When would you want a 95% confidence interval and when would you be interested in a 99% confidence level ?

Results from surveys or opinion polls often report a range of values—the sample statistic plus or minus a margin of error . This tells us that the range is likely to contain the population parameter. How much wiggle room we provide is based on how much confidence we wish to have that the range contains the actual population mean. That confidence level is directly related to the middle “truth” area we will accept versus the dubious tail area we will reject–also known as alpha .

The more confidence we wish to have—the more middle ground we will need to accept thus a smaller tail area. If we insist on a larger alpha we narrow the middle ground we will accept and thus provide less wiggle room—so the more likely it is that we will miss the true average . A 95% confidence level leaves 5% alpha. A 99% confidence level leaves 1% alpha.

Now, without calculating a mean or margin of error or a confidence level, provide an example from your current professional or personal life that describes a measurement that is normal—and how much wiggle room on either side would be appropriate. When would you want a 95% confidence interval and when would you be interested in a 99% confidence level ?