I always thought loading was to dampen the stylus against any 'ringing' which may occur...
@quincy Nope! Especially with LOMC cartridges, which have a very high 'Q' factor in their coils (so as to minimize mass) the result is that the peak I mentioned can be as much as 30dB higher than the signal level. That will mess with a lot of phono sections- it can cause distortion and overload, which is why they sound better when the 'loading' resistor is installed.
When an SUT is involved you have a different situation! The RFI can't make its way past the transformer, but the loading of the transformer becomes critical to prevent it 'ringing' (distorting). If the load is too high an impedance value this will happen; if the value too low the transformer will roll off. If you change the source impedance (the cartridge) that impedance is magnified by the transformer in this case (because its a stepup) so the load value will change with the source impedance.
@lewm I suspect this is why small differences in loading the SUT can be heard.
This process is thus best done with a squarewave generator and an oscilloscope to see how well the squarewave is passing thru the transformer. Most audiophiles don't have access to such equipment and if they did, may not understand how to do this particular critical measurement. So as a result, most of the time its by gosh and by golly unless a lot of care is taken to eliminate variables (like cable capacitance) and also if the manufacturer was careful to specify exact values.