A scatter chart shows the correlation between two variables in a process. These variables can be a Critical To Quality (CTQ) trait and a factor influencing it, two factors influencing a CTQ, or two related quality traits. Points representing data points are scattered on the diagram. The degree to which the points clump together in a line across the chart shows the strength to which the two factors are related.
Sometimes the scatter plot shows little correlation when all data is considered at once. Stratifying the data, that is, dividing it into two or more groups based on a difference, such as equipment used, time of day, some variation in materials, or differences in the individuals involved, can yield surprising results.
The reverse can also be true. Significant correlation may appear to exist when all data are considered at once, but stratification could show no correlation. One way to look for stratification effects is to make your points in different colors or to use different symbols if you suspect that there may be differences in stratified data.
Why use scatter charts?
To study and identify the possible relationship between the observed changes in two different sets of variables. Scatter charts are especially useful in the measurement and analysis phases of the Lean Six Sigma methodology.
What do scatter charts do?
– Provides the data to confirm a hypothesis that two variables are related.
– Provides both a visual and statistical means of testing the strength of a possible relationship
– Ensures proper follow-up of a Cause & Effect diagram to find out if there is more than just a consensus relationship between the causes and the effect. However, you must be careful not to misinterpret what your chart means.
What Can Scatter Charts Do For You?
Understanding which causes generate which effects is absolutely essential to control variation in any process. A cause and effect diagram can help you identify possible causes. Scatter charts can help you test them. By knowing which elements of your process are related and how they are related, you know what to check or what to vary to influence a quality attribute.
Scatter charts can be used to prove that there is a suspected cause-effect relationship between two process variables. With the results of these diagrams, you can design experiments or make adjustments to center your processes and control variation.