Analysis of Variance
In order to take account of interactions, the AIAG manual
recommends the use of ANOVA. Mesa
also generates a fully linked ANOVA worksheet and the output for the AIAG data
is shown below (the AIAG version is shown on page 74 of the 2nd
Note that Mesa signifies whether or not a particular
effect is significant. We can see
that the AIAG data exhibits a statistically significant interaction effect.
The Wheeler & Lyday approach regards this as a problem which should
be eliminated, rather than a parameter which should be estimated.
The ANOVA worksheet in Mesa is fully linked to the values
on the Raw Data sheet and therefore updates automatically.