Well, some FR difference is to be expected between samples, though it should be minor if QC is high. You allude to gear that a common consumer would use - does this mean he was using low cost cartridges? I would not expect QC to be as high on a $60 cartridge, as compared to a much more costly unit.
Another (cheaper and MOST accurate) way is to use the same cartridge and install it on both tables(take care of the mass issue, of course) and record the output from each one to a high quality sound card. Then level adjust and time synch the recordings with an audio editor. Now use a software ABX program to blind test the samples.
Most accurate, yes, but harder to do an A/B comparison on the fly. We synched the turntables and switched between them to decide which sounded better. The two cartridges we decided on were actually very close, and we listened to quite a few albums to verify this.
They were about $100 list price, which means they sold for about $60 and unless a comparison is being made, it's usually not going to be noticed as easily, except for the fact that they intentionally had track 1 EQd and placed the same, as a reference. He used a Niles A/B switch and while it would be a bad thing to do if he had one TT in one Phono in with the switch on only TT, since they were both running through it, it was a fair test. While it wasn't meant for purists, it wasn't just a casual test. He has a good number of Grammy Awards for his production work and owns a studio in the San Francisco area. It was interesting to be asked for opinions on slight changes made to the mixes, too.
"Keep in mind that one can easily employ DSP correction hardware and measure each sample and program a correction curve as compensation. Transparent high quality DSP hardware is readily available today. Behringer makes a couple of very high quality devices suitable for the purpose, that are around only $300 each, such as the DCX2496 and DEQ2496."
So much for keeping the signal path short and simple.