I have a very simple question regarding parallel RLC networks (or any similar circuit). When given a problem where the reference voltages and current directions are allocated 'strangely', is my intuition correct to make all voltage references positive on one side (top) and the current directions consistent (down)? No equations or dependencies are given.
You can (in general) select the polarity of current sources and voltage sources any way you want. Then, when you solve the circuit (via nodal or mesh analysis) if you guessed wrong on any of the sources the voltage or current value will come out negative. This tells you the actual direction is opposite from what you picked.
Good practice is to choose initial polarities to be consistent (as you suggested) and then solve the circuit!
Draw arrows on your circuit diagram to represent potential differences (voltages between two points), do same for voltage sources, put arrow heads on to wires for currents , label them all, solve the equations and evaluate what the currents and voltages are. And what previous poster says is correct, if any if the currents and voltages come out negative, then if you want, you can turn the arrows around on your circuit diagram.
I tend to think about it logically first, rather than assigning the orientation of the arrows randomly, so if I know I have a positive voltage across the input between two input terminals, then I will orientate the arrow for the potential difference correctly with the arrow head pointing to the higher voltage, and current flows from a higher voltage to a lower voltage so you can usually assign some of the arrow heads for the currents correctly, though some currents you may not know which way they are flowing until you solve the equations and do the calculations.
It doesn't matter if you get the current and voltage orientations wrong before you start the analysis.