We've found that a 3uV full scale Null meter is your best friend here to look at the differential voltage across the connector, generally easier, faster and more accurate than any DVM for this test. Generally you test close-mounted pairs of connectors because that's how they are most often used in-circuit.
It's important to note that you want the connectors mounted to a whatever substrate they will be mounted to because the only thing that is going to generate thermal EMF trouble is if a pair of connectors are at DIFFERENT temperatures. Generally you don't give a whit what the connector thermal EMF is if a pair of connectors in the same circuit are at same temp. In other words mount the connectors on a good thermal conductor and you'll have far less thermal problems, if any.
It also makes a difference what wire you're using at the connector and how fast its thermal flow is contributing to a temperature difference at each node.
Of course air drafts, ambient temperature, and so on all make things better or worse. The mass of the connection joint area also sets the reaction time of the join.
So the concept of "Testing a Connector" really applies to the whole connector, wiring and enclosure SYSTEM and how and where it will be used. Sometimes you find it doesn't matter a whole bunch - and there are much bigger errors in the system.
Suggestion: While testing connectors, jacks sockets etc., if they are threaded: use a torque wrench for repeatable torque force on the connection. The clamping force on the conductor and how compression force is distributed makes a difference. Generally resistance measure circuits want a smaller contact area (to reduce changes in R especially in lower ranges), higher amperage flow wants a larger contact area (to lower self-heating) - but use what works best for your application. When testing connectors, you also want to see how -repeatable- the connection is, not just thermal EMF. A clean pure copper to pure copper connection is best, snug and hopefully gas-tight in the metal to metal zone but so the wire isn't extruded or bent. Everything else goes downhill from there.
You'll probably find out that one of the best connections ever invented is just two ea. 24AWG solid -clean- copper wires well twisted together...
Telco wire / shielded Cat5e wire can work for lots of stuff at DC. If chasing PPM's then teflon insulation will certainly help solve leakage current problems, but silver-plated wire may or may not be an asset.
RE: Banana plug - not normally used for low PPM work - they really aren't that repeatable as they age and the plug spring force drops off and inside plating wears away from inside the socket. Binding post is usually preff. for high accuracy / repeatability.