It seems to me that this would not be real easy to test. You'd need to build up a number of pairs of identical rims and then run them with measured lateral loads or overloaded vertically until they failed. Just doing one or two rims wouldn't tell you much. Comparing Wheel A built up by Bob to Wheel B built up and tied and soldered by Jim wouldn't tell you anything at all (about the tying, that is, might still show which wheel was superior, but wouldn't show why).

I wonder, too, if maybe improvements in wheelbuilding, spoke strength, and available rims haven't made this less necessary. I know over in the Clydesdale forum, you'll hear different recommendations for wheels, but tying and soldering is never part of the recommendation.