sjPlot (version 2.8.4)

tab_model: Print regression models as HTML table


tab_model() creates HTML tables from regression models.


  show.intercept = TRUE,
  show.est = TRUE, = 0.95,
  show.ci50 = FALSE, = NULL,
  show.std = NULL,
  show.p = TRUE,
  show.stat = FALSE,
  show.df = FALSE,
  show.zeroinf = TRUE,
  show.r2 = TRUE,
  show.icc = TRUE, = TRUE,
  show.ngroups = TRUE,
  show.fstat = FALSE,
  show.aic = FALSE,
  show.aicc = FALSE, = FALSE,
  show.loglik = FALSE,
  show.obs = TRUE,
  show.reflvl = FALSE,
  terms = NULL,
  rm.terms = NULL,
  order.terms = NULL,
  title = NULL,
  pred.labels = NULL,
  dv.labels = NULL,
  wrap.labels = 25,
  bootstrap = FALSE,
  iterations = 1000,
  seed = NULL,
  robust = FALSE, = NULL,
  vcov.type = c("HC3", "const", "HC", "HC0", "HC1", "HC2", "HC4", "HC4m", "HC5", "CR0",
    "CR1", "CR1p", "CR1S", "CR2", "CR3"),
  vcov.args = NULL,
  string.pred = "Predictors",
  string.est = "Estimate",
  string.std = "std. Beta", = "CI", = "std. Error",
  string.std_se = "standardized std. Error",
  string.std_ci = "standardized CI",
  string.p = "p",
  string.df = "df",
  string.stat = "Statistic",
  string.resp = "Response",
  string.intercept = "(Intercept)",
  strings = NULL,
  ci.hyphen = " – ",
  minus.sign = "-", = FALSE, = FALSE,
  linebreak = TRUE,
  col.order = c("est", "se", "std.est", "", "ci", "", "ci.inner",
    "ci.outer", "stat", "p", "df.error", "response.level"),
  digits = 2,
  digits.p = 3, = 2,
  emph.p = TRUE,
  p.val = c("wald", "kenward", "kr", "satterthwaite", "ml1", "betwithin"), = c("numeric", "stars", "numeric_stars", "scientific", "scientific_stars"),
  p.threshold = c(0.05, 0.01, 0.001),
  p.adjust = NULL,
  case = "parsed",
  auto.label = TRUE,
  prefix.labels = c("none", "varname", "label"),
  bpe = "median",
  CSS = css_theme("regression"),
  file = NULL,
  use.viewer = TRUE



One or more regression models, including glm's or mixed models. May also be a list with fitted models. See 'Examples'.


A character vector, naming a function that will be applied on estimates and confidence intervals. By default, transform will automatically use "exp" as transformation for applicable classes of model (e.g. logistic or poisson regression). Estimates of linear models remain untransformed. Use NULL if you want the raw, non-transformed estimates.


Logical, if TRUE, the intercepts are printed.


Logical, if TRUE, the estimates are printed.

Either logical, and if TRUE, the confidence intervals is printed to the table; if FALSE, confidence intervals are omitted. Or numeric, between 0 and 1, indicating the range of the confidence intervals.


Logical, if TRUE, for Bayesian models, a second credible interval is added to the table output.

Logical, if TRUE, the standard errors are also printed. If robust standard errors are required, use arguments, vcov.type and vcov.args (see standard_error_robust and this vignette for details).


Indicates whether standardized beta-coefficients should also printed, and if yes, which type of standardization is done. See 'Details'.


Logical, if TRUE, p-values are also printed.


Logical, if TRUE, the coefficients' test statistic is also printed.


Logical, if TRUE and p.val = "kr", the p-values for linear mixed models are based on df with Kenward-Rogers approximation. These df-values are printed. See p_value for details.


Logical, if TRUE and model has a zero-inflated model part, this is also printed to the table.


Logical, if TRUE, the r-squared value is also printed. Depending on the model, these might be pseudo-r-squared values, or Bayesian r-squared etc. See r2 for details.


Logical, if TRUE, prints the intraclass correlation coefficient for mixed models. See icc for details.

Logical, if TRUE, prints the random effect variances for mixed models. See get_variance for details.


Logical, if TRUE, shows number of random effects groups for mixed models.


Logical, if TRUE, the F-statistics for each model is printed in the table summary. This option is not supported by all model types.


Logical, if TRUE, the AIC value for each model is printed in the table summary.


Logical, if TRUE, the second-order AIC value for each model is printed in the table summary.

Logical, if TRUE, shows the deviance of the model.


Logical, if TRUE, shows the log-Likelihood of the model.


