Remembering DOE Resolution

Engineering often relies on Design of Experiments (DOE) to effectively understand significant dependencies of different input or control parameters to a desire measurable outcome. It is critical to properly design your experiments in a way that maximizes your ability to distinguish between the impact or contribution that each input variable has on your outcome. Different types of DOE designs have different resolutions. As stated here, a DOE resolution describes the degree to which main effects are confounded with 2-level interactions, 3-level interactions and so on. For example, in a resolution III design main effects are confounded with 2-level interactions. Ideally, we would like to have the highest resolution possible DOE. Full factorial designs have no confounding and thus are considered resolution “infinity”. However, having full factorial designs is often not possible due to the very large number of runs required. As a result, I often end up working with different resolution DOEs and often I need to remember which factors are confounded. For this, I always use a hand trick that was thought to me during my Design for Lean Six Sigma Training. I wanted to share this hand trick for those looking to remember this. However, rather than explaining it, I put together the following picture:


The nice thing about this trick is that it also holds true when you are dealing with say resolution VI or greater DOE. For DOE resolutions higher than V, just use your two hands, and work with the number of fingers equal to the DOE resolution as indicated above.

No comments:

Post a Comment