|
Analysis of Variance (ANOVA) is a statistical technique that is used to compare whether samples are drawn from populations that have the same mean (average), or whether the population means are significantly different. For example a manufacturer might have five potential suppliers of a steel for an application where strength is important and want to know if the average strength of the steel from all five suppliers is essentially the same.
The ANOVA method analyses the variation in the data set to discover the amount of variation that can be attributed to each factor. In the example there is only one factor, the strength of the steel. The ANOVA method also calculates the amount of ‘common cause variation'.
A hypothesis test is then carried out to see if the amount of variation from each source is statistically significant. The test uses the F Test to compare the variation due to each factor to the amount of 'common cause variation'. See One Way ANOVA and Two Way ANOVA for worked examples.
The ANOVA method requires the process conforms to a normal distribution. Refer to Nonparametric Tests for nonparametric equivalents.
For more information on ANOVA refer to MiC Quality Six Sigma Glossary and Advanced Statistics course.
|
