"the results are meaningless"
Meaningless??? That is a bit harsh, don't you think? How about, "not entirely reflective of reality"? Besides, wouldn't the meaninglessness of the results depend upon how many idiots are choosing impossible pairs, and statistical significance? (Also, even if the software rejected such impossible pairs, it still wouldn't detect lying, bias, very high expectations, very low expectations, etc.). It is what it is. Not meaningless, however.