Overview

A 90% confidence interval gives a range of plausible values for a population's mean, based on a sample. It's narrower than a 95% interval, but it's less likely to contain the true value.

Enter the sample's mean, standard deviation and size. Keep the confidence code at 1 for 90%, or use 2 for 95% or 3 for 99%.

Worked examples

InputsResult
Sample Mean54.2Sample Standard Deviation12.4Sample Size (n)120Confidence level90%90% CI: [52.34, 56.06]Margin of error: 1.86
Show 2 more examples
Sample Mean100Sample Standard Deviation15Sample Size (n)50Confidence level90%90% CI: [96.51, 103.49]Margin of error: 3.49
Sample Mean54.2Sample Standard Deviation12.4Sample Size (n)120Confidence level95%95% CI: [51.98, 56.42]Margin of error: 2.22

How it works

The standard error is the standard deviation divided by the square root of the sample size. The margin of error is that value times 1.645 for 90%. The interval runs from the mean minus the margin to the mean plus the margin. It uses z values, so for samples under about 30, a t-based interval would be slightly wider.

The formula

CI = x̄ ± 1.645 × s ÷ √n
x̄Sample mean
sSample standard deviation
nSample size

Practical tips

  • With a mean of 54.2, s = 12.4 and n = 120, the 90% interval is 52.34 to 56.06.
  • A 90% interval is about 16% narrower than a 95% one from the same data.
  • If two intervals overlap, that doesn't always mean there's no significant difference, so run a proper test.

Frequently asked questions

What does a 90% confidence interval mean?

If you repeated the sampling many times, about 90% of the intervals would contain the true mean. It doesn't quite mean there's a 90% chance for this particular interval.

When would I use 90% rather than 95%?

In early-stage research or quick business decisions, when you can accept more risk in return for a narrower range.

Is it the same as a one-sided 95% test?

They're connected. A 90% two-sided interval corresponds to one-sided tests at the 5% level.