At this point, we don't know any different. The change in mean Doppler shift of some stars with transiting planets as the planet covers pieces on different sides tells us that their orbits are more or less lined up with the star's equators (at least as opposed to being polar or retrograde). I've been told that the data exist to do slightly more in a statistical sense from the Doppler linewidths of stars with planets, but the analysis is tedious enough that the planet-hunting teams haven't been motivated to grind through it at this point (having plenty to do with new planets and all).