Day 11 Applied Linear Mixed Models II

June 24th, 2025

11.1 Review: Designs

When we model data generated by designed experiments, we typically put the information from the treatment as affecting the mean (\(\mu\)), while the elements from the design (topographical, design, or the logistics of the experiment) indicate which observations are independent and which ones are not.

  • Completely randomized designs: independent EUs that allow us to make the assumption of independence.
  • Randomized complete block designs: groups of similar EUs: the assumption of independence does not stay the same anymore, but we can say that observations from a same block are all related to each other.
  • Other types of designs: split-plot and strip-plot

11.1.1 Let’s look at the anova table

Table 11.1: Treatment ANOVA for a two-way treatment structure
Source df
Factor A a-1
Factor B b-1
A x B (a-1)(b-1)
Parallels N-(ab)
Total N-1

11.2 Tomorrow

  • Zoom classes – we will use the link for the office hours