Front Air Diffusers - do they work?

I disagree somewhat. The front end aero of the early models is meant to be fairly poor, creating lift and giving a wandering feel at highway speeds. This is why they added the 'speed flaps' to the front wheel arches, which from most accounts make a noticeable difference even at legal speeds.
A properly made front splitter will give you a little bit more downforce (or at least eliminate lift) whilst cleaning up the front end airflow to potentially reduce drag. Combining this with a flat underbody and potentially a diffuser can also work in your favour with almost no drawbacks. Its fairly simple theory, so you don't need to be doing CFD to get something that works, and if it makes your car look pretty bitchin' then win/win :D
 
As for being strongly secured, that is something most people do overlook. Downforce only occurs if its transferred to the car. For inspiration:

According to "Competition Car Aerodynamics" - A Practical Handbook, 2nd Ed. By Simon McBeath pg 190 (referred to here as "Dive Planes" but given many different names):
 

What general conclusions can we draw about dive planes, then? We can perhaps say that by adding dive planes like those featured here:
 
-Drag increases
-Front downforce increases
-rear downforce might decrease
-the aerodynamic balance shifts to the front
-Efficiency (-L/D) might decrease if rear downforce reduces but might increase if not
-The effects will depend on the dive plane size and it's steepness
In short, dive planes can be very useful but need to be deployed with care

 
The examples "featured" in the book ranged from the Britcar Noble M400, the Eco Racing Radical SR10 LMP1, a 1991 IMSA Jaguar XJR-16, and an ALMS GT2 Porsche with modified (extended) dive planes. Full scale wind tunnels were used, and the results seemed to be inconsistent with ricer logic - the Jag (which had arguably the most "extreme" looking aero) had large dive planes that created lift at the rear, reducing overall downforce. The Noble had some of the most subdued aero of all the cars tested, but the result was the same, despite differences in Dive Plane angle and size. 
 
*HOWEVER On both cars the front gained slight downforce, shifting balance marginally towards the front - but front downforce change was in the order of a couple of percent (Noble was higher, due to lower initial downforce). Having said that, a road car may see a larger percentage increase due to the lower initial downforce of the car when stock, but at the speeds required to feel that difference you probably won't be in possession of your car for very long anyway - and whilst it's a large percentage change (I'm talking 5-6% for the Noble), it's a large percentage of a very small number :)*
 
The Radical saw an increase in downforce at the front, and no loss at the rear, seeing a shift in balance towards the front of 2%, and the Porsche, whilst having much smaller dive planes than any of the other cars, saw an increase in front and reduction in rear downforce resulting in a shift in balance by 5% to the front. 
 
 
Overall, dive planes are fickle and need to be used *very* carefully IF they are to have an effect - and the effect they have varies from car to car, and require extensive testing to actually improve handling. Their wake can interfere with the front wing, and it can interfere with the rear wing, but in all honesty,  a reduction in rear downforce isn't what you want at high speed on public roads anyway. Even on race cars we're seeing a change in downforce balance in the order of a few percentage points - something very few drivers will notice on the road (and shouldn't notice in an MR2 anywhere other than the track), particularly when airflow is so disorganised on a road car in the first place (placement of the dive planes in a turbulent low pressure zone might not do a thing to balance, nor downforce).
 
Results are entirely dependent on interactions with the rest of the aero package, and it's not as simple as "angled down = downforce", or "bigger is better". Their purpose on race cars is more about shifting downforce balance forwards, rather than simply creating more downforce - on a street car that shift in balance, if it can be achieved, won't necessarily help handling on a car prone to oversteer. To get any rear downforce in an MR2 obviously the rear wing needs to be out of the low pressure area behind the cabin, either up high or off to the sides - and the wider the wing, the greater effect dive planes will have on it. Even the dude in the Scion in the photo above^^ will probably find that his massive front wing has completely ****ed the airflow over the rear - a bit like an F1 car following closely the one in front.
 
The one thing they're guaranteed to do is marginally increase fuel consumption, due to increased drag in all instances, and probably get asked by police to remove at best, due to protruding bodywork, and get defected at worst, I would imagine. 
 
By all means give it a go - but for them to be truly effective, it's not as simple as buying something off ebay and sticking them on - even on fully fledged race cars in most cases the difference in balance appears to be marginal. If I was going to try it, Id go to a track day armed with a bunch of different designs, along with a front bumper from the wreckers so I could drill a bunch of holes in it and mount different ones in different positions, and see if I could feel the difference. It's the only way to really know their effectiveness on the MR2.
 
[edit]: Ok that turned out a bit longer than I expected - 
tl;dr version: They make a difference, but it's marginal, extremely variable, has unintended side-effects on downforce, only matters at (very) high speed, and needs fine tuning and experience to get it right - not as simple as buying off ebay.