In the formulation of automotive paint, particular attention is given to what additives will minimize the fading of the paint color in sunlight. Recently there have been complaints about paint fading and the chemists were instructed to find a different formula to address this problem.
In the past when the chemists conducted experiments they focused on studying only one factor at a time, thinking that they would find clearer answers to their questions. Unfortunately, studying only one factor at a time prevents a chemist or any other experimenter from learning about interactions among the factors which occur frequently in paint formulations.
Working with the group of paint chemists, we tackled the issue of which additives and the amount of each would work best to prevent fading at the lowest cost. Together we designed an experiment with multiple factors to study the additives. After the experiment was conducted and the data were gathered, we analyzed the data together and discovered clear and convincing results as to which additives at which amounts did the best job of preventing fading. Several combinations of additives and amounts worked equally well. We proposed the one with the least cost which also turned out to have the least variability.
For several of the chemists, this experiment was their first experience using multiple factors and a small number of observations. As a result, they were hesitant to believe that such a critical decision could be made with such a small number of data values and consequently were reluctant to sign on to this recommendation. I suggested that we set up several confirmation runs to see if the results were the same. Data from the confirmation runs supported the results from the original experiment and the chemists made a recommendation for a new formula.