|Print a Handout: Scatter Plot (PDF)|
What is a Scatter Plot?
A scatter plot identifies a possible relationship between changes observed in two different sets of variables. It provides a visual and statistical means to test the strength of a relationship between two variables. Scatter plots can be effective in measuring the strength of relationships uncovered with a fishbone diagram (QI Toolbox: Fishbone Diagram).
How to Create a Scatter Plot
You will need at least 50-100 paired samples of data that you think might be related for a scatter plot. Enter the data into a spreadsheet, and plot the data points on a diagram (if you have created your spreadsheet in MS Excel, you can use the program to build a scatter plot with your data).
Interpreting a Scatter Plot
Example: Income and Metro Health Many levels of analysis can be applied to the diagram. You might find it helpful to consult a statistical process control guide or other texts for assistance with analysis, in order to ensure you're correctly identifying a positive or negative correlation (or absence thereof).
It's important to note that scatter plots show correlation between two variables, from which causation only may be inferred.
Example Scatter Plots
A Screening Survey to Assess Local Public Health Performance
[Public Health Reports (PDF)]
This scatter plot, from Miller, Moore, Richards, and McKaig (PDF), shows the correlation between survey responses and screening queries for an assessment of local public health performance.
Why Some Cities are Healthier than Others [The Atlantic Cities]
This scatter plot from The Atlantic Cities (2012) plots a city's "Metro Health Index" (a factor measuring the share of people who smoke or are obese) as it correlates to the city's median income. Note that the plot does not prove causation between income and health in this instance—just that the two are related.
- Scatter Diagram
American Society for Quality
- Scatter Plot
NIST/SEMATECH e-Handbook of Statistical Methods
- Public Health Memory Jogger
Public Health Foundation, GOAL/QPC