A scatter diagram consists of pairs of numerical data containing one variable on each axis. The diagram is used to find a relationship between the data pairs. Points that create a line or curve indicate correlated variables.
To show what happens to a variable when another is changed.
When analyzing whether two variables are related.
When working with paired, quantitative data.
1. DEFINE THE X VARIABLE, typically the "cause variable," on a graph paper scatter diagram form.
2. DEFINE THE Y VARIABLE, typically the "effect variable," on the diagram.
3. NUMBER THE PAIRS OF X AND Y VARIABLE MEASUREMENTS CONSECUTIVELY. Put each data pair for x and y in their respective columns. Ensure the data remains paired.
4. PLOT THE X AND Y DATA PAIRS on the diagram. Place a point on the graph where the two values intersect.
5. STUDY THE SHAPE that is created by the series of data points plotted. If there is a data association the shape should resemble an elliptical shape or a straight line.
a. There is little or no correlation if the points form a circular shape.
b. The correlation is positive if the points for both values are increasing.
c. The correlation is negative if one variable increases and another decreases.
Spreadsheet software makes this method easy.
Reveals patterns in data.
Is an easy way of determining whether a cause-and-effect relationship between two variables exists.
Provides the first step towards analyzing the relationship between two variables.
If the data cover too small of a range the results of the analysis can be skewed.
Patterns can appear or vanish based on how the data are stratified.
If there is a pattern, it does not always indicate that there is a relationship.
University Research Co. LLC. Health care improvement project: scatter diagram. 2008 [cited 2009 July 28];
George M, Rowlands D, Price M, et al. Identifying and verifying causes. The lean six sigma pocket toolbook. New York: McGraw - Hill; 2005. p. 141-96.
American Society for Quality. Cause analysis tools: scatter diagram. 2009 [cited 2009 July 23]; Available from: http://www.asq.org/learn-about-quality/cause-analysis-tools/overview/scatter.html