Tidy ANOVA Tables

Andrew Zieffler

6/30/2019

The tidy_anova() function can be used to create an ANOVA table that is outputted as a data frame. This is useful for implementation with kable() and other R functions. The function takes as its input an lm object.

# Fit model with lm()
lm.1 = lm(mpg ~ 1 + wt + hp, data = mtcars)

# Load educate package
library(educate)

# Obtain ANOVA table
tidy_anova(lm.1)

##      Source df        SS         MS         F            p
## 1        wt  1 847.72525 847.725250 126.04109 4.488360e-12
## 2        hp  1  83.27418  83.274183  12.38133 1.451229e-03
## 3 Residuals 29 195.04775   6.725785        NA           NA

Model-Level Output

The function takes an optional argument of model=TRUE. This argument will create the model-level ANOVA table that is commonly presented in APA publications.

# Obtain model-level ANOVA table
tidy_anova(lm.1, model = TRUE)

##      Source df       SS         MS        F            p
## 1     Model  2 930.9994 465.499716 69.21121 9.109047e-12
## 2 Residuals 29 195.0478   6.725785       NA           NA

Pretty Tables with kable()

Since tidy_anova() outputs a data frame, you can use knitr’s kable() function to make it look great in a knitted document.

#Load libraries
library(dplyr)
library(knitr)

# Set knitr options to remove NA values
options(knitr.kable.NA = '')

# Obtain model-level ANOVA table and make it pretty
tidy_anova(lm.1, model = TRUE) %>%
  kable(digits = 2)

Source	df	SS	MS	F	p
Model	2	931.00	465.50	69.21	0
Residuals	29	195.05	6.73