ECO220Y1 Lecture Notes - Lecture 13: Statistical Inference, Sampling Distribution, Random Variable
27
ECO220Y1 Full Course Notes
Verified Note
27 documents
Document Summary
Claim that proportion of fake bas is 0. 10 (p=0. 10) Potentially, need more info on claim and or/data. Non-sampling error not reflected in sampling distribution graph. Starting presumption is not based on any direct evidence. The only way to know for sure if a coin is fair is to toss coin infinite number of times. Is easy to prove if coin is unfair. Null hypothesis (h0): initial presumption on unknown parameter, not based on any evidence (data) If reject h0, have sufficient evidence to infer h1. One tailed test: h1: p>0. 50 or h1: p<0. 50. Does not mean there is no evidence of guilt, simply not enough evidence. Failing to reject h0 does not mean it is true. Reject null hypothesis in favor of research hypothesis. The research hypothesis is what we hope to show is true. The point is to reject null hypothesis so we can infer research hypothesis. Test statistic: random variable created using sample statistics.
