If this were presented in a series of selection tests and there were no pattern to which was chosen, it would suggest free will is operating. If there were anything driving selection that was beyond the scope of will, then it would suggest that choice was determined by other factors than free will.
I think one could make a lot of interesting experiments with this.
If we take a series of babies which have never tastes M&Ms, and you place two red ones in front of them and let them choose. I would expect that the would choose rather random, maybe those that are right handed would prefer the one to the right slightly more than the left one, simply because of this. So to remove the chance of that, you could switch the M&Ms around from their initial position, even though they are identical but to see if that would change how the babies chooses. Obviously while the they ain't looking.
Now next experiment, you have a lot of babies that have already made that decision, so now you could replace one of the red M&Ms with a yellow and see whether they stick to what they know, meaning the red or if this doesn't matter and you could test this with them being in varies positions as well.
You could even place them slightly apart so for instance it would be slightly easier for the baby to choose the yellow one over the red, an see if there is some preferences to go for the red now that they have tastes it and know it tastes good.
So in the first test if they choose completely random, but doesn't in the next one where they have already gotten experience with the red one, whether that seem to affect how they choose.
Obviously you would and should expand on these experiments to set up some well thought out senaries as these are just from the top of my head. But think it could be quite interesting to see how they would behave.