Tidying methods for a spicy_categorical_table

Standard broom::tidy() and broom::glance() interfaces for an object returned by table_categorical(). They re-shape the underlying long-format data (stored on the object as the "long_data" attribute) into the two canonical broom views so the table can be consumed by any downstream tidyverse-stats pipeline.

Usage

# S3 method for class 'spicy_categorical_table'
tidy(x, ...)

# S3 method for class 'spicy_categorical_table'
glance(x, ...)

Arguments

x

A spicy_categorical_table returned by table_categorical().

...

Currently ignored. Present for compatibility with the broom::tidy() / broom::glance() generics.

Value

A tbl_df.

Details

tidy() returns one row per (variable x level x group), the groups of by and the Total margin included, with columns variable (the column name), label (the label the rows display for it, see labels), level, .row_role (the role of level: "missing" on the row of missing values the table adds, "total" on the Total row of total_row = TRUE, "level" otherwise), group (NA without by), .group_role (the role of group: "missing" on the column of missing by values, "total" on the margin, "group" otherwise, NA without by), total (TRUE on the rows of the margin and of the Total row), n, and proportion (between 0 and 1, the share the table displays: within the group, or within the level under percent = "row", as attr(x, "percent") records; NA on the row of missing values under percent_base = "valid", which leaves them out of the base, as attr(x, "percent_base") records). Filter on the role columns and total rather than on the displayed labels, which follow the display language.

glance() returns one row per variable, in the order of select, with its test (when by is used), the requested association measure and the standardized mean difference: variable, label, test_type ("chi_squared", "n1_chi_squared", "fisher_exact", "fisher_midp", or "fisher_exact_mc", as in the long output of table_categorical()), statistic (the chi-squared, Pearson's or the N-1, NA for an exact test), df, p.value, es_type (the key of the measure the variable carries: "phi", "cramer_v", ...), es_value, es_ci_lower, es_ci_upper, smd_type, smd_value, and nobs (the observations of the variable, the margin and the Total row not counted twice). Without by, only variable, label and nobs are populated; the other columns are NA. The interval bounds are NA unless the table was built with assoc_ci = TRUE, and the SMD columns unless it was built with smd = TRUE.

See also

as.data.frame.spicy_categorical_table() for the raw wide-format access; tidy.spicy_continuous_table() for the continuous-descriptive companion.