I reiterate: Don't forget to perform a residual analysis, recall the Anscombe dataset. Attached is a data table with 3 scripts: Fits a covariance model (interaction b12 * X * Color) Fits independent models drawn on the same graph Fits independent models drawn on two side-by-side graphs (shown below) The slopes intercepts, RSquare and RMSE are essentially the same, however Blue is not a good fit. Same summary statistics, Blue is not a good fit!
... View more
Don't worry it's possible. You just have to make sure that you pick the right batches. To be on the save side you just need two batches of each relevant color. One of those batches in each color should have a relatively large particle size and the other one should have a small particle size. You could for example try to find 8 batches with the following parameters: Batch Color Particle Size A Yellow small (maybe ~1000) B Yellow large (maybe ~2000) C White small D White large E Red small F Red large G Green small H Green large The particle sizes do not need to be perfectly 1000 or 2000, but try to find batches that are on the higher end (or lower end) of particle sizes and you will be fine. If you are able to find these specific kinds of batches you can actually treat "Color" and "Particle Size" just as a factor that you can control. If you just have some given batches with known color and particle size you could actually treat "Color" and "Particle Size" as a covariate as well. There is a nice article about that in the JMP help: https://www.jmp.com/support/help/Experiments_with_Covariates.shtml If I understand you correctly the example there is pretty much what you are trying to do. Kind regards, Sebastian
... View more