Summarize a MAIHDA Analysis — summary.maihda

Returns the variance summary (VPC/ICC, variance components, stratum estimates) of one of the analysis's models. The per-group comparison, when present, is attached as the "groups" attribute.

Usage

# S3 method for class 'maihda_analysis'
summary(object, which = c("null", "adjusted"), ...)

Arguments

object: A maihda_analysis object from maihda.
which: Which model to summarize: "null" (default) or "adjusted". A crossed-dimensions analysis has a single model and accepts only "null".
...: Additional arguments (not used). The summaries are computed by maihda, so arguments that would change them are rejected rather than ignored: set bootstrap, n_boot, conf_level, response_vpc and seed on maihda itself, and for a different fixed-effect reference summarise the fitted model directly, as in summary(x$model_adjusted, df_method = "bootstrap").

Value

The maihda_summary for the requested model.

Which model

A two-model maihda analysis holds a null model (strata as the only predictor) and an adjusted model (the dimensions' additive main effects added as fixed effects), and this returns the null model's summary by default – the one the headline VPC/ICC comes from. The strata dimensions are the random-effect grouping there, not fixed effects, so the printed fixed-effect table shows only the intercept and any covariates. Use which = "adjusted" for the model whose fixed effects carry the dimensions' additive main effects (the same object as object$summary_adjusted), matching tidy's which argument.