Significance testing is still wrong, and damages real lives: a brief reply to Spreckelsen and Van Der Horst, and Nicholson and McCusker