spec_grid() builds a tibble in which each row is one fully specified,
defensible analysis. Every named argument is one analytic choice; the grid
is the Cartesian product of all choices (via tidyr::expand_grid()), with a
leading .spec_id column added.
Value
A tibble with one row per specification and a
leading integer column .spec_id.
See also
adjuster_sets() for the common case of toggling covariates.
Examples
spec_grid(
adjust_age = c(TRUE, FALSE),
transform = c("identity", "log"),
missing = c("complete", "impute")
)
#> # A tibble: 8 × 4
#> .spec_id adjust_age transform missing
#> <int> <lgl> <chr> <chr>
#> 1 1 TRUE identity complete
#> 2 2 TRUE identity impute
#> 3 3 TRUE log complete
#> 4 4 TRUE log impute
#> 5 5 FALSE identity complete
#> 6 6 FALSE identity impute
#> 7 7 FALSE log complete
#> 8 8 FALSE log impute