The 25% rule of thumb is just that. So far, no scientific evidence of it, just some personal experience. I also noticed that if you do the standard navigation planning exercise of flying between two points 100 nautical miles apart at 100 KIAS and start adding a wind that is directly along the track (to show how things don't always cancel out), that the effect really starts to show when you get to 25 knots of wind. Got me thinking and observing more closely and it appears to be a reasonable rule of thumb.
The folks who really know about this effect are ultra-light pilots - ask one of them what height he would climb to if taking off with a left cross wind of lets say 10 knots crosswind component and wants to do a right turn after takeoff. It's quite high, because when the right turn starts the aircraft will probably descend (their airspeeds are pretty low).
The folks who absolutely know and understand this are the model airplane folks. Their airspeeds are very low and even the slightest wind affects them when they turn.
Who knows, might be a good master's thesis subject...