"The established literature fails to predict X, Y, Z and doesn't account for N. My new model takes this into account and provides better predictive power than the old one, please peer review my work" is a far cry from "I only believe in two genders" or whatever.