Hi all. So I'm working on a project where I have 2 x component classes, each of which has 4 items within the class, which are mixed together at varying concentrations over 48 combinations. Figured the best Factor way to do this was to use 2 x Continuous for the max/min concentrations for each class, then 2 x 4-level Categoricals for each of the items. Each of the combinations was made twice, so after running the rest of the experiment for my output, I've ended up with duplicate responses for each combination.
I'd like to use this to identify how much each of the component class concentrations and items mattered for my output, and predict the optimum mixture, but I'd rather not just plug in the mean of the responses since for some samples there was some significant variation between the duplicates. Can anyone advise on which analysis method is best and how I should organise the data table for this? Thanks!
