STATS 13 Chapter Notes - Chapter 3.1-3.3: Confidence Interval, Standard Error
Document Summary
Estimation: tells how large the effect is by giving us an interval of values. Confidence interval: interval of plausible values for the size of the effect we want to know about. Confidence level: measure of reliability, typically 95% Observed statistic gives an estimate of the actual parameter value but does not take into account random variation. To determine whether values considered are plausible: We consider a value of the parameter to be plausible if the two-sided p-value for testing that parameter value is larger than the level of significance. Confidence level = 100% - significance level. If we want to be more confident that our interval contains the parameter value increase interval of plausible values. Need to make it harder to reject values of the parameter specified. As confidence level increases, the width of the confidence interval increases. Values that are not considered plausible by the two-sided test of significance should not be contained in the corresponding confidence interval.