Ever wonder why a survey of the American population often has what seems like a very small sample size? 2000 people can tell us what the American population is thinking about. We have formulas to calculate an appropriate sample size for both estimation of the mean and estimation of the proportion. How do they work and why do we need so few to tell us what so many might be thinking?
Confidence intervals is a range of values of which the true value is in which comes from sample statistics and contain atleast 95% true and 5% will not. It gives us more information then estimates. Three pieces of data to compute the confidence intervals which are for continuous data (mean) or binary data (proportion), and the sample size. An example that can be used is obtaining surveys from customers on customer service and have them rate it from 1 to 10 and 10 being the best. Any thoughts?