Logical, if TRUE, the number of observations per model is printed in the table summary.


Logical, if TRUE, an additional row is inserted to the table before each predictor of type factor, which will indicate the reference level of the related factor.


Character vector with names of those terms (variables) that should be printed in the table. All other terms are removed from the output. If NULL, all terms are printed. Note that the term names must match the names of the model's coefficients. For factors, this means that the variable name is suffixed with the related factor level, and each category counts as one term. E.g. rm.terms = "t_name [2,3]" would remove the terms "t_name2" and "t_name3" (assuming that the variable t_name is categorical and has at least the factor levels 2 and 3). Another example for the iris-dataset: terms = "Species" would not work, instead use terms = "Species [versicolor,virginica]".


Character vector with names that indicate which terms should be removed from the output Counterpart to terms. rm.terms = "t_name" would remove the term t_name. Default is NULL, i.e. all terms are used. For factors, levels that should be removed from the plot need to be explicitly indicated in square brackets, and match the model's coefficient names, e.g. rm.terms = "t_name [2,3]" would remove the terms "t_name2" and "t_name3" (assuming that the variable t_name was categorical and has at least the factor levels 2 and 3).


Numeric vector, indicating in which order the coefficients should be plotted. See examples in this package-vignette.


String, will be used as table caption.


Character vector with labels of predictor variables. If not NULL, pred.labels will be used in the first table column with the predictors' names. By default, if auto.label = TRUE and data is labelled, term_labels is called to retrieve the labels of the coefficients, which will be used as predictor labels. If data is not labelled, format_parameters() is used to create pretty labels. If pred.labels = "" or auto.label = FALSE, the raw variable names as used in the model formula are used as predictor labels. If pred.labels is a named vector, predictor labels (by default, the names of the model's coefficients) will be matched with the names of pred.labels. This ensures that labels always match the related predictor in the table, no matter in which way the predictors are sorted. See 'Examples'.


Character vector with labels of dependent variables of all fitted models. See 'Examples'.


Numeric, determines how many chars of the value, variable or axis labels are displayed in one line and when a line break is inserted.


Logical, if TRUE, returns bootstrapped estimates..


Numeric, number of bootstrap iterations (default is 1000).


Numeric, the number of the seed to replicate bootstrapped estimates. If NULL, uses random seed.


Logical, shortcut for arguments and vcov.type. If TRUE, uses = "vcovHC" and vcov.type = "HC3" as default, that is, vcovHC with default-type is called (see standard_error_robust and this vignette for further details).

Character vector, indicating the name of the vcov*()-function from the sandwich or clubSandwich package, e.g. = "vcovCL", if robust standard errors are required.


Character vector, specifying the estimation type for the robust covariance matrix estimation (see vcovHC() or vcovCR() for details).


List of named vectors, used as additional arguments that are passed down to


Character vector,used as headline for the predictor column. Default is "Predictors".


Character vector, used for the column heading of coefficients. Default is based on the response scale, e.g. for logistic regression models, "Odds Ratios" will be chosen, while for Poisson models it is "Incidence Rate Ratios" etc. Default if not specified is "Estimate".


Character vector, used for the column heading of standardized beta coefficients. Default is "std. Beta".

Character vector, used for the column heading of confidence interval values. Default is "CI".

Character vector, used for the column heading of standard error values. Default is "std. Error".


Character vector, used for the column heading of standard error of standardized coefficients. Default is "standardized std. Error".


Character vector, used for the column heading of confidence intervals of standardized coefficients. Default is "standardized std. Error".


Character vector, used for the column heading of p values. Default is "p".


Character vector, used for the column heading of degrees of freedom. Default is "df".


Character vector, used for the test statistic. Default is "Statistic".


Character vector, used for the column heading of of the response level for multinominal or categorical models. Default is "Response".


Character vector, used as name for the intercept parameter. Default is "(Intercept)".


Named character vector, as alternative to arguments like or string.p etc. The name (lhs) must be one of the string-indicator from the aforementioned arguments, while the value (rhs) is the string that is used as column heading. E.g., strings = c(ci = "Conf.Int.", se = "std. Err") would be equivalent to setting = "Conf.Int.", = "std. Err".


Character vector, indicating the hyphen for confidence interval range. May be an HTML entity. See 'Examples'.


string, indicating the minus sign for negative numbers. May be an HTML entity. See 'Examples'.

Logical, if FALSE, the CI values are shown in a separate table column.

Logical, if FALSE, the SE values are shown in a separate table column.


