Screening experiments are lower resolution designed experiments (DOE) for investigating main effects, usually involving several factors. Screening experiments often use Fractional Factorial designs.

One of the methods used in the design and execution of experiments is the screening DOE. This technique works best when there are multiple independent variables. If you use a standard DOE design, you would need to run many experiments.

As the name suggests, the purpose of the screening DOE is to reduce the number of total runs of an experiment by screening and removing variables that are statistically insignificant. You can also use the Plackett Burman Design, which is a fractional factorial. This allows you to filter out variables that have little effect on the response.

A full factorial DOE allows you to test all possible combinations of your independent variables and their impact on your response variable. The main effects as well as all interactions are a great way to get more information.

You lose some information by using a screening DOE, especially regarding the interactions between your variables. Depending on the purpose of your experiment, this information may be or not be relevant to you. This is known as “losing resolution”.

DOE screening experiments

  • Screening experiments are the ultimate fractional factorial experiment. These experiments assume that all interactions are significant, even those in two-way relationships.
  • They screen all variables and factors in the process to determine the most critical variables that have an impact on the output.
  • There are two main types of screening experiments:
  • Drs. Plackett and Burman created the first family of screening experiment matrices in 1940.
  • Dr. Taguchi modified the Plackett–Burman screening design. Plackett-Burman was modified so that experimenters could assume interactions are not significant and still test for two-way interactions.

What is the importance of a screening DOE?

DOE by itself is quite complicated. To properly execute it, you will need to have a lot of experience. You can make your experiment run more smoothly by using a screening DOE.

It takes less effort to complete

Understanding how screening DOEs work will allow you to better manage the process.

2. Important importance of identifying variables

You may be able to reduce the number of variables by doing some statistical work (e.g. regression analysis) before you screen your DOE. You should only explore variables that have a high chance of being significant.

3. Information reduced

Screening DOEs will not provide as much information as a full factorial experiment. You should have some knowledge about the potential existence, importance, and resolution of interactions if you plan to conduct a screening DOE.

Three best practices for screening DOE

Any type of DOE can be difficult and dangerous. These tips will help you avoid many of the most common pitfalls.

1. Eliminate all noise

A DOE is intended to examine the relationship between process variables, and the output or response of the process. The experiment may be contaminated if other variables are introduced to the equation.

2. What is the purpose of screening DOEs?

Because of the potential complexity and high resource consumption, it is important to clearly explain why you are doing a DOE. Before you consider doing a DOE, be clear about your purpose and desired outcome.

3. Understanding interactions

When you conduct a screening DOE, you will lose information (or resolution). To determine whether the loss of higher-order interactions is important or not, you should be able to understand the significance of those interactions.