Errata
The following updates to the text have been made in the online version only:
May 2025
- Created this Errata section
- Replaced exclude = F with useNA = “ifany” in a few places in Section 6.22 (“Ordinal logistic regression)
- In a pre-published version, datasets had variable labels created using Hmisc::label(). These caused errors in some places, so were removed, however a reference to these labels for the COVID dataset remained in the Appendix. This reference has been removed.
- Updated Section 5.19 (“Box-Cox outcome transformation”) to use (Yλ−1)/λ instead of Yλ.
February 2025
- Chapter 4, Exercise 7: Changed the first “hypothesis” to plural
- Chapter 4, Exercise 17
- What is the predicted mean heart rate (and 95% CI) for those with a BMI of 20 kg/m2? For those with a BMI of 35 kg/m2?
- Changed “predicted” to “estimated”
- Section 5.9.4
- Equation 5.2, the coefficient for the gender main effect is positive, not negative
- “females have an average fasting glucose that is 1.746 mmol/L lower (since the coefficient is negative) than males.”
- Lower changed to greater
- negative changed to positive
- The Section “Visualizing the Unadjusted” moved up a level so it is easier to find
- Section 5.19 (Box-Cox), Example 5.1
- Deleted the following: “Additionally, the errors were not normally distributed, although this by itself is not an issue due to the large sample size.”
- Modified the beginning of “Re-check normality”
- Multiple Linear Regression exercises 30 and 33: replaced “estimated” with “estimated mean” and “for a country” with “among countries”