Chapter 4

[YinghuiZhouu]

• [x] create diagnosis files for chapter 4.
• [x] add the elements into book chapter 4 scripts (refer to chapter 3)
• [x] Update the diagnoses figures for chapter 4. Don't use legend, put sd and mean labels.
• [x] Cite FEDAI pages
• [x] Include definitions (Chapter2 diagnostic and diagnosands, Chapter 1 inquiry.)
• [x] Make progress on diagnoses code

[YinghuiZhouu]

• Suggestions from Don: look at DD chapter 18, perhaps we need to bring in some probability adjustment to chapter 4

Questions:

• Should pages be added when citing from FEDAI? (Yes)
• Include definition from DD? Chapter2 diagnostic and diagnosands, Chapter 1 inquiry. (Yes, will add)
• What's the code generated diagnoses.rds files? E.g. where does diagnosis <- read_rds("diagnoses/diagnosis_2_1.rds") the rds files come from (In every diagnoses R code, save a rds. In the book chapter, also include a chunk {r, eval=FALSE} to show the readers how we did the diagnoses. Similarly, show the readers what we did for simulations )
• (solved)chapter 3 line 346-351 Error: object 'designs' not found

[YinghuiZhouu]

Questions: table4.2 only includes the less reliable version of the pre-test $X{weak{i}}$, resulting very similiar variations of DIM and Block randomization in figure4.2. Should we add the stronger predictor as well?

• Why is the DIM biased here? (Because the probabilities of assignment are different!)
• declaration_4.2 only includes block_ra, should we also include complete_ra?

[YinghuiZhouu]

• [ ] Blocking vs complete vs complete + covariate adjustment

• [ ] Include a seperate Lin simulation (two examples, OLS and Lin), show the residual plot (we can do it later) (download the lin book and take a look at it)