George SavvaNov 21, 2023
Nearly got caught out by stats::anova() doing type 1 ANOVA but lmerTest::anova() doing Type 3. #rstats.
Show threadDenis Meuthen
@georgemsavva There's also car::Anova(). To avoid this issue, simply get used to define the type of anova as I, II or III within your formula rather than relying on the package picking the right one for you.
Nov 21, 2023 at 1:12pmWeb
Show threadGeorge SavvaNov 21, 2023
@DenisMeuthen would you trust car::Anova for mixed models?
Show threadDenis MeuthenNov 21, 2023
@georgemsavva Yes, check https://www.bookdown.org/rwnahhas/RMPH/mlr-distinctions.html where the differences to stats::anova() are dicscussed.
5.12 Which to use when? car::Anova(), anova(), gmodels::estimable(), and predict() | Introduction to Regression Methods for Public Health Using R

An introduction to regression methods using R with examples from public health datasets and accessible to students without a background in mathematical statistics.