The Correlation and Regression Analysis is used to show the relationship of one
characteristic to another. If a dependent relationship exists the regression
plot will indicate strong correlation. If the two characteristics analyzed have
little or no relationship, the regression plot will indicate little
correlation. The amount of correlation is visually determined by observing how
the data points behave. A tightly grouped pattern over the range of the plot
generally indicates a stronger relationship than a loosely grouped pattern.
Uses of Regression Analysis
-
To determine the relationship between two process variables. If a cause and
effect is present the regression analysis will indicate strong correlation.
-
Identification of the key variable or variables which contribute to variation
in other characteristics. Isolating the causes of variation provides knowledge
of which characteristics should be continuously monitored for process control.