Compare Model Fit Indices Across Multiple Models

compare_model_fit() compares the fit indices of two or more models. It extracts the fit indices using model_fit() and combines them into a single data frame for easy comparison. It supports any fitted lavaan model objects (including CFA, SEM, growth, multigroup, and multilevel variants).

Usage

compare_model_fit(
  ...,
  type = NULL,
  metrics = "essential",
  verbose = TRUE,
  test = "default",
  standard_test = FALSE,
  test_details = FALSE
)

Arguments

...: Two or more model objects to be compared. For lavaan, pass any fitted objects created with lavaan::cfa(), lavaan::sem(), lavaan::growth(), or related functions.
type: A character string specifying the type of fit indices to extract. Options are "standard", "scaled", and "robust". Defaults to NULL, which automatically selects "scaled" if a robust estimator is used, otherwise "standard".
metrics: A character vector specifying which fit indices to extract. Defaults to "essential", or a custom vector of indices.
verbose: Logical. If TRUE, prints messages about the indices being adjusted.
test: A character string or vector specifying which lavaan tests to report. Use "default" (the default) to return all non-standard tests from lavInspect(fit, "options")$test, excluding "standard", "default", and "none". Provide a character vector to request specific tests; unavailable entries are dropped with an informational message when verbose = TRUE. When supplying a list, it must be named and each name must match the model labels shown in the MODEL column. If you supply named arguments, you may also reference the original object names when they are unambiguous.
standard_test: A logical value indicating whether to include a standard-test row in addition to non-standard tests. When TRUE, the standard row is shown first and always uses standard indices. When supplying a list, it must be named and each name must match the model labels shown in the MODEL column. If you supply named arguments, you may also reference the original object names when they are unambiguous.
test_details: Logical. If TRUE, include TEST and SE columns (for lavaan fits) that describe the test and standard error settings used to compute each row. When standard_test = TRUE adds the standard row, SE is set to NA for that row because fit indices do not depend on standard errors.

Value

A data frame containing the fit indices for each model, with an additional column identifying the models.

Examples

if (requireNamespace("lavaan", quietly = TRUE)) {
  library(lavaan)
  library(psymetrics)
  hs_model <- 'visual  =~ x1 + x2 + x3
               textual =~ x4 + x5 + x6
               speed   =~ x7 + x8 + x9'
  fit1 <- cfa(hs_model, data = HolzingerSwineford1939, estimator = "ML")
  fit2 <- cfa(hs_model, data = HolzingerSwineford1939, estimator = "MLR")
  compare_model_fit(fit1, fit2)
  sem_model <- 'visual  =~ x1 + x2 + x3
                textual =~ x4 + x5 + x6
                speed   =~ x7 + x8 + x9
                textual ~ visual
                speed ~ textual'
  sem1 <- sem(sem_model, data = HolzingerSwineford1939, estimator = "ML")
  sem2 <- sem(sem_model, data = HolzingerSwineford1939, estimator = "MLR")
  compare_model_fit(sem1, sem2)
  compare_model_fit(fit1, fit2, standard_test = TRUE, test_details = TRUE)
  compare_model_fit(fit1, fit2, metrics = c("cfi", "tli", "rmsea"))
  fit3 <- cfa(
    hs_model,
    data = HolzingerSwineford1939,
    test = c("satorra.bentler", "mean.var.adjusted")
  )
  compare_model_fit(fit1, fit3, test = "satorra.bentler")
  compare_model_fit(
    fit1 = fit1,
    robust = fit3,
    test = list(robust = "mean.var.adjusted"),
    standard_test = list(robust = TRUE),
    test_details = TRUE
  )
} else {
  message("Please install 'lavaan' to run this example.")
}
#> Standard-test row uses standard indices for estimator ML. Affected model: fit2.
#> cfi, tli, and rmsea were adjusted to their scaled version.
#> If you want to control the specific metric type used, specify it explicitly
#> (e.g., `cfi.robust`) or modify the type argument.
#> Requested tests not found in the fit1 model and were dropped: satorra.bentler.
#> Standard-test row uses standard indices for estimator ML. Affected model:
#> robust.
#> MODEL  | NOBS | ESTIMATOR  |       TEST        |    SE    | NPAR | Chi2 
#> ------------------------------------------------------------------------
#> fit1   | 301  |     ML     |     standard      | standard |  21  | 85.31
#> robust | 301  |     ML     |     standard      |          |  21  | 85.31
#> robust | 301  | ML_variant | mean.var.adjusted | standard |  21  | 67.23
#> 
#> MODEL  | Chi2_df | p (Chi2) |  CFI  |  TLI  | RMSEA |   RMSEA  CI    | SRMR 
#> ----------------------------------------------------------------------------
#> fit1   |  24.00  |  < .001  | 0.931 | 0.896 | 0.092 | [0.071, 0.114] | 0.065
#> robust |  24.00  |  < .001  | 0.931 | 0.896 | 0.092 | [0.071, 0.114] | 0.065
#> robust |  19.95  |  < .001  | 0.734 | 0.887 | 0.089 | [0.068, 0.110] | 0.065
#>

Usage

Arguments

Value

See also

Examples