Choose the confidence level (e.g., 90%, 95%, 99%).
Find the critical value (z^*) (or (t^*)) for that confidence level.
Determine whether the population standard deviation is known:
If known: use (z^*)
If unknown: use (t^*) with (df = n-1)
Compute the standard error:
If using a known (sigma): (text{SE} = sigma/sqrt{n})
If using a sample standard deviation (s): (text{SE} = s/sqrt{n})
Calculate margin of error:
If using (z^*): (text{ME} = z^* cdot text{SE})
If using (t^*): (text{ME} = t^* cdot text{SE})
For a proportion:
(text{SE} = sqrt{dfrac{p(1-p)}{n}}) (use (p) or the sample (hat{p}))
(text{ME} = z^* cdot text{SE})
For an interval:
Use (text{Estimate} pm text{ME})