bimodal distribution

n. A statistical distribution of data in which there are two distinct peaks or modes (as contrasted to the unimodal, or single-peak, distribution depicted by a bell curve). A bimodal distribution may indicate that there are actually two different samples or distributions being summed together. For example, many gender studies produce a bimodal distribution. When information (i.e., attitudes, opinions, traits, values) from males and females is averaged together, it generally produces a normal distribution, but when the data from males and females is considered separately, there are often two different means with their own standard deviations. These different means produce the two peaks, or modes, characteristic of a bimodal distribution. Bimodal distributions may be used to demonstrate how deceptive simple descriptive or summary statistics can be, or may be an impetus to further investigation and improved exploratory modeling of a target phenomenon.

- BJM