QTM 100 Lecture 19: Inference for Two Proportions
7 Dec 2017
School
Department
Course
Professor
Document Summary
Took samples from different regions to try and understand its transmission. Sample may not be representative of the whole disease. Based on sample, need to decipher if a region is actually worse than another. Take sample proportions in regions and assess if the data is due to chance of random sample or do to a cause. Categorical variable 2: 2 presents vs. absence. Estimate the difference in proportions with a 95% ci for p1 p2. If this ci includes 0, no effect/no association. Otherwise, there is an association (and we can determine the direction of the association) the association) Question: we want to do a 95% ci for p1 p2. Data are obtained by randomization (such as a random sample or a randomized experiment) Made best effort, but constraints made it impossible to achieve perfect randomization. The two groups or populations are independent, whereas the observations are independent. Yes, ashanti and greater accra are independent regions.
