Its actually very easy to do real World comparisions of antennas, you do need some LoRa nodes, one as a transmitter and another as a receiver.
No need for an Anechoic Chamber either, your not going to put the Gateway in an anechoic chamber so why test it in one. All you need is a large open field, a transmitter and receiver and some long bamboo poles, oh and its helpful if you have built a 1/4wave vertical with radials antenna as reference.
In simple terms you mount the transmitter, which sends out test packets, on a large bamboo pole, say 2M off the ground, or more if you want. Then with the referance antenna in place go 50m or 100m away and measure the RSSI of the received packets, the antenna on the receiver does not matter as long as you use the same one for each comparison. Put the receiver on a pole too so you can stand out of the way. A display on the receiver helps, or you can just use a long serial lead and laptop to read the receiver output.
Swap antennas and do the measurement again. You now know the real world performance differance, in dBm, between the two antennas. And your measuring what you want to know, the performance at roughly horizontal, outdoors, in the real world. Want to put the test transmitter on top a roof, then do so, go a long distance away, read the signal RSSI, swap antennas and repeat.
Some descriptions of past testing endevours are described here;
https://stuartsprojects.github.io/2016/04/01/testing-and-comparing-transmitters-receivers-and-antennas.html
The link testing software, including the decending power level stuff is in my library.
If you use the same large open field, referance antennas, heights of antennas, software etc, you can go back a year later and run the same tests so you can compare the performance of a new antenna.
There was once an antenna design touted as an easy to build good gain job. It was designed by ‘someone’ who has given talks on propagation at TTN conferances. The antenna was indeed fairly easy to build, just bits of wire some straight and some coiled. The SWR match was very low indeed, so must be a real good antenna ? So over to my local playing field I went, and compared it to my 1/4wave vertical with radials. This easy build antenna was rubbish, somewhere around 4-6dBm down on the much smaller 2dBi referance. Why such a poor real world antenna, no idea, maybe it was radiating up at the heavens.
And if you reduce the power output of the transmitter enough, put it in a tin or wrap it in foil, you can use a receiver, with a beeper fitted, that you fit with the antenna to test and walk away until the receiver stops beeping. The distance comparison between antennas gives you the relative gains. The technique is describe for 2.4Ghz antennas here.
https://stuartsprojects.github.io/2019/04/27/Antennas-for-SX1280-2-4Ghz-LoRa-tranceivers.html
In that test it literally took 10 minutes to do a real worl comparison between 6 antennas.
There is more examples of using ‘ping’ testing for antenna\receiver comparisons here;
https://stuartsprojects.github.io/2015/01/05/Semtech-LoRa-Transceivers-a-KISS-approach-to-Long-Range-Data-Telemetry.html