OPER1135 Lecture Notes - Lecture 16: Type I And Type Ii Errors, Statistical Hypothesis Testing
Document Summary
If you have a two-tailed f-test can use the function f. test(array 1, array2). The output is the p- value. (you don"t have to double it. : or , p-value method. Two-tail: =2*f. dist. rt(f-stat, numerator df, denominator df: critical value method. For the f-distribution there will only be one critical value, the upper critical value, even for a two-tailed test. Ways you can be wrong in hypothesis testing: in hypothesis testing you either reject h0 or fail to reject it, so you can be wrong if you. Fail to reject h0 when you should (type ii: i claim that for every hypothesis test we"ve done you can tell me the probability of rejecting h0 when you shouldn"t. It"s , which you set for every test you do: the probability of failing to reject h0 when you should is denoted , let"s look at the relationship between and . For a given n to decrease we increase .