Extract and combine estimates and goodness-of-fit statistics from several statistical models.
extract(
models,
statistic = "std.error",
statistic_override = NULL,
statistic_vertical = TRUE,
conf_level = 0.95,
coef_map = NULL,
coef_omit = NULL,
gof_map = modelsummary::gof_map,
gof_omit = NULL,
add_rows = NULL,
add_rows_location = NULL,
stars = FALSE,
fmt = "%.3f",
...
)
a single model object or a (potentially named) list of models to summarize
string name of the statistic to include in parentheses
Typical values: "conf.int", "std.error", "statistic", "p.value"
Alternative values: any column name produced by `broom::tidy(model)`
manually override the uncertainy estimates. This argument accepts three types of input:
a function or list of functions of length(models) which produce variance-covariance matrices with row and column names equal to the names of your coefficient estimates. For example, `R` supplies the `vcov` function, and the `sandwich` package supplies `vcovHC`, `vcovHAC`, etc.
a list of length(models) variance-covariance matrices with row and column names equal to the names of your coefficient estimates.
a list of length(models) vectors with names equal to the names of your coefficient estimates. Numeric vectors are formatted according to `fmt` and placed in brackets, character vectors printed as given.
TRUE if statistics should be printed below estimates. FALSE if statistics should be printed beside estimates.
confidence level to use for confidence intervals
named character vector. Names refer to the original variable names. Values refer to the variable names that will appear in the table. Coefficients which are omitted from this vector will be omitted from the table. The table will be ordered in the same order as this vector.
string regular expression. Omits all matching coefficients from the table (using `stringr::str_detect`).
data.frame with four columns: `raw`, `clean`, `fmt`, and `omit`. See `modelsummary::gof_map`
string regular expression. Omits all matching gof statistics from the table (using `stringr::str_detect`).
list of character vectors, each of length equal to the number of models + 1.
integer or NULL. custom rows will be added to the bottom of the table if this parameter is NULL, or after the position set by this integer.
to indicate statistical significance
FALSE (default): no significance stars.
TRUE: *=.1, **=.05, ***=.01
Named numeric vector for custom stars such as `c('*' = .1, '+' = .05)`
string which specifies how numeric values will be rounded. This string is passed to the `sprintf` function. '%.3f' will keep 3 digits after the decimal point with trailing zero. '%.5f' will keep 5 digits. '%.3e' will use exponential notation. See `?sprintf` for more options.
all other arguments are passed to the `tidy` method used to extract estimates from the model. For example, this allows users to set `exponentiate=TRUE` to exponentiate logistic regression coefficients.
tibble
# NOT RUN {
library(modelsummary)
data(trees)
models <- list()
models[['Bivariate']] <- lm(Girth ~ Height, data = trees)
models[['Multivariate']] <- lm(Girth ~ Height + Volume, data = trees)
extract(models)
# }
Run the code above in your browser using DataLab