Logical, if TRUE and = FALSE or = FALSE, inserts a line break between estimate and CI resp. SE values. If FALSE, values are printed in the same line as estimate values.


Character vector, indicating which columns should be printed and in which order. Column names that are excluded from col.order are not shown in the table output. However, column names that are included, are only shown in the table when the related argument (like show.est for "estimate") is set to TRUE or another valid value. Table columns are printed in the order as they appear in col.order.


Amount of decimals for estimates


Amount of decimals for p-values

Amount of decimals for random effects part of the summary table.


Logical, if TRUE, significant p-values are shown bold faced.


Character, for mixed models, indicates how p-values are computed. Use p.val = "wald" for a faster, but less precise computation. For p.val = "kenward" (or p.val = "kr"), computation of p-values is based on conditional F-tests with Kenward-Roger approximation for the degrees of freedom. p.val = "satterthwaite" uses Satterthwaite's approximation and "ml1" uses a "m-l-1" heuristic (see degrees_of_freedom for details). Use show.df = TRUE to show the approximated degrees of freedom for each coefficient.

Character, indicating if p-values should be printed as numeric value ("numeric"), as 'stars' (asterisks) only ("stars"), or scientific ("scientific"). Scientific and numeric style can be combined with "stars", e.g. "numeric_stars"


Numeric vector of length 3, indicating the treshold for annotating p-values with asterisks. Only applies if = "asterisk".


Character vector, if not NULL, indicates the method to adjust p-values. See p.adjust for details.


Desired target case. Labels will automatically converted into the specified character case. See to_any_case for more details on this argument. By default, if case is not specified, it will be set to "parsed", unless prefix.labels is not "none". If prefix.labels is either "label" (or "l") or "varname" (or "v") and case is not specified, it will be set to NULL - this is a more convenient default when prefixing labels.


Logical, if TRUE (the default), and data is labelled, term_labels is called to retrieve the labels of the coefficients, which will be used as predictor labels. If data is not labelled, format_parameters() is used to create pretty labels. If auto.label = FALSE, original variable names and value labels (factor levels) are used.


Indicates whether the value labels of categorical variables should be prefixed, e.g. with the variable name or variable label. See argument prefix in term_labels for details.


For Stan-models (fitted with the rstanarm- or brms-package), the Bayesian point estimate is, by default, the median of the posterior distribution. Use bpe to define other functions to calculate the Bayesian point estimate. bpe needs to be a character naming the specific function, which is passed to the fun-argument in typical_value. So, bpe = "mean" would calculate the mean value of the posterior distribution.


A list with user-defined style-sheet-definitions, according to the official CSS syntax. See 'Details' or this package-vignette.


Destination file, if the output should be saved as file. If NULL (default), the output will be saved as temporary file and openend either in the IDE's viewer pane or the default web browser.


Logical, if TRUE, the HTML table is shown in the IDE's viewer pane. If FALSE or no viewer available, the HTML table is opened in a web browser.


Invisibly returns

  • the web page style sheet (,

  • the web page content (page.content),

  • the complete html-output (page.complete) and

  • the html-table with inline-css for use with knitr (knitr)

for further use.


Standardized Estimates

Default standardization is done by completely refitting the model on the standardized data. Hence, this approach is equal to standardizing the variables before fitting the model, which is particularly recommended for complex models that include interactions or transformations (e.g., polynomial or spline terms). When show.std = "std2", standardization of estimates follows Gelman's (2008) suggestion, rescaling the estimates by dividing them by two standard deviations instead of just one. Resulting coefficients are then directly comparable for untransformed binary predictors. For backward compatibility reasons, show.std also may be a logical value; if TRUE, normal standardized estimates are printed (same effect as show.std = "std"). Use show.std = NULL (default) or show.std = FALSE, if no standardization is required.

How do I use CSS-argument?

With the CSS-argument, the visual appearance of the tables can be modified. To get an overview of all style-sheet-classnames that are used in this function, see return value for details. Arguments for this list have following syntax:

  1. the class-names with "css."-prefix as argument name and

  2. each style-definition must end with a semicolon

You can add style information to the default styles by using a + (plus-sign) as initial character for the argument attributes. Examples:
  • css.table = 'border:2px solid red;' for a solid 2-pixel table border in red.

  • css.summary = 'font-weight:bold;' for a bold fontweight in the summary row.

  • css.lasttablerow = 'border-bottom: 1px dotted blue;' for a blue dotted border of the last table row.

  • css.colnames = '+color:green' to add green color formatting to column names.

  • css.arc = 'color:blue;' for a blue text color each 2nd row.

  • css.caption = '+color:red;' to add red font-color to the default table caption style.