# To the editors

We hereby submit for your consideration a brief manuscript regarding a statistical error that is particularly common in experimental research into the neurobiology of language: testing that stimulus sets or participant groups are equivalent on nuisance parameters (e.g., conducting a $t$ test on the difference in word length between an animate and an inanimate condition; conducting a $t$ test to establish the equivalence of aphasic patients and matched controls regarding age). In sum, these tests refer to *any* difference in a *population*, whereas the experimenter's intention supposedly refers to a *meaningful* difference in the *sample*. We elaborate why this method is worse than useless, discuss alternatives, and show how common it is in issues of *Brain & Language*.

Hoping to contribute to a minor, but essential correction of our statistical repertoire, we thank the editors for hopefully giving us the opportunity to discuss this issue in this venue. We would be delighted to hear from you,

Jona Sassenhagen