Document drop1.loglm() vs. stats::drop1.glm() in @details

Brief "Comparison with drop1.glm()" subsection summarizing the findings
from dev/anova-like/drop1-compare.R: default scopes agree, Df/LR statistic
match exactly, and what drop1.loglm() adds (table/array input, Pearson X^2,
fitted-model reuse, scope validation against the generating class).

Author: friendly

R/drop1.loglm.R

@@ -50,6 +50,17 @@
 #' (still implied by `A:C`/`A:D` and `B:C`/`B:D` respectively), so the
 #' reduced model stays hierarchical.
 #'
+#' ## Comparison with `drop1.glm()`
+#' A hierarchical `loglm()` model has an equivalent fit via
+#' `glm(family = poisson)`, and `stats::drop1.glm()`'s default `scope`
+#' already restricts to the same generating-class terms as `object$margin`,
+#' so `Df` and the LR statistic agree exactly between the two. `drop1.loglm()`
+#' works directly on table/array or `Freq ~ ...` data, adds the Pearson X^2
+#' column, and returns the fitted drop-one models for reuse; its `scope`
+#' validation also guards against testing a non-generating-class term, which
+#' `drop1.glm()` will silently allow with a misleading result. See
+#' `dev/anova-like/drop1-compare.R` for a worked comparison.
+#'
 #' @param object a fitted `MASS::loglm` model
 #' @param scope character vector of term labels (colon-separated, as in
 #'   `"A:B"`) to test; must be a subset of the model's generating class.