Feature request: Two-way ANOVA documentation

[MauKappel]

I was looking at a two-way ANOVA analysis and was confused by the p-value in the table. Usually the p-values are reported for the independent variables and not the model p-value. When I looked at the tests documentation I was still confused as it mentioned the lm() call and I was expecting anova() but not broom::glance(). However, it makes sense to do it like this for the format of the table.

gtsummary/data-raw/gtsummary_tests.csv

Line 45 in d9b1952

tbl_continuous,TRUE,FALSE,,stats,anova_2way,,gtsummary:::add_p_test_anova_2way,FALSE,lm(continuous_variable ~ by + variable),Two-way ANOVA,

It may be obvious to others, but I wanted to point out my experience.

Thanks for providing gtsummary, I love using it!

Warm regards,
Maurice

[ddsjoberg]

Dear @MauKappel thank you for the post and the kinds words!

Reviewing the documentation, I agree that it's not 100% clear which p-value from lm() is being reported.

If we provided lm(continuous_variable ~ by + variable) |> broom::glance() for the pseudo code instead, would that be more clear?

[MauKappel]

Yes then it would be clear. Thanks!