QUERI – Quality Enhancement Research Initiative

Go to the HSRD website

Quality Improvement Methods

41. Scatter Diagram

a. Definition: A two-axis graph is used to display values with the intent to quickly and visually determine if a relationship exists. Pairs of measures, such as price and quality or age and weight, can be plotted against one another to identify the direction and strength of the relationship between them. Statistical tools exist to quantify such correlations but often the relationships are evident from just looking at such a diagram. The Scatter Diagram is useful during a group discussions or brainstorming, along with other methods such as the fishbone diagram, to determine causes of problems and opportunities for improvement.

b. Literature: Many books and articles discuss scatter diagrams along with other quantitative tools for quality improvement.

  • Quality Management for Organizational Excellence: Introduction to Total Quality (6th Edition) by David L. Goetsch and Stanley Davis (2009)
  • Bamford, David R., and Richard W. Greatbanks. "The use of quality management tools and techniques: a study of application in everyday situations." International Journal of Quality & Reliability Management 22.4 (2005): 376-392.
  • Reynard, Sue, ed. Scatter Plots: Plain & Simple. Oriel Incorporated, 1995.

c. Example:

scatter plots

d. Steps:

1) Determine two measures in which you are interested. For example, incidents of a quality problem and another type of incident, such as use of a certain procedure. For this method, the data should be a range of values such as a count of the number of events or positive integer quantities. It is not as useful for continuous or binary data.

2) Gather data on occasions where the two measures are recorded.

3) Plot as an X-Y graph of data points as shown above. MS Excel will do the scatter plots.

4) Merely observe the relationship, if any, and consider what strength and direction for the relationship the data implies. In some cases there may be little or no relationship between the variables but that can also be useful information.

5) Optionally, calculate the statistical correlation to quantify the direction and certainty of the relationship.