20.0 Experimental Design Applications

Part X. Application to Experimental Design and Statistical Process Control

The analytical capabilities of Process Behavior extend to experimental design applications.^[52] In this case the “process design factors” are the experimental variables under the control of the analyst that define the “experimental treatments” that may affect the response variable of interest. The production time variable defines the number of times that the experiment is “replicated”. If there is no “replication”, then there is no production time variable associated with the available data.

20.0.1 Example of a Single Factor Experiment¶

As an example of the analysis of an experiment using the Process Behavior capabilities consider the data in Table 20-1. The response variable is “coating weight”, the treatment group is taken as a type of coating material and there were N=4 observations per treatment group.^[53] The process behavior charts produced by PB for the subgroup means and standard deviations are presented in Figure 20-1 and Figure 20-2. These charts indicate that Treatment Group 2 yields a detectably higher coating weight. The within subgroup variation appears to be comparable across the five treatment groups.

Table 20-1. Coating Weight Data

20.0.2 Example of a Multi-Factor Study¶

An experiment was conducted to study bacteria growth under several growth medium environments defined by Factor 1 – Type of Fatty Acid at 4 levels, Factor 2 – Isolate at 3 levels and Factor 3 – Concentration at 7 levels. A sample of N = 3 petri dishes were used under each experimental condition. A partial list of the data is presented in Table 20-2. The process behavior charts produced by Process Behavior are presented in Figure 20-3 through Figure 20-10 which gives a detailed analysis of the experimental design factors affecting both the average bacteria growth rate and the standard deviations associated with bacteria growth rate.

Table 20-2. Partial List of Bacteria Growth Data

`