height is 185 centimeters?
My understanding was always kelvins plural, though i suppose i’m in gross error about the whole capitalization thing. A quick survey of the literature (thank you google) seems to suggest both are in usage.
I always thought the SI rule was that, strictly speaking, the plural is always the same as the singular. Hence 273 kelvin, 110 kilogram, 1.95 metre, 4.3 joule, 12 volt, etc. Apparently I was wrong and there is no such hard and fast rule. The National Institute of Standards and Technology says
I’ve typically seen it when referring to a combined unit of measure to append the plural to the numerator.
eg. 5 Watts per square meter (W/m[sup]2[/sup]), 15 kilogram-meters per second squared (kg·m/s[sup]2[/sup]), etc.
This is not necessarily official SI notation (I have no idea), but rather the use convention through practicality. We tend to think of anything over 1 as being plural, and so try to pluralize the units. So 1.0000000000001 meters. Because of that type of thought, often people will refer to fractional amounts in plural units, i.e. 0.4 meters, not 0.4 meter. I think it’s a linguistic thing where we think of 1 as “a” something or other, but when we append other numbers to the front, we think as plural.
I can think of several reasons to throw in the "degrees:. E.g.,
When writing to a general audience. They might not know what a Kelvin is. Thus I support Cecil’s usage.
When talking about temperature differences vs. absolutes. “This tea is 5 degrees Kelvin too warm.” Leaving out the “degrees” might leave a few people thinking the tea is near absolute zero and yet somehow still to warm.