When you look at a frequency histogram, you will see it has a certain shape. This fact is so useful that statisticians talk about the shape of the distribution of data in terms of the shape the histogram would make. On this page, we will look at some of the most common shapes and learn a little about their qualities.
Here are 6 different histograms, each with a different shape.
Flat (or Even)
A flat histogram indicates that every value appears in the data the same number of times. If you rolled a die often enough, you would get a flat histogram.
A perfectly flat histogram is rare in practice unless you have very a large sample. A nearly flat histogram suggests that the population distribution is flat, but that the sample is not large enough to reflect that fact.
A normally shaped histogram indicates that the sample data has a normal distribution. There is a topic covering normal distributions in this tutorial.
A bimodal histogram has two peaks - showing two modes. This might suggest that there are two distinct populations, each with a different mode, represented in your one sample.
A histogram with skew to the left (also called negative skew) indicates that the majority of the data has values towards the upper end of its range.
A histogram with skew to the right (also called positive skew) indicates that the majority of the data has values towards the lower end of its range.