You don't measure output power directly by, for instance, measuring the temperature gradient across incubator walls of known physical properties and so on.
I would call that a first principle method. That's what Fleischmann and Pons did. They calibrated, but they also did the physics. So did Melvin Miles, McKubre and others. It is more rigorous to do both. But it can be daunting to understand! See the great long equation on p. 3:
https://www.lenr-canr.org/acrobat/Fleischmancalorimetra.pdf
My discussion of that equation, pages 21 - 23, 40 - 41:
https://www.lenr-canr.org/acrobat/RothwellJreviewofth.pdf
Fleischmann emphasized that you always have to calibrate, even when you do a first principle analysis.