I would add to that:
Perform 3-4 successive tests on the same sample each time you test, at least for a while, then do that occasionally ongoing to keep things honest. This will help identify test-to-test inconsistencies in the operator-device system. Learn from such inconsistencies, e.g., Did bubbles on the device cause it? - then remidiate the procedure accordingly. This is not exactly a full measurement system analysis, but imo will get to the heart of what we need as aquarists from our salinity measurements.
I was able to get pretty consistent results from my IO swing arm, but it was a bit finicky (swing arm bubble entrapment) and I much prefer the TM glass precision hydrometer.