Just to reinforce what Bruce has said, I built a series of F1A gliders that featured Whitcomb winglets. Nicknamed by the locals as 747-400's because of their size (they were high aspect ratio in the late 1970's, standard now)

You could see these winglets working from the flightpath, a very flat fugoid - from the top (slowest - no stall) the model nosed slowly over, accelerating, then the winglets would cut in and the model would climb, again very slowly, with the nose coming up as the speed reduced. A full cycle would take perhaps 3 or 4 seconds, and would have the model moving through an altitude change of perhap 12 inches. The speed differential was not noticeable but has been deduced as the explanation of the flight path.

How did we know that they worked? By knocking them off one side to see what happened to the glide. In every case the model turned torward the side with no winglet. And I mean turned - the standard circle for these models was about 50m diameter, without one winglet, they were turning in about 20m.

Did they make any difference to the performance of the models? In my head - a big yes. My heart does still have reservations.