1. Learn
    2. /
  2. Courses
    3. /
  3. Experimental Design in R

Exercise

RCBD Model Validation

As we did in the last chapter (and when building any model!) it's a good idea to validate the results. We'll examine the Residuals vs. Fitted and Normal Q-Q plots, though now we'll only see a Constant Leverage plot in place of the other two. A good model has a Q-Q plot showing an approximately normal distribution and no clear patterns across blocks or treatments in the others.

We can also look at Interaction plots. We hope to see parallel lines, no matter which of the block or the treatment is on the x-axis. If they are, they satisfy a key assumption of the RCBD model called Additivity.

The nhanes_rcbd model object from the last exercise has been loaded for you. Examine the results with summary(nhanes_rcbd) in the console if you need a refresher.

Instructions 1/3

undefined XP
  • 1
    • Plot the nhanes_rcbd model object, being sure to set up a 2x2 grid of plots beforehand.
  • 2
    • Run the code to view the interaction plot between the treatment and gender and observe if the lines are parallel.
  • 3
    • Run the code to view the interaction plot between gender and the treatment (it'll be a little different!) and observe if the lines are parallel.