facets_positioning_guide() gives user-facing wording for the relationship
between mfrmr and FACETS. Use it when a report, migration note, or
methods appendix must make clear that mfrmr is not a FACETS numerical
clone.
Details
The guide separates five ideas that are easy to conflate:
estimation authority: fitted values come from
mfrmrunless external FACETS output is explicitly supplied;compatibility purpose: FACETS-style names and files are transition, handoff, and report-organization surfaces;
external comparison: FACETS comparisons require a supplied external table and should separate MnSq differences from df/ZSTD convention differences;
current model boundary: one ordered-categorical response-model family and one observed score scale are used per fit; nominal and count-response families, mixed families, multiple independent scales, general threshold anchoring, and importing FACETS or third-party frozen calibrations are not part of the current public estimator; native portable calibration is available within the scope of
mfrm_calibration_capabilities(), and posterior scoring from an existing fitted object remains separate;extension surface: native R tables, plot data, GPCM diagnostics, network views, and G/D-study helpers are package extensions, not promises of FACETS menu-level reproduction.
Examples
facets_positioning_guide()
#> Topic
#> 1 Estimation authority
#> 2 Compatibility purpose
#> 3 External FACETS comparison
#> 4 Current model and calibration boundary
#> 5 Reporting source of truth
#> 6 Extension beyond FACETS
#> Position
#> 1 mfrmr estimates are package-native; FACETS-style names do not mean that FACETS estimated the model.
#> 2 FACETS-style wrappers, table labels, and files support transition, handoff, and report organization, not optimizer-level reproduction.
#> 3 Numerical comparison requires an explicit external FACETS output table supplied by the user.
#> 4 Each fit uses one ordered-categorical response-model family and one observed score scale; nominal and count-response families, mixed families, multiple independent scales, general threshold anchors, and importing FACETS or third-party frozen calibrations are not current public capabilities. Native portable calibration follows the model, estimator and source-check scope in mfrm_calibration_capabilities(), and posterior scoring from an existing fitted object is separate. Positive observation weights act on row-level conditional ordered-rating contributions; they are not a general collapsed-person frequency-table interface and do not change the response family.
#> 5 Inference and reporting should be based on native fit, diagnostics, review, table, and plot-data objects.
#> 6 GPCM, D-study, network, and reusable visualization data are extension routes rather than FACETS menu clones.
#> RecommendedWording
#> 1 The model was estimated with mfrmr; FACETS-style output names are used only to organize the report.
#> 2 FACETS-style outputs were generated for handoff or reader familiarity; they are not evidence of FACETS numerical equivalence.
#> 3 When external FACETS output is supplied, compare MnSq first and report df/ZSTD convention sensitivity separately.
#> 4 Describe mfrmr as a native R RSM/PCM analysis, diagnostic, and reporting environment, not as a general FACETS operational-calibration replacement.
#> 5 Report estimates, standard errors, fit summaries, and plots from documented mfrmr objects.
#> 6 Use package-native extensions as additional evidence and label them as mfrmr analyses.
#> PrimaryRoute
#> 1 fit_mfrm(); diagnose_mfrm(); reporting_checklist()
#> 2 facets_feature_coverage(); run_mfrm_facets(); facets_output_file_bundle()
#> 3 read_facets_fit_table(); facets_fit_review(); fit_measures_table(df_sensitivity = TRUE)
#> 4 fit_mfrm(); facets_feature_coverage(); mfrmr_output_guide()
#> 5 build_summary_table_bundle(); build_visual_summaries(); plot_data()
#> 6 gpcm_capability_matrix(); mfrm_d_study(); mfrm_network_analysis(); plot_data_components()