F-ratio

a ratio of two independent estimates of experimental error. If there isn't an effect, the ratio should be close to 1.