Whenever you measure two different groups, you find a difference. Can it be solely ascribed to the one variable? Doubtful. You would have to match the groups for all possible contributing factors, and here not even the basic demographics have been matched? Another statistical effect for the irreproducibility bin.