Step 2: Identifying peaks (modality)

A peak is the highest point in a section of the distribution, where the bars are taller than their neighbors on both sides. Peaks show us where values tend to cluster.

The number of peaks in a distribution tells us about its modality. Here is what to look for:

Types of distributions by peaks

No peaks

(Uniform)

All bars are roughly the same height. No clear high point.

Single peak

(Unimodal)

One clear highest region where values cluster.

Multiple peaks

(Bimodal/Multimodal)

Two or more distinct high regions.

No peaks (uniform)

All bars roughly equal

Single peak (unimodal)

One clear high point

Multiple peaks (bimodal)

Two distinct high points

Figure 2. Three distributions showing no peaks (uniform), single peak (unimodal), and multiple peaks (bimodal).

How to identify peaks

  1. Look for the tallest bars in the distribution
  2. Check if there is one main high area or multiple separate high areas
  3. If all bars are about the same height, there are no clear peaks

Example 1

Looking at this distribution, how many peaks does it have?

  • The bar(s) in the middle are taller than bars on either side
  • There is one clear high region, not multiple separate peaks
  • Values cluster around the center

Answer: "The distribution has a single peak."

Example 2

Looking at this distribution, how many peaks does it have?

  • There are two separate high regions (one on each end)
  • The bars dip lower in the middle between the peaks
  • Values cluster in two distinct groups

Answer: "The distribution has multiple peaks."

Example 3

Looking at this distribution, how many peaks does it have?

  • All bars are roughly the same height
  • There is no clear high point or region
  • Values are spread evenly across the range

Answer: "The distribution has no peaks."

Previous Next: Describing Shape