That would make a great Hallmark card for statisticians: a twee little teddy bear with caption "I <3 YOU (p < 0.05)"
And when things do turn out, you can honestly call that person...your significant other. ;-)
...and there's no such thing as generating enough power to limit type 1 error. :-p
And Muffin, your method of correlating responses of mutual acquaintances has another name: the seventh-grade technique.