Summary

Due to the random nature of processes, each time an estimate is made using a finite number of measurements (e.g., "tosses" in our "experiment"), the "true" result can be in a range of values determined by the error around the observed value.

• The result can never be determined "exactly" - only "approximately".

• The range (around the average value) that contains 95% of the results is known as the "95% confidence level".
  We also use the less scientific term "Margin of Error" to describe this range.

• One in 20 results can be expected to be outside this range!

• The range (error) can be determined by repeating the experiment many times

In practice, statistical theory can compute the error, so that we don't have to actually repeat the experiment.

Likewise, computer simulation can determine the error.

If you benefitted from this site, please consider supporting us.