That 'study' didn't quantify the effect of no lubrication or even the effect of sprocket size or chain tension, for that matter, which they claim is a bigger factor for bicycle efficiency.
We are talking about - at best - a 1 watt incremental gain with a ceramic bearing set from a manufacturer of the highest repute. I'm sure that would not have fallen within their materiality threshold.
At any rate, seems silly to question the efficacy of chain lubrication. There's a reason not to ride rusty chains, despite what a largely irrelevant, unrigorous article tells you.
Fairly obvious that an unlubricated rusted chain will be less efficient than a lubricated chain. It's not so obvious that a "well lubricated" chain will outperform a "poorly lubricated" chain.
As for the rigor of the Study, it was done at Johns Hopkins, and reported in a peer reviewed journal (Journal of Mechanical Design).
I think you'll be hard pressed to find much data to show that the quality or type of lubrication affects the efficiency of a chain.
Keeping the chain clean and well lubricated appears to be more about reducing wear, than increasing efficiency.
Also do you see the irony in your assertion given that you're rejecting any potential benefit from Ceramic bearings out of hand, while you're asserting an efficiency benefit to a clean chain, without data?