With the Engine Control Module's modified to take full advantage of ethanol's higher octane mileage is improved significantly.
Fuel Freedom Foundation did a study with several FFVs and one non-FFV using various blends of Ethanol and Methanol. They modified the ECU's to take advantage of Ethanol's and Methanol's higher octane than gasoline.
http://www.fuelfreedom.org/wp-content/uploads/white-paper_GGE-Sept-2013.pdf
They tested the cars with various blend levels from E0 to E85 (similar mix of blends of Methanol). They found the cars on Ethanol E85 averaged 17% better fuel efficiency than the BTU content estimate of mpg predicted (and BTU estimate does give a pretty good rough estimate of what FFVs with detuned engines - for low octane gasoline - achieve in terms of mpg). Using Methanol the improvement was 20%.
This was without doing anything to boost compression (as in turbo-charging) - just using optimal ignition timing. By using turbo-charging to get higher combustion chamber pressures (ethanol's higher octane will allow greater pressures than gasoline) you can produce more power per unit displacement and downsize the engine and achieve greater improvements in fuel efficiency.
Any difference in mileage for E10 is barely discernable (maybe 2%). This has been tested multiple times by Government labs.
THere really is no reason to get the poor mileage the FFVs get except to get along with (keep political contributions coming) the Oil industry.
Anybody involved in auto racing knows that alcohol is a better fuel than gasoline. burns cleaner and has higher octane and greater latent heat (another advantage when super-charging or turbo-charging).
Getting back to OP, the tests this group did were pretty straight-forward. Don't see any reason to doubt the results. Of course, engines do have to have seals and hoses made of material that can hold up to alcohol. Without those changes you can have problems. Autos since around 2001 are capable of handling E15 with no trouble.