STAT 201B Lecture Notes - Lecture 17: Type I And Type Ii Errors, Emo, Jerzy Neyman

• School: Berkeley
• Department: Statistics
• Course: STAT 201B
• Professor: Elizabeth Purdom

Document Summary

7l type 1 error ) = flo cx er ) 7ctype ii error ) = th , cx hr) Tito ( x e r) tu=o cxer ) 7th ( x er) = ? different for each problem. Look at error as fns of o. Po (x er) tf o c- . Type i error b co) for a g ho. Type ii error: size test x = snf to cxer) = sup bc -07, c- ho. Tex) > c size test ed level x c ca ) #o power function ideal ho. Reject tcm > c where ci too 5174=2 then g. c) is most powerful among all test of size a ( max bcom ) L = tmrw ( t cx) > c ) No cin > w9""nt#o; ) under ho. { text > c)(cid:12200) { in > not 3 ,- a ohh} them. Idea how we set up and choose between test.