Almost every statistical tests generates a P value (or several). Yet many scientists don't really understand what P values are. This section explains the principles, and also the difference between one- and two-tail P values.