Great question. Unrelatedly, mind if I correlate your medical history and a genetic profile of you with your purchase history at three major retail chains and the contents of your inbox?
Bad point. Violating the privacy of individuals is worlds different from studying gender differences on average.
(... everything has a political dimension, including data accumulation. Because ultimately, humans use the data, and what humans do is politics).
But from that, it can't follow that we should alter the results to avoid hurt feelings. That would render the "data" no better than mythical just-so stories. Politics should be subservient to empirical data and science. We've seen what doing it the other way results in historically.