Does it make sense to use variance to describe non normal distributions? I would like to know if it makes any sense to use the variance or standard deviation to describe non gaussian distributions
 A: Yes, the variance can still be a useful descriptor for non-Gaussian distributions, but it will also depend on your goal. For example, a Gaussian distribution is uniquely characterized by its mean and variance (i.e. if we know the mean and variance, then we know everything about the distribution), and this is true for some others, but certainly not all distributions.
A: You can compute many statistics for many distributions. They have a used.
Some of the most common ones include:

*

*Quantiles

*moments (including mean and variance)

*Extreme values

If I rearrange this list it is still useful:

*

*measures of central tendency.

*measures of tendency of variation or spread

*Measures of other geometric tendencies.

*Ordered measures vs moments

You can see how under the heading of the measures of Tendency of variation, also called measures of tendency of spread, the variance could be quite useful. In fact, being a moment, it is not necessarily derived or contingent on a distribution being normally distributed.
