Type:	Package
Title:	Post-Estimation Utilities for 'lavaan' Fitted Models
Version:	0.3.4
Description:	Companion toolbox for structural equation models fitted with 'lavaan'. Provides post-estimation diagnostics and graphics that operate directly on a fitted object using its estimates and covariance, and refits auxiliary models when needed. The package relies on 'lavaan' (Rosseel, 2012) <doi:10.18637/jss.v048.i02>.
URL:	https://github.com/g-corbelli/lavinteract
BugReports:	https://github.com/g-corbelli/lavinteract/issues
License:	GPL-3
Encoding:	UTF-8
Imports:	lavaan, rlang, ggplot2, stats
Suggests:	testthat (≥ 3.0.0), knitr, rmarkdown
Language:	en-US
NeedsCompilation:	no
Maintainer:	Giuseppe Corbelli <giuseppe.corbelli@uninettunouniversity.net>
Config/testthat/edition:	3
RoxygenNote:	7.3.2
Packaged:	2025-11-13 11:09:18 UTC; giuse
Author:	Giuseppe Corbelli [aut, cre]
Repository:	CRAN
Date/Publication:	2025-11-13 17:40:10 UTC

Post-Estimation Utilities for 'lavaan' Fitted Models

Description

Companion toolbox for structural equation models fitted with 'lavaan'. Operates directly on a fitted object using its estimates and covariance. Refits auxiliary models when needed to compute estimates, diagnostics, and plots.

Details

The functions are:

• lav_slopes: simple slopes and interaction plots from a fitted 'lavaan' model.

• lav_vif: variance inflation factors for structural predictors with measurement preserved.

• lav_cv: repeated holdout (Monte Carlo) cross-validation of R^2 for SEM outcomes.

Note

The development of this package grew from ongoing discussions and interactions (sic) with colleagues, in particular Dr. Cataldo Giuliano Gemmano, whose steady feedback and support helped shape it.

Author(s)

Giuseppe Corbelli (<giuseppe.corbelli@uninettunouniversity.net>)

See Also

Useful links:

• https://github.com/g-corbelli/lavinteract

• Report bugs at https://github.com/g-corbelli/lavinteract/issues

Repeated holdout (Monte Carlo) cross-validation of R^2 for structural equation models ('lavaan' objects)

Description

Estimate out-of-sample predictive performance for structural relations in a fitted 'lavaan' model using repeated holdout (Monte Carlo cross-validation, leave-group-out CV). At each repetition, the model is refitted on a random training subset and evaluated on a disjoint test subset.

Usage

lav_cv(
  fit,
  data = NULL,
  times = "auto",
  train_prop = 0.8,
  seed = 42L,
  quiet = TRUE,
  digits = 3L,
  plot = TRUE,
  tol = 0.001,
  window = 50L,
  max_times = 3000L,
  min_r2_for_pct = 0.05
)

## S3 method for class 'lav_cv'
print(x, digits = x$digits %||% 3L, ...)

## S3 method for class 'lav_cv'
summary(object, ...)

Arguments

Details

For observed outcomes, R^2 is computed by comparing test-set observed values with predictions obtained by applying the training-set structural coefficients to the test-set predictors.

For latent outcomes, the outcome is not directly observed in the test set. Factor scores for the outcome are first computed in the test set using the measurement model learned on the training set; these scores serve as the outcome values. Predictions are then formed by applying the training-set structural coefficients to the test-set predictors (including factor scores for any latent predictors). R^2 is computed by comparing the test-set factor scores of the outcome with these predicted scores.

The in-sample baseline R^2 is computed on the full dataset using the same metric as in cross-validation: observed outcomes use observed-versus-predicted R^2; latent outcomes use score-versus-predicted-score R^2.

By default, repetitions continue until the running mean R^2 for each outcome stabilizes within a specified tolerance over a trailing window of successful splits, or until a maximum number of splits is reached.

The summary table reports the in-sample baseline R^2, the median cross-validated R^2, its standard deviation, and the percent drop (baseline vs. median CV) with heuristic threshold markers. The percent drop is suppressed when the in-sample R^2 is very small.

Value

A list with class 'lav_cv' and elements:

table

Data frame with columns: outcome, type ("observed" or "latent"), r2_in, r2_cv_mean, r2_cv_median, r2_cv_sd, drop_mean_pct, drop_med_pct, splits_used.

split_matrix

Matrix of split-wise test-set R^2 values (rows = splits, columns = outcomes).

times

Character or integer indicating the number of splits used (e.g., "auto(534)" or 500).

train_prop

Numeric. Training proportion used in each split.

N

Integer. Number of rows in the input data.

seed

Integer. Random seed used to generate the splits.

tol

Numeric. Tolerance used by the auto-stop rule.

window

Integer. Trailing window size for the auto-stop rule.

min_r2_for_pct

Numeric. Minimum in-sample R^2 required to compute percent drop.

call

match.call() of the function call.

digits

Integer. Default number of digits for printing.

