Quality Improvement Zone

A histogram shows how frequently different values occur and can help identify patterns, variation and unusual observations.

Histograms are commonly used with continuous or numerical data, such as:

• waiting times
• ages
• costs
• workload measures
• satisfaction scores
• clinical measurements

The horizontal (X) axis shows the values being measured, grouped into ranges known as bins. The vertical (Y) axis shows how many observations fall within each range.

Histograms help teams understand variation within a dataset and explore the overall shape of the data. They can reveal patterns that may not be obvious when looking at individual data points or summary statistics such as averages.

Using a histogram can help you:

• understand the distribution of a measure
• identify the typical range of values
• understand the spread and variation within the data
• identify unusual observations or outliers
• detect clusters or multiple groups within the data
• compare distributions across different teams, services or populations
• generate questions for further investigation

Histograms help teams understand what variation looks like within a dataset, but they do not explain why the variation exists. Further investigation may be needed to understand the factors contributing to the patterns observed.

Histograms are most useful when you have a reasonable amount of numerical data to analyse. As a guide, histograms tend to become more meaningful when there are at least 30 observations.

1. Define the measure

Identify the numerical measure you want to explore, such as waiting time, age, cost or satisfaction score.

2. Gather the data

Collect the data and ensure it is accurate and complete. The more observations available, the easier it becomes to understand the distribution.

3. Create the histogram

Histograms can be created easily in modern versions of Excel using the built-in histogram chart.

1. Select your data.
2. Choose ‘Insert’.
3. Select ‘Statistic Chart’.
4. Choose ‘Histogram’.

Excel will automatically create the histogram and select an initial set of bins. These can be adjusted if required.

4. Review the distribution

Examine the overall shape of the histogram and consider what it tells you about the data. Look for:

• where most observations occur
• how much variation exists
• whether values are concentrated or widely spread
• whether there are unusual observations

5. Look for common patterns

Common patterns include:

• symmetrical distribution: The data is spread evenly around the centre and may resemble a bell shape.
• skewed distribution: The data has a longer tail on one side than the other, suggesting values are not evenly distributed.
• multiple peaks: Two or more peaks may suggest different groups or sub-populations within the data.
• gaps: Gaps between groups of values may indicate different processes, populations or data collection issues.
• outliers: Individual values that sit far away from the main distribution may warrant further investigation.

6. Compare groups where appropriate

It can be helpful to compare histograms from different teams, services, locations or populations. Differences in shape, spread or clustering may highlight variation that requires further investigation.

7. Use findings to inform improvement

Use the insights from the histogram to develop a better understanding of the system and identify areas for further exploration. The findings may highlight:

• unusual variation
• data quality issues
• opportunities for deeper analysis
• areas that may benefit from additional investigation

Interpreting histograms

A histogram provides a snapshot of the distribution of data at a point in time. It can help you understand:

• the spread of values
• the shape of the distribution
• the presence of unusual observations

Important note

Histograms show the distribution of data but do not show changes over time. If understanding performance over time is important, a run chart or SPC chart may be more appropriate.