Power correction factors for glass-lined vessels and different baffle arrangements are very limited. The typical power data for glass-lined vessels has just one baffle and  shows gradual decreases in power number as the Reynolds number increases in the  turbulent range.  That reduction in power is associated with swirling of the vessel contents.

For the effect of two baffles, the attached graph shows different numbers of baffles and different fractions of  baffle width, with the standard baffles being four (4) baffles with a width  of 1/12 the tank diameter.  Your data would suggest that you have the effect of two (2) baffles 3/4 of the standard width, giving you about 70%  of full power.
 
The effect of baffles in most glass-line reactors is limited to upper tank effects, including surface motion.  The baffles have almost no effect below the ends of the baffles, where the impeller is located typically.  A paper from a few years ago looked at flow patterns in glass-lined reactors:

"Understand Flow Patterns in Glass-Lined Reactors," Chemical Engineering Progress, pp. 21-25, (November 2004), Dickey, D. S., Kevin J. Bittorf, Christopher J. Ramsey, and Keith E. Johnson.

I hope this information helps answer your question.  The measurements obtained for your system seem quite reasonable and are probably better than predicted values would be.

Baffle Effect graph: http://www.chemicalprocessing.com/Media/MediaManager/Baffle_Effect.pdf.