An oscilloscope is completely and totally the wrong tool for RF power measurements. "Utterly useless for that" is a phrase I would use.
Define "accurate" measurement. Accurate to what percentage?
I no longer have the setup because I no longer need it and sold it for a fair profit, but I used an Agilent E4431B RF power meter with sensor (I forget which one, E series model, anyway) and the setup was calibrated by a cal lab along with the 100 watt 30 dB pass-thru attenuator that was
always incorporated into the system. The entire system was calibrated as a system, and used as a system, and was recalibrated yearly but it received a "special" calibration, being calibrated only from 10 MHz to 1 GHz. I never needed more than that so that kept the cost of calibration down.
It was accurate to, if I remember correctly, 0.1 percent over its calibrated range.
Now as far as opinions go on test equipment, for precision power measurement, Agilent and Rohde & Schwarz are the only names that matter. Bird meters are suitable for field use if they are checked carefully against calibrated standards.
MFJ is adequate for CB and amateur radio but I wouldn't use them for any testing on commercially licensed equipment.
The big thing about using proper calibrated equipment, power meters in particular, is that a lot of radios are not very tolerant of having their output power turned up beyond rated limits. A poorly calibrated power meter could easily be off by as much as 3 dB (I've seen worse) and 3 dB is equivalent to halving, or doubling, power output. Use that meter to set power output on a 5 watt radio and depending on the direction of the error, you might end up with the meter saying 5 watts with actual power output being anywhere between 2.5 watts ( possible range issue complaints) and 10 watts ( if it'll do it, it won't do it for long!) so using an accurate power meter is essential.
Frankly, if you don't have a GOOD and CALIBRATED meter that's still within its calibration period, don't even BOTHER to do a power measurement and adjustment. The measurement error is likely to be more than the amount that the transmitter drifted since it left the factory.