Look Elsewhere Effect

The problem received formal statistical treatment in the 1950s, as researchers including John Tukey and Henry Scheffé developed correction procedures to control the inflated false-positive rates that arise when many hypotheses are tested simultaneously. The canonical remedy — the Bonferroni correction — is named for Italian mathematician Carlo Emilio Bonferroni (1892–1960), though it was first applied in this statistical context by Olive Jean Dunn in 1959. The underlying logic is ancient — more searches mean more chances of finding something by luck — but the formal mathematical treatment arrived alongside the computing capacity to run large numbers of simultaneous tests.