xmu_savalei_polycor_blocks: Build OpenMx-native polycorrelation blocks for the Savalei...

In umx: Structural Equation Modeling and Twin Modeling in R

Build OpenMx-native polycorrelation blocks for the Savalei sandwich

Description

Assembles the \Delta, \Gamma, and W ingredients for the polycorrelation subsystem of Savalei (2021) \hat{c}_3 (c.hat3), using matrices extracted from the same fitted OpenMx WLS model at mutually consistent scales.

Usage

xmu_savalei_polycor_blocks(
  asymCovAlignedPerObs,
  weightMatAlignedPerObs,
  jacTargetAligned,
  polyNames,
  nVal
)

Arguments

asymCovAlignedPerObs	Aligned per-observation asymptotic covariance (asymCov subset to common moments, before n scaling).
weightMatAlignedPerObs	Aligned per-observation WLS weight (before n scaling).
jacTargetAligned	Full aligned implied Jacobian (all moment rows).
polyNames	Character vector of polycorrelation moment names.
nVal	Sample size n (from model@data@numObs).

Details

Why a separate builder? Savalei's correction is evaluated on polycorrelations only (threshold rows excluded). Each matrix must refer to the same moment ordering (rownames from aligned asymCov) and the same sample size scaling.

Objects returned:

jacPoly (\Delta_{poly})

Rows of model@output$implied_jacobian corresponding to polycorrelation moments (names not ending in t1, t2, etc.). Columns are model parameters in OpenMx order. This is \partial s / \partial \theta for the polycor block.

gammaPoly (\Gamma_{poly})

Asymptotic covariance of the sample polycorrelation moments at sample scale: \Gamma_{poly} = n^2 \times V_{poly,per-obs}, where V_{poly,per-obs} is the polycorrelation block of OpenMx observedStats$asymCov before the global n scaling applied to the full aligned matrices. This is the moment sampling covariance; it is not solve(useWeight).

wPolyWiU (W_{poly} in W_i U)

Polycorrelation block of the WLS weight matrix at sample scale: W_{poly} = n \times W_{poly,per-obs}, from useWeight. This is paired with E^{-1} = (\Delta_{full}' W_{full} \Delta_{full})^{-1} computed from the identically scaled full weight matrix.

Value

List with gammaPoly, wPolyWiU, and jacPoly, or NULL if any block is empty.

See Also

xmu_savalei_scaling_factor(), xmu_WLS_polycor_names()

Other xmu internal not for end user: mx_make(), umxModel(), umxRenameMatrix(), umx_APA_pval(), umx_fun_mean_sd(), umx_get_bracket_addresses(), umx_make(), umx_standardize(), umx_string_to_algebra(), xmuHasSquareBrackets(), xmuLabel_MATRIX_Model(), xmuLabel_Matrix(), xmuLabel_RAM_Model(), xmuMI(), xmuMakeOneHeadedPathsFromPathList(), xmuMakeTwoHeadedPathsFromPathList(), xmuMaxLevels(), xmuMinLevels(), xmuPropagateLabels(), xmuRAM2Ordinal(), xmuTwinSuper_Continuous(), xmuTwinSuper_NoBinary(), xmuTwinUpgradeMeansToCovariateModel(), xmu_CI_merge(), xmu_CI_stash(), xmu_DF_to_mxData_TypeCov(), xmu_PadAndPruneForDefVars(), xmu_WLS_align_jacobian(), xmu_WLS_align_weight(), xmu_WLS_polycor_names(), xmu_ace_de_apply_censor_thresholds(), xmu_ace_de_parse_censor_meta(), xmu_ace_de_parse_censor_rule(), xmu_ace_de_release_binary_v1(), xmu_bracket_address2rclabel(), xmu_catml_df3_diagnostic(), xmu_catml_discrepancy_at_WLS(), xmu_catml_eval_model(), xmu_catml_implied_correlation(), xmu_catml_wls_v(), xmu_cell_is_on(), xmu_check_levels_identical(), xmu_check_needs_means(), xmu_check_variance(), xmu_clean_label(), xmu_compare_constraint_note(), xmu_compare_robust_ML(), xmu_data_missing(), xmu_data_swap_a_block(), xmu_describe_data_WLS(), xmu_dot_make_paths(), xmu_dot_make_residuals(), xmu_dot_maker(), xmu_dot_move_ranks(), xmu_dot_rank_str(), xmu_extract_column(), xmu_get_CI(), xmu_has_summary_mxData(), xmu_invert_matrix(), xmu_is_ordinal_WLS(), xmu_lavaan_process_group(), xmu_make_TwinSuperModel(), xmu_make_bin_cont_pair_data(), xmu_make_mxData(), xmu_match.arg(), xmu_mxData_summary(), xmu_mxRefModels(), xmu_name_from_lavaan_str(), xmu_path2twin(), xmu_path_regex(), xmu_print_algebras(), xmu_ram_de_apply_censor_thresholds(), xmu_ram_de_parse_censor_meta(), xmu_rclabel_2_bracket_address(), xmu_relevel_factors(), xmu_require_summary_mxData(), xmu_robust_ML_fit(), xmu_safe_run_summary(), xmu_savalei_fit_cfi(), xmu_savalei_fit_rmsea(), xmu_savalei_fit_tli(), xmu_savalei_null_scaling_factor(), xmu_savalei_scaling_factor(), xmu_set_sep_from_suffix(), xmu_show_fit_or_comparison(), xmu_simplex_corner(), xmu_standardize_ACE(), xmu_standardize_ACEcov(), xmu_standardize_ACEv(), xmu_standardize_CP(), xmu_standardize_IP(), xmu_standardize_LISREL(), xmu_standardize_RAM(), xmu_standardize_SexLim(), xmu_standardize_Simplex(), xmu_start_value_list(), xmu_starts(), xmu_summary_RAM_group_parameters(), xmu_threshold_id_RAM(), xmu_threshold_id_twin_check(), xmu_topo_sort(), xmu_twin_add_WeightMatrices(), xmu_twin_check(), xmu_twin_get_var_names(), xmu_twin_make_def_means_mats_and_alg(), xmu_twin_upgrade_selDvs2SelVars(), xmu_update_covar(), xmu_wls_display_chi(), xmu_wls_extract_WV()

umx documentation built on Oct. 11, 2026, 5:07 p.m.