It is simultaneously possible for the results to have statistical issues while also being good enough to provide utility as evidence for drawing some conclusions. You can both be right.