Package index • drmSEM

Building a distributional SEM

Assemble a piecewise SEM from drmTMB nodes.

drm_node(): Specify one endogenous node of a distributional SEM
drm_sem(): Fit and assemble a distributional piecewise SEM
drm_psem() summary(<drm_sem>): Assemble a distributional piecewise SEM from fitted drmTMB models

Graph and paths

Inspect the component-labelled structural graph.

paths(): Component-labelled path table for a distributional SEM
standardize(): Standardized component-labelled path coefficients
plot(<drm_sem>): Plot the distributional SEM as a component-labelled DAG

Covariance edges (bivariate)

Declare and inspect residual (rho12) and higher-level (corpair) covariance edges, and declare a bivariate (joint two-response) node. The joint fit that estimates rho12 is the 0.4 engine step; the grammar and accessors are pure-R.

covary(): Declare a covariance edge between two responses
covariances(): Covariance edges of a distributional SEM (residual rho12 and higher-level corpair)
drm_pair(): Declare a bivariate (joint two-response) node
drm_expand_pair(): Expand a bivariate pair onto the covariance-edge grammar
rho12(): Residual response-response correlation (rho12)
corpairs(): Higher-level random-effect correlations (corpairs)

Latent constructs

Build composite (formative) constructs from observed indicators.

drm_composite(): Declare a composite construct from observed indicators
loadings(): Indicator loadings of a SEM's composite constructs

Feedback / cyclic motifs

Declare a feedback motif so specific cycles are allowed (every undeclared cycle stays an error). The equilibrium effect engine and consistent estimation are staged in docs/design/10-cyclic-feedback.md.

drm_cycle(): Declare a feedback (cyclic) motif
cycles(): Declared feedback motifs of a distributional SEM

d-separation

Test missing arrows on any distributional component.

basis_set(): Basis set of independence claims for a distributional SEM
dsep(): Test directed-separation claims by likelihood-ratio refits
fisher_c(): Fisher's C statistic for a fitted distributional SEM

Effects

Simulation-based direct, indirect, and total effects.

direct_effects(): Response-scale direct (controlled) effect of a predictor on a node
indirect_effects(): Indirect effect of a predictor on a node, with a distributional decomposition
total_effects(): Total effect of a predictor on a node by simulation
path_effects(): Path-specific (per-mediator) decomposition of an indirect effect
plot(<drm_effect>): Plot an effect decomposition as a forest plot

Symbolizer integration

Render per-node equations from fitted drmSEM objects.

symbolize(<drm_sem>): Symbolize a distributional piecewise SEM
as_latex(<symbolized_drm_sem>) equations(<symbolized_drm_sem>) assumption_table(<symbolized_drm_sem>): Render the symbolized SEM as LaTeX, equations, or an assumption table

Model comparison

Compare competing causal DAGs by Fisher’s C / CBIC.

drm_dag(): Specify a candidate causal model (an unfitted DAG)
drm_model_set(): Assemble a set of candidate causal models
compare(): Fit and rank a set of candidate causal models
best(): Best-supported candidate model
average(): Criterion-weighted average of standardized path coefficients

Interop (graph interchange)

Translate a drmSEM graph to/from neighbouring ecosystems’ text formats (lavaan model syntax, Graphviz DOT). Graph interchange only — drmSEM never fits its own likelihoods, and non-mean (distributional-component) paths are dropped-with-notice when emitting lavaan syntax, never misrepresented as a mean regression.

as_lavaan(): Export a distributional SEM as lavaan model syntax
from_lavaan(): Parse lavaan model syntax into a drmSEM graph skeleton
as_dot(): Export a distributional SEM as Graphviz DOT

Phylogenetic SEM

Build evolutionary covariance matrices for phylogenetic nodes.

drm_phylo_cov(): Build an evolutionary phylogenetic covariance / relatedness matrix

Diagnostics

check_sem(): Diagnose a fitted distributional SEM