I am evaluating SPDT switches with off isolation as a key criterion. I will be using a ±12V supply and switching between GND and an input signal. My test circuit is as follows:

Simulating with these conditions indicates slightly different off isolation to the datasheets (though this is not my primary concern):

What is astounding is that when setting the SA resistor to 0Ohm, the ADG1419 goes from the worst to best performing part.

Summary of results where the best performance for each set of conditions is BOLD:
- ADG1219
- Datasheet: -77dB
- Simulation: -74dB
- Simulation w/ grounded input: -74dB
- ADG1419
- Datasheet: -64dB
- Simulation: -70dB
- Simulation w/ grounded input: -95dB
- ADG5236
- Datasheet: -85dB
- Simulation: -83dB
- Simulation w/ grounded input: -84dB
Question: Is this a spice model error, or, can I expect this to be the best part for my application with a directly grounded input?