Geography 2210A/B Lecture Notes - Lecture 11: Standard Deviation, Test Statistic, Type I And Type Ii Errors
14 Dec 2017
School
Department
Course
Professor
Document Summary
Selection of sample statistic/distribution: statistic we are going to test > test statistic. Based on question: distribution depends on population sampled, assumed to be normal for tests in this course. Important: in hypothesis sampling distribution is based on null hypothesis being true: does a neighbourhood exhibit the city wide mobility? n (proportion) = 0. 2, normal (large sample size) Selection of a level of significance: how to select a, depends on nature of problem and consequence of being wrong (i. e. , making a. Type i error: medical sciences usually use very small a (e. g. , 0. 0001) Recall: ho: n = 0. 2, ha: n s 0. 2: alternative is to calculate probability. Value (p: p-value indicates degree of belief that ho should be rejected. Equivalent to type i error rate: obtain from test-statistic table, avoids clear cut rejection/acceptance of null even if just outside critical region, note: does not preclude setting of a priori a.
