ECO220Y1 Lecture Notes - Lecture 12: Confidence Interval, Non-Sampling Error, Type I And Type Ii Errors
27
ECO220Y1 Full Course Notes
Verified Note
27 documents
Document Summary
Estimator: random variable based on sample statistics that is used to estimate parameter. Unbiased estimator: expected value equals population parameter it estimates. We can use to make inferences about p. We know 95% of will be within 2 standard errors of p. We are 95% certain that p is within 2 standard errors of any. Margin of error: extent of interval on either side of. The more confidence we want, the larger the required margin of error. Every confidence interval balances between certainty and precision. Consistency: as sample size approaches infinity, sampling error decreases and converges to population parameter. Confidence interval is point estimate +/- margin of error. Margin of error = measure related to desired confidence level * measure of sampling error. Start with desired confidence interval when designing sample size for survey. Critical value: number of standard errors we must stretch out on either side of. Success/failure condition for and: if target me( with (1 - ) confidence.
