Put estimates from several ties side by side

The simplest and least committal of the three combinations: line the ties up on the same cells and look at them. Nothing is averaged and nothing is assumed — which is why it is worth doing first, and often worth doing instead.

Usage

compare_ties(..., estimator = c("ind", "agg"))

Arguments

...

estimator results, named by tie: compare_ties(siblings = a, cousins = b)

estimator

"ind" (individual visibility) or "agg" (aggregate)

Value

a tie_comparison: a tibble with one row per cell per tie, carrying each tie's estimate, interval and visibility provenance

Details

Where two ties disagree by more than their intervals allow, that is information: the ties differ in who they reach, in how completely they are reported, or in how well their visibility rule holds. Averaging that away before looking at it would be a mistake.

See also

pool_ties() to average them, ties_union_check() for the third combination

Examples

if (FALSE) { # \dontrun{
compare_ties(siblings = sib_est, cousins = cousin_est)
} # }