It sort of depends on exactly what you are testing and why.
For instance;
Just want a 'reference' reading to see how your signal is doing today? Any sized antenna would work, but it would always have to be the same for each time you did a check, today, tomorrow, next week.
Want repeatable and 'standardized' numbers for determining how 'much' signal is radiated by some antenna? How about a 1 meter antenna at a known distance. You would have all the information necessary for computing the E/I of the signal received at X distance through a 'standard' length of conductor. 'Course, that would depend on your' knowing what the inputed signal was to start with, and the characteristics of all devices/appliances/feed lines/etc/etc. between the transmitter and the point where the antenna ends.
The 'bad' part is that you have to know all of that stuff to get any kind of meaningful reading, which would provide some 'meaningful' numbers for making comparisons. ... ... Or! Do like the rest of us (mostly), do like 'M.C.' said, use a 'standard' sized antenna at the frequency in use, get the readings, then let the 'other' guy do the math. Much easier! What if the 'other' guy makes a wrong calculation or assumption? Hey! That's his problem, not yours! - LOL
- 'Doc