Statistical models as flextables

Convert lm, glm, mixed models, hypothesis tests, and clustering objects directly to formatted flextables with as_flextable().

as_flextable() is a generic function. For supported model classes, it extracts the relevant statistics and produces a ready-to-use flextable with appropriate column labels, formatting, and footer notes. All methods require the broom or broom.mixed package.

Enabling automatic printing in R Markdown

Call use_model_printer() once in a setup chunk to register as_flextable() as the knitr print method for all supported model classes:

# In a setup chunk
flextable::use_model_printer()

After this call, any supported model object printed in a chunk renders as a flextable automatically. Supported classes include lm, glm, lme, merMod, gls, nlme, brmsfit, glmmTMB, glmmadmb, gam, kmeans, and pam.

Linear models — lm

as_flextable.lm() uses broom::tidy() and broom::glance() to produce a coefficient table with a footer showing model fit statistics.

Requires: broom

library(flextable)

if (require("broom")) {
  lmod <- lm(
    rating ~ complaints + privileges + learning + raises + critical,
    data = attitude
  )
  ft <- as_flextable(lmod)
  ft
}

Columns produced:

Column	Header label
term	(blank)
estimate	Estimate
std.error	Standard Error
statistic	t value
p.value	Pr(>|t|)
signif	(blank, optional)

The signif column is added only when getOption("show.signif.stars") is TRUE. Remove it with options(show.signif.stars = FALSE).

Footer lines added automatically:

• Residual standard error and degrees of freedom
• Multiple R-squared and Adjusted R-squared
• F-statistic, degrees of freedom, and p-value

Generalized linear models — glm

as_flextable.glm() uses broom::tidy() and summary() to produce a coefficient table.

Requires: broom

if (require("broom")) {
  dat <- attitude
  dat$high.rating <- (dat$rating > 70)
  probit.model <- glm(
    high.rating ~ learning + critical + advance,
    data = dat,
    family = binomial(link = "probit")
  )
  ft <- as_flextable(probit.model)
  ft
}

Columns produced:

Column	Header label
term	(blank)
estimate	Estimate
std.error	Standard Error
statistic	z value
p.value	Pr(>|z|)
signif	(blank, optional)

Footer lines added automatically:

• Dispersion parameter
• Null deviance and degrees of freedom
• Residual deviance and degrees of freedom

Mixed models — merMod, lme, gls, nlme

as_flextable.merMod() handles objects from lme4 (merMod), nlme (lme, gls, nlme), brms (brmsfit), glmmTMB, and glmmadmb. The same method is registered for all of these classes.

Requires: broom.mixed

if (require("broom.mixed") && require("nlme")) {
  m1 <- lme(distance ~ age, data = Orthodont)
  ft <- as_flextable(m1)
  ft
}

as_flextable(x, add.random = TRUE, ...)

Parameter	Description
x	A mixed model object.
add.random	If TRUE (default), random effects are appended to the table as a grouped section below fixed effects.

Columns produced:

Column	Header label
effect	(group label row)
group	(shown for random effects)
term	(blank)
estimate	Estimate
std.error	Standard Error
df	(degrees of freedom)
statistic	(blank)
p.value	p-value (if available)

Fixed and random effects are separated into grouped rows using as_grouped_data(). If the model does not produce p-values, those columns are omitted.

Footer lines added automatically:

• Square root of estimated residual variance (sigma)
• Log-likelihood
• AIC and BIC

Hypothesis tests — htest

as_flextable.htest() converts any object of class htest (e.g. from t.test(), chisq.test(), wilcox.test()) into a single-row flextable.

if (require("stats")) {
  M <- as.table(rbind(c(762, 327, 468), c(484, 239, 477)))
  dimnames(M) <- list(
    gender = c("F", "M"),
    party = c("Democrat", "Independent", "Republican")
  )
  ft_1 <- as_flextable(chisq.test(M))
  ft_1
}

The function extracts estimate, statistic, p.value, and parameter from the htest object. Additional fields included when present:

Field	Condition
conf.low, conf.high	When conf.int is non-null
method	Always included when present
alternative	Always included when present

For Welch two-sample t-tests, the difference of means is prepended as estimate.

When show.signif.stars is TRUE, significance codes are appended to the p-value cell and a footer note is added.

Clustering — kmeans and pam

kmeans

as_flextable.kmeans() uses tabulator() internally to lay cluster statistics across columns:

if (require("stats")) {
  cl <- kmeans(scale(mtcars[1:7]), 5)
  ft <- as_flextable(cl)
  ft
}

as_flextable(x, digits = 4, ...)

Each cluster becomes a column. Rows show per-cluster statistics:

• withinss — within-cluster sum of squares
• size — number of observations
• Cluster centers for each variable

Footer lines:

• Total sum of squares
• Total within-cluster sum of squares
• Between-cluster sum of squares
• BSS/TSS ratio
• Number of iterations

pam

as_flextable.pam() works similarly for cluster::pam() objects:

if (require("cluster")) {
  dat <- as.data.frame(scale(mtcars[1:7]))
  cl <- pam(dat, 3)
  ft <- as_flextable(cl)
  ft
}

as_flextable(x, digits = 4, ...)

Each cluster column shows statistics such as size, max.diss, avg.diss, diameter, separation, and avg.width. Medoid rows are flagged with * in the row label. The footer shows the average silhouette width.

Significance stars

All model methods check getOption("show.signif.stars"). When it is TRUE:

• A signif column is added with "***", "**", "*", ".", or " " codes.
• A footer note with the significance legend is appended.

To suppress significance stars globally:

options(show.signif.stars = FALSE)

Supported model classes

Class	Source package	Method
lm	stats	as_flextable.lm()
glm	stats	as_flextable.glm()
merMod	lme4	as_flextable.merMod()
lme	nlme	as_flextable.lme() (alias for merMod)
gls	nlme	as_flextable.gls() (alias for merMod)
nlme	nlme	as_flextable.nlme() (alias for merMod)
brmsfit	brms	as_flextable.brmsfit() (alias for merMod)
glmmTMB	glmmTMB	as_flextable.glmmTMB() (alias for merMod)
glmmadmb	glmmADMB	as_flextable.glmmadmb() (alias for merMod)
htest	stats	as_flextable.htest()
kmeans	stats	as_flextable.kmeans()
pam	cluster	as_flextable.pam()