p1db measurement network analyzer

I measured the ouput P1dB of an LNA using a power meter, spectrum analyzer, and network analyzer. All three methods gave different results. The differences are greater then 1dB which is very significant when it comes to P1dB. What could cause the discrepancy? Which method is the most accurate?

Which is most accurate? Probably the one that measures P1db of the fundamental frequency and ignores any harmonic power. Spectrum analyzer is good, network analyzer IF it is a vector analyzer is good (assuming it can do power [dBm] measurements, not just ratio measurements).

Network analyzer that is a scalar analyzer or a power meter will meaure total power, including harmonics. Harmonic power will be pretty high around the P1db point, leading to a false measurement.

At least the power meter and the spectrum analyzer should give exactly the same result. I would first check the accuracy of these devices with a synthesizer.

Harmonics may explain the differences, but often the cause is the instrumental accuracy.
Well calibrated power meter measurement is the most accurate.
Traditional SA give an accuracy +/-1.5 dB (agilent PSA is better)

In fact , to measure a power very accurate is to use a well calibrated power meter and a sharp filter around carrier. If this filter and power meter are calibrated, the harmonics may not disturb the measurements.

I guess it's a well method..