If you're running a second-order on the woofer, the output from the crossover measured across the woofer voice coil probably has a significant peak somewhere near the crossover point. Look at pages 14 and 15 of this crossover lab document to see what I mean.

If the driver and horn are falling off about the same place, then crossover peaking may help boost the top end, I suppose. That would make summing through the crossover difficult to predict, so the only way you'd know is to check it and see. This is all very implementation specific, and that was the point I wanted to make.