W-subtype W UMa variables
I have 2 years of photometric data (B, V and Ic) for a variable star known to be a W-subtype W UMa variable. The spectral class of this system has been variously assigned in the literature between G4V and G9V. For the purposes of Roche modeling, the effective temperature (Teff = 5520 K) was estimated from the mean color index determined during quadrature (0.25P and 0.75P) in 2008 and 2011 along with supporting data from other several other surveys. My initial inclination was to assign this value to the hotter component, which for a W-subtype system is the least massive star (qsp = 0.38) yet this appears to be the exact opposite of that reported in the literature this binary for many other similar W UMa systems. A back of the envelope calculation using the Stefan-Boltzmann law (L=4πR2σT4) would argue in this case size wins (1/0.38)0.8 given the relatively small difference in temperature between W UMa stars (<500 K). Is this reasoning correct or am I missing some other key piece of information which is needed to properly assign a Teff based on color index to the appropriate star in a W-subtype system?