Thread exporting tracks from Reason - dither or not

I have just started to export tracks from reason(I just started using Reason a few months ago, its great but the sequencer is pants) to put into cubase as audio for seqeuncing and final mixdown and to take advantage of a couple I vsts i want to apply to various elements. I cant rewire anymore with this project, cpu cant cope.
In Reason I am working at 48khz samplerate at 24bits. I'm exporting indivisual parts, and I notice a common loss in quality of sound when reviewing the audio. It cannot be anything to do with my soundcard. Its a M-Audio Vgood quality firewire thingy.
At first I was dithering at export and use the same settings when imported in cubase(48k/24b).
I have experimented without dither(48k/24b), at 90khz 24b, and at various other settings with and without dither but find it confusing now when comparing the results (more like a strain on my brain and tears in my ears).

I have tried playing reason, recording in cubase. Applications fight for sound card. Tried recording into wavelab, but have pain in the ass timing trouble when bringing into cubase.

I have tried researching this, but have not got a degree in communications or computer science or physics.
I know dithering is a process to retain as much quality as possible when putting something to cd or reducing the the bits per sample rate. However I am not reducing either until the very very end of the project.
I need a very simple dumbed down explanation of what to do or a good method, I'm sure someone has a method for this activity.

Should I dither or not in this circumstance?

Whats the best mehod of exporting/importing to retain the best quality samples?