Previous Page

|

Next Page

Interaction Plot

The interaction plot is also available for regular fractional factorial experiments. This plot is not recommended for use with Resolution 3 designs, because two-factor interactions are aliased with main effects. For Resolution 4 designs, it is possible to estimate some two-factor interactions, but you should refer to the alias structure when using the interaction plot. In Resolution 5 designs, all two-factor interactions are free of each other (though they might be aliased with even higher-order interactions). For this reason, it is safest to use the interaction plot in the Resolution 5 case.

To view the interaction plot, click the Interaction Plot tab.

There is an apparent interaction between HTIME and GSIZE. Right-click, select Zoom, and click on the HTIME*GSIZE cell to magnify it.

Now, zoom in on the MTEMP*SSPD cell. If you were to calculate the difference in slopes between these two lines, it would be the same as the difference in slopes between the two lines in the HTIME*GSIZE cell.

Recall that MTEMP*SSPD and HTIME*GSIZE are aliased. Therefore, it is impossible to determine whether one or both of these effects are significant, even though their interaction plot strongly suggests a significant effect. These effects could be separated through a follow-up. See Chapter 13 on design augmentation to break an alias chain.