References

Cudeck, R., & Browne, M. W. (1983). Cross-Validation Of Covariance Structures. Multivariate Behavioral Research, 18(2), 147-167. doi:10.1207/s15327906mbr1802_2

Hastie, T., Friedman, J., & Tibshirani, R. (2001). The Elements of Statistical Learning. In Springer Series in Statistics. Springer New York. doi:10.1007/978-0-387-21606-5

Kvalseth, T. O. (1985). Cautionary Note about R2. The American Statistician, 39(4), 279-285. doi:10.1080/00031305.1985.10479448

Shmueli, G. (2010). To Explain or to Predict? Statistical Science, 25(3). doi:10.1214/10-sts330

Yarkoni, T., & Westfall, J. (2017). Choosing Prediction Over Explanation in Psychology: Lessons From Machine Learning. Perspectives on Psychological Science, 12(6), 1100-1122. doi:10.1177/1745691617693393

See Also

sem, lavPredict, inspect

Examples

library("lavaan")
model <- "
ind60 =~ x1 + x2 + x3
dem60 =~ y1 + y2 + y3 + y4
dem65 =~ y5 + y6 + y7 + y8

dem60 ~ ind60
dem65 ~ ind60 + dem60

y1 ~~ y5
y2 ~~ y6
"
fit <- lavaan::sem(
model = model, 
data = lavaan::PoliticalDemocracy,
std.lv = TRUE, 
estimator = "MLR", 
meanstructure = TRUE)
  
result <- lav_cv(
fit = fit, 
data = lavaan::PoliticalDemocracy, 
times = 5)
print(result)

Simple slopes and interaction plots for fitted 'lavaan' models

Description

Computes conditional (simple) slopes of a focal predictor across values of a moderator from a fitted 'lavaan' model that includes their explicit product term. Plots predicted lines with Wald confidence ribbons, and print an APA-style test of the interaction for easy reporting and interpretation, plus a simple-slopes table.

Usage

lav_slopes(
  fit,
  outcome,
  pred,
  modx,
  interaction,
  data = NULL,
  modx.values = NULL,
  modx.labels = NULL,
  pred.range = NULL,
  conf.level = 0.95,
  x.label = NULL,
  y.label = NULL,
  legend.title = NULL,
  colors = NULL,
  line.size = 0.80,
  alpha = 0.20,
  table = TRUE,
  digits = 2,
  modx_n_unique_cutoff = 4L,
  return_data = FALSE
)

## S3 method for class 'lav_slopes'
print(x, ...)

## S3 method for class 'lav_slopes'
summary(object, ...)

Arguments

Details

The model should include a main effect for the predictor, a main effect for the moderator, and their product term. The simple slope of the predictor at a given moderator value combines the predictor main effect with the interaction term. The moderator can be continuous or categorical. Standard errors use the delta method with the model covariance matrix of the estimates.

Value

A list with elements:

plot

ggplot object with lines and confidence ribbons.

slope_table

Data frame with moderator levels, simple slopes, SE, z, and CI.

plot_data

Only when return_data = TRUE: data used to build the plot.

Notes

Estimates are unstandardized; a standardized beta for the interaction is also reported for reference. Wald tests assume large-sample normality of estimates.

Examples

set.seed(42)
X <- rnorm(100); Z <- rnorm(100); X_Z <- X*Z
Y <- 0.6*X + 0.6*Z + 0.3*X_Z + rnorm(100, sd = 0.7) 
dataset <- data.frame(Y, X, Z, X_Z)
fit <- lavaan::sem("Y ~ X + Z + X_Z", data = dataset)
lav_slopes(
fit = fit, 
data = dataset,
outcome = "Y", 
pred = "X", 
modx = "Z", 
interaction = "X_Z")

Variance Inflation Factors for 'lavaan' Structural Predictors

Description

Compute VIF for each predictor that appears in structural regressions with two or more predictors, refitting the necessary sub-models so that latent predictors are handled at the latent level (i.e., with their original measurement models). It returns also the R^2 of each eligible endogenous variable from the original fit for context.

Usage

lav_vif(
  fit,
  data = NULL,
  quiet = TRUE
)

## S3 method for class 'lav_vif'
print(x, digits = 3, cutoff = c(5, 10), ...)

## S3 method for class 'lav_vif'
summary(object, ...)

Arguments

Details

Each auxiliary refitted model:

• includes the original measurement model for any latent predictors;

• includes any residual covariances among those indicators that were specified in the original model;

• regresses the focal predictor on the remaining predictors at the latent level when applicable.

VIF_i = 1 / (1 - R^2_i) generalizes VIF to SEM while respecting measurement models.

The function reuses the estimator, missing-data handling, and several options from fit.

Value

A list with:

• vif_table: data.frame with columns outcome, predictor, group, r2_predictor, vif, k_predictors.

• outcome_r2: data.frame with R^2 per eligible endogenous outcome and group from the original fit.

Examples

set.seed(42)
x1 <- rnorm(100); x2 <- 0.85*x1 + rnorm(100, sd = sqrt(1 - 0.85^2)); x3 <- rnorm(100)
y  <- 0.5*x1 + 0.3*x2 + 0.1*x3 + rnorm(100, sd = 0.7)
dataset <- data.frame(y, x1, x2, x3)
fit <- lavaan::sem("y ~ x1 + x2 + x3", data = dataset)
lav_vif(
fit = fit,
data = dataset)