QTM 100 Lecture Notes - Lecture 17: Test Statistic, Confidence Interval, Null Hypothesis

• School: Emory University
• Department: Quantitative Theory & Methods
• Course: QTM 100
• Professor: Jacobson Jeremy

Document Summary

Relationship of confidence intervals and hypothesis tests (41/46) A hypothesis test allows you to quantify the evidence about a null hypothesis with a p-value. Doesn"t matter which test you do first, they will always agree with each other. Results of hypothesis tests and confidence intervals usually agree. Possible that a rounding error in the formula causes a difference. Hypothesis test may say fail to reject because it"s not in the ci. Not going to get exact decimal places in agreement within ci and hypothesis tests when working with a proportion. A confidence interval allows you to estimate the true but unknown parameter value value. A confidence interval is much more informative than a p-value. Hypothesis testing results should always be accompanied by a confidence interval for the parameter of interest. Sometimes more complex questions can be answered by just one p-value, whereas you could not answer the question with one ci.