Low-level constructor for the `csem` S3 class. This is an internal building block used by `csem_gt()`. End users should not call this directly.
Usage
new_csem(
estimates,
by_score,
call,
paradigm,
methods,
error_types,
arguments,
variance_components = NULL,
smooth_fits = NULL,
diagnostics = NULL,
bootstrap = NULL,
scale_transform = NULL,
n_persons,
n_items,
version = .csemGT_version()
)Arguments
- estimates
A data frame, one row per person, in wide format. Must contain at minimum the identifier columns `person_id`, `observed_score`, `group_size`.
- by_score
A data frame, one row per distinct value of the conditioning variable, also in wide format.
- call
A captured `match.call()` of the user-facing function.
- paradigm
Character scalar; one of `"gt"`, `"split_half"`, `"anova"`, `"binomial"`.
- methods
Character vector of methods applied.
- error_types
Character vector of error types reported.
- arguments
Named list with the resolved arguments of the user-facing call.
- variance_components
Named list with ANOVA-based variance components (used by the GT paradigm). May be `NULL`.
- smooth_fits
Named list of smoother diagnostics (one element per smoothed column). May be `NULL`.
- diagnostics
Named list of smoother sample-size diagnostics (`n_floor`, `n_ceiling`, `n_fit`). May be `NULL`.
- bootstrap
Named list with bootstrap metadata and replicates, or `NULL` when bootstrap is not performed.
- scale_transform
Scale transformation specification, or `NULL`.
- n_persons, n_items
Integer scalars.
- version
Package version. Defaults to the installed version of `csemGT`; falls back to `"0.0.0"` when the package is being built.