Side-by-side modifier-level treatment effects from DA, SWADA, and CAMS

Compares pooled modifier-level effects from DA, SWADA, and CAMS (da() or the DA part of conventional(), swada(), cams()). These are not three equivalent statistical models: DA pools each modifier series separately, SWADA applies one common weight vector, and CAMS is the contribution-adjusted model. Average Difference (ad()) does not pool modifier-specific treatment effects, so it is omitted here.

Usage

compare_modifiers(da, swada, cam = NULL, exponentiate = TRUE, type = "Overall")

Arguments

da

A da() or conventional() fit, or a cams() fit when comparing two CAMS approaches.

swada

A swada() fit, or a second cams() fit.

cam

A cams() CAMS fit. Optional when da and swada are already CAMS fits.

exponentiate

If TRUE (default), report on the exponentiated (ratio) scale. Use FALSE for the fitted log scale.

type

CAMS layer ("Overall" default).

Value

A tibble with one row per approach × subgroup and class interactmeta_compare_subgroups (plottable).

Details

Warns if mixed constructions do not share the same engine and method. A like-for-like frequentist comparison uses metafor / REML throughout (cams(..., method = "REML")). Two CAMS fits of the same model with different methods are labelled (CAMS — Bayesian vs CAMS — REML) and are not treated as a mismatch.

Two or more cams() fits may be passed instead of DA / SWADA / CAMS, in which case plot() draws the compared subgroup estimates.