Tidying methods for a spicy_multiresponse_table

Standard broom::tidy() and broom::glance() interfaces for an object returned by table_multiresponse().

Usage

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

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

Arguments

x

A spicy_multiresponse_table returned by table_multiresponse().

...

Ignored, for S3 compatibility.

Value

A tbl_df (or data.frame when tibble is not installed).

Details

tidy() returns one row per (item x group), the margin included and flagged, and the items only: the rows of none_row and the base row are left out. Columns: variable (the item's column), label (its displayed label), group (the by level or the margin, NA without by), .group_role (the role of group: "group" or "total", NA without by), total (whether the row is the margin), n (the respondents who chose the item), base (the respondents of the group's base; in the margin, those of the groups with data for the item), proportion_cases (n / base, the share of respondents), proportion_responses (n over the choices made in the group), and the item's test: test_type ("chi_squared", "n1_chi_squared", "fisher_exact", "fisher_midp" or "fisher_exact_mc"), statistic (the chi-squared, Pearson's or the N-1, NA for an exact test), df, p.value (adjusted under p_adjust) and p_unadjusted.

glance() returns one row for the question. Columns: n_items, n_items_tested (the items the global test kept), n_groups (the groups of by, NA without it), nobs (the respondents of the base), test_type ("rs2" or "bootstrap", from global_test), statistic, df, p.value and boot_n of the global test, rs_cv (the \(\hat a\) of the second-order Rao-Scott correction, see Tests in table_multiresponse()), and p_adjust, the adjustment of the item p-values.

Both column sets are FIXED: a column the table does not show (a test it did not run, an adjustment it did not make) is present and NA. Index by name rather than by position.