Use of Standard Deviation: Chebyshev’s Theorem
Once we know the mean and standard deviation of a data set,
we can estimate what proportion of all observations lies in an interval around the mean.
Chebyshev’s theorem does this for any distribution shape
(not just bell-shaped curves).
Chebyshev’s theorem
Let \(\mu\) be the mean and \(\sigma\) the standard deviation of a data set.
For any number \(k > 1\),
In words: at least \(1 - \dfrac{1}{k^{2}}\) of the data values must lie
within \(k\) standard deviations of the mean, in the interval
\([\mu - k\sigma,\; \mu + k\sigma]\).
This percentage is a minimum; the real proportion can be larger.
Important special cases
-
For \(k = 2\): at least
\[
1 - \frac{1}{2^{2}} = 1 - \frac14 = 0.75
\]
so at least 75% of the data lies within \(2\sigma\) of the mean.
-
For \(k = 3\): at least
\[
1 - \frac{1}{3^{2}} = 1 - \frac19 \approx 0.89
\]
so at least 89% of the data lies within \(3\sigma\) of the mean.
Using this calculator
-
Enter the mean \(\mu\) and standard deviation \(\sigma\) of your data
(for a sample, the calculator uses the sample mean and sample standard deviation).
-
Specify either:
- a symmetric distance \(k\) (number of standard deviations) from the mean, or
- an interval \([a,b]\) around the mean.
-
The tool computes
\[
k = \frac{\max(|a-\mu|,\;|b-\mu|)}{\sigma}
\]
when needed, then applies Chebyshev’s theorem to give the
minimum proportion and percentage of observations
that must lie in the chosen interval.
Chebyshev’s theorem is conservative: it works for any distribution shape,
but the actual percentage of observations within the interval is often larger than
the bound it provides.