vue - Experiments 1B - Some basic terminology

Videos used in the Coursera course: Experimentation for Improvement. Join the course for FREE at https://www.coursera.org/learn/experimentation These videos are also part of the free online book, "Process Improvement using Data", http://yint.org/pid Full script for the video: http://yint.org/scripts/1B -------------------- In the previous class we learned why experiments are so important. If you're not sure why we should be running experiments, take a moment and review the previous lesson. But in today's class we're going to learn some new terminology that we're going to use throughout the course. There are some important words we have to be clear about. Every experiment you run will always have an outcome. An outcome is what will happen as a result of the experiment. It is what we are interested in improving. That's the subject of this course: experimentation for improvement, so we need to have something to improve, called the "outcome". Many experiments will also have one or more factors. Factors are the things you change to influence the outcome. In this course we consider cases with two or more factors, but it is totally possible to run experiments with just one factor, where you are only changing one thing to see the effect on the outcome. Let's use an example of growing plants to see this terminology at work now. The outcome of growing these plants could be the height of the plant. An important aspect about the outcome is that it is always measurable, at least in some way. In other words, once you have finished your experiment, you must have some measurement. In the plant example a different outcome that you could have used might have been the average length of the leaves, or it might have been the number of flowers on the plant. Both of these would be numeric measurements. In most experiments, the outcome is a number, a quantitative measurement. But, a qualitative measurement is also possible. For example, perhaps the outcome is the colour of the flower: light red, red, or dark red. That's a qualitative measurement; a description of what happened. When we combine an outcome and the need to adjust the outcome, we get what we call an objective. In other words, outcome plus a desire to adjust the outcome equals the objective. Here are some examples of objectives. Maximize the height of a plant, minimize the amount of pollution. We most often want to maximize or minimize our outcomes, but sometimes we want our outcome to remain the same even though we are changing the factors. For example, let's say you want to change your recipe for your favourite pastry to be gluten free. Now your objective is, you want the taste to be the same as the regular recipe. Your outcome is taste and your objective is the same. Note that you don't always need an objective to run an experiment. Every experiment always has an outcome. Not every experiment has an objective, though. But we usually have one in mind. One more thing. Sometimes you'll hear the word response used instead of the word "outcome". In the area of experiments, the word response and outcome mean the same thing. Our next piece of terminology is the word "factors". In the growing plants example there could've been three factors that you changed. The amounts of water that you gave the plant each day. The amounts of fertilizer that you gave the plant each week. And using a soil type A or a soil type B. All experiments must have at least one factor that is changed. However, as you'll learn in the next few classes that you should always consider as many factors as possible, not just one or two. So, to summarize, every experiment will always have two components: an outcome and one or more factors. Let's try this quick exercise. At your company, your manager asks you to change the recipe for biodegradable plastic. You've been making this plastic for two years now, but recently customers have been complaining the plastic does not break down fast enough. The recipe to make the plastic uses a specific chemical call polylactic acid, or PLA. Next, I'd like to discuss a bit more about the factors. We can distinguish between two main types of factors, numeric factors and categorical factors. Numeric factors are referred to as quantitative, since they can be measured as a numeric quantity. Let's go back to the plants example. We may decide to use 15 millilitres of water, or 30 millilitres of water. We could use six drops of fertilizer, or we could add ten drops of fertilizer. In these examples, we can numerically measure the factor. A key point about numeric factors is that we are able to select and adjust their numeric level. We also have categorical factors. These are factors that can take on a limited number of values and are usually not quantified numerically. For example, we could use soil type A or soil type B. We could put a bag over the plants to create a greenhouse effect, or we could leave the plants exposed. The factors in your ...