Welcome back. In this tutorial, we're going to talk about a filtration rate experiment with a block. This is a factorial experiment with a block. We're going to talk about what happens when we don't have enough runs to perform an entire replica of our design in each block. So in this experiment, we have a two to the four factorial, and we're interested in the filtration rate of a resin. We also talked about this example back in Module 6. There were four factors, temperature, pressure, mole ratio, and stirring rate, and the experiment was performed in a pilot plant. But now we're going to suppose that the experimenter can only run eight treatment combinations in a single batch of raw material. So we're going to use raw material as a block, but we only can perform eight runs in this block. So we'll need to split our two to the or 16 run experiment into two blocks. But in order to do that, we're going to have to compound a higher-order interaction term. So this situation will confound the A, B, C, D interaction with block. If we look at this experiment, we decide that half of the treatment runs. The ones where the A, B, C, D interaction is positive will go into block one, and the other half of the 16 runs, the other eight runs are going to go into block 2. This is when A, B, C, D is at the low-level and that's how we compound block with A, B, C, D. I'm going to show you how we can create this design easily and JMP. We're going to go to our DOE menu, and this time we're going to go to classical, and we're going to go to screening Design. In this platform, we'll have our filtration rate response, and then our four two-level continuous factors. So I'm just going to put a four here and click continuous. This adds four continuous two-level factors at once. We can just call these A, B, C, and D. When I click continue, I have the option of choosing the screening type. We're going to keep it at the first bullet that says choose from a list of fractional factorial designs. When I click continue, I get a list of designs. We know that our design has 16 runs. So I'm going to go down to the last four here, and we can know that our block size is eight. So I'm going to select that design, the 16 run design with blocks of size eight. So I'll click Continue, and now I have some more options. In this display and modify design, we can go to change generating rules, and if I expand this, we see that block is confounded with A, B, C, D. This is what we want, so we don't need to change anything. But if you had a different situation where you want it to control what you are compounding with the block, this is how you would do it. I'm going to randomize within the blocks. We have no center points or extra replicates, so I'm going to make the table. Now, I have a data table that has my 16 runs. Block 1 are all the runs where A, B, C, D interaction is low, and block 2 is where all the runs where the A, B, C, D interaction is high. So if we enter our data into this table, we can use our model script here, and we'll see that we can check out all these runs and we'll include a block in our model. We cannot include block and A, B, C, D because they're the same effect. So you can only enter one when you're analyzing this model. That's how we can create a design with a block that's confounded with a higher-order interaction term. Thanks for checking out this tutorial.
