A bimodal distribution is a probability distribution that has exactly two modes. Here by mode, we do not mean the value with the highest frequency but the value where the frequency attains a “peak”.

It attains a relative maximum at the peaks as opposed to an absolute maximum. A bimodal distribution might look somewhat like this:

A bimodal distribution is not normal. The normal distribution is an example of a unimodal distribution since it has only one “peak”.

A bimodal distribution may be symmetric or skewed depending on the shape of the data.

The presence of two modes generally indicates that two different kinds of data may have been improperly mixed. For example, the physical strength of men and women may well be bimodally distributed with two peaks because women have lesser strength on average.

Another example of a bimodal distribution is the number of customers at a restaurant which may peak during lunch and dinner hours.

**Multimodal Distribution**

Sometimes a distribution can have even more than two modes. Such distributions are called multimodal distributions. In such a case, the data might very well look like a sine wave.

It may also happen that the data is uniformly distributed and the graph is a straight horizontal line in which case each value may be thought of as the mode.