COMM 162 Lecture Notes - Lecture 15: F-Test, F-Distribution, Test Statistic

• School: Queen's University
• Department: Commerce
• Course: COMM 162
• Professor: Tatsiana Levina

Document Summary

Lecture 15: testing with two populations part 2. Independent samples: two samples are completely independent from one another, can have the same or different sample size. Paired samples: cannot be separated or sorted in the same way, as the samples are connected to one another, must have the same sample size for both. Testing hypotheses for two samples: still have h0 and h1, h1 is >, < or =/, comparing u1 and u2, if sample 1 and sample 2 are both normally distributed, then their difference is also normally distributed. If o1 and o2 are given, then you can calculate z-value: Z = [(x1-x2) - (u1-u2)] / sqrt[(o1^2/n1) + (o2^2/n2)] P-value = 1- norm. dist (x1-x2, u1-u2, sqrt[(o1^2/n1) + (o2^2/n2)], 1) If o1 and o2 are not given, then you will use t-distribution: use f test: to determine if o1=o2 or o1=/=o2.