Understanding Confidence Intervals and Hypothesis Tests in Descriptive Statistics - Prof. , Study notes of Statistics

Download Understanding Confidence Intervals and Hypothesis Tests in Descriptive Statistics - Prof. and more Study notes Statistics in PDF only on Docsity! 4/7/2010 1 1STAT 110 – Introduction to Descriptive Statistics Chapter 23 Use and Abuse of Statistical Inference 2STAT 110 – Introduction to Descriptive Statistics Assumptions The two forms of inference we’ve learned are confidence intervals and hypothesis tests. • For either of these to be valid we need for the data to come from an unbiased random sample • These methods on inference for p require a large enough sample size to approximate the distribution of with a normal distribution 3STAT 110 – Introduction to Descriptive Statistics Confidence Intervals The confidence level tells us how often the interval catches the true parameter in the long run if we were to take many samples from a population and compute a confidence interval from the resulting data for each one Once you take your one sample and compute a single confidence interval, it either was one of the correct intervals or one of the incorrect intervals….so it’s either right or wrong (and so it has probability of 0 or 1 of containing the true parameter at that point) This is why we say “With 95% confidence” instead of “With 95% probability” 4/7/2010 2 4STAT 110 – Introduction to Descriptive Statistics Confidence Intervals For 95% confidence, which of the following is a true statement? a. 5% of the intervals in the long run contain the true parameter b. 95% of the intervals in the long run contain the true parameter c. We don not know what percent of the future intervals will contain the parameter 5STAT 110 – Introduction to Descriptive Statistics Confidence Intervals True or False: It is OK to start your confidence interval interpretation with “With 95% probability” a. TRUE b. FALSE 6STAT 110 – Introduction to Descriptive Statistics Confidence Intervals If computed on the same sample, a 99% confidence interval will be a. Wider that an 95% interval b. Narrower than a 95% interval c. The same width as the 95% interval d. Don’t know until we compute the intervals