Don’t use the ‘one factor at a time’ (OFAT) approach when designing experiments

Home / News, Views & Updates / Don’t use the ‘one factor at a time’ (OFAT) approach when designing experiments

Don’t use the ‘one factor at a time’ (OFAT) approach when designing experiments

Today’s blog post is written by Albert Plant who delivers a wide range of Continuous Process Improvement courses.

I am greatly surprised at the number of experimenters I meet who still don’t understand the basics of designing experiments. This includes engineers and scientists, many with PhD’s in leading multinational companies with household names, who are using outdated methods to design experiments. It is odd to see experimenters making calls on their most up to date iPhones, but using experimental design techniques from the nineteenth century. The most common error is the use of the OFAT approach. OFAT stands for “one factor at a time”; essentially, the experimenter has several factors in which they have an interest, but they study the factors one at a time.

There are two serious downsides to using OFAT; (a) the method is grossly inefficient, leading to an unnecessarily large number of experimental runs, (b) more seriously, the experimenter is unable to study interactions among the factors. The alternative and correct approach is to use factorial methods, in which the levels of all factors of interest are varied simultaneously. I think my students are surprised to learn that the factorial approach to designing experiments has been in use since the 1920’s, and, therefore, the OFAT method they are using is almost a century out of date!! It is true that in the early days many people would have found the analytical methods of the factorial approach complex to use, but this is no longer the case with an array of special DOE (Design of Experiments) computer software programmes readily available at reasonable prices.

Engineers and scientists can be readily trained to use the most up to date and most mathematically advanced software tools to design and analyse their experiments, and thereby gain the maximum amount of information from their work.

Share this Article

Blog Sign up

Sign up to receive the latest industry and company news direct to your inbox.