It's tricky to try to guess which configuration is better or worse.
But there are a few things you can try to figure this out experimentally:
1) Tuft test the rear glass. Easy to do! You're looking to see whether flow off the roof remains attached all the way to the base of the glass, or whether it separates at some point, causing reverse flow/vortices (tufts pointing the wrong direction, or lifting off the glass).
EG: This video shows the air flow pattern (attached flow) on the rear glass of a 1st gen. Honda Insight
FROM: http://ecomodder.com/forum/showthrea...ght-19299.html
If you've got attached flow (all tufts streaming back) all the way to the base of your hatch glass, as the Insight does, then a big-ass spoiler isn't going to make things any better - it will more likely hurt.
2) Another approach is to do high speed A-B-A coastdown testing (if you can find a road where you can do the testing): Do a set of runs with the spoiler in place (A); Remove the spoiler & do another set of runs (B); Re-install the spoiler and do one last set. The higher the initial speed, the more likely you can tease out "small" aero changes.