ORIGINAL: Aeroengineer1
I currently do not have the link, but researchers at Arizona State University showed that winglets on a small wing did improve performance over the baseline, but also showed that performance degradation with a non-optimized design.
My final year project was about this subject, and I used a wing that was approxmatly the scale of an R/C model wing, I managed to design a winglet that decreased drag by approximatly 8% at positive AoA vs baseline, and about 6% -AoA vs baseline, it wasnt an endplate and was half the chord length, and blended into the front leading edge.
Considering the results I saw from the wind tunnel tests I will say it could be applied quite successfully to R/C, one thing thats important though is blending into the aerofoil, if you dont then you MAY loose drag from the winglet effect but get it all back plus some from the increased interference drag.