Estimates and evaluates latent factor scores for scales

Usage

elfs(
  data,
  f1_cols,
  f2_cols = NULL,
  f3_cols = NULL,
  f4_cols = NULL,
  f5_cols = NULL,
  f6_cols = NULL,
  fg_cols = NULL,
  f1_name = NULL,
  f2_name = NULL,
  f3_name = NULL,
  f4_name = NULL,
  f5_name = NULL,
  f6_name = NULL,
  fg_name = NULL,
  ordered = FALSE,
  missing = "listwise",
  dynamic = FALSE,
  meas_invar = NULL,
  modify = NULL,
  lfs_method = NULL,
  lfs_transform = TRUE,
  chrome_bypass = FALSE
)

Arguments

data: data.frame object
f1_cols, f2_cols, f3_cols, f4_cols, f5_cols, f6_cols: Character vector(s) listing column names included in (each) latent factor. Compatible with dplyr::starts_with() code. Only f1_cols is required.
fg_cols: Character vector listing column names included in additionally-specified "general factor" (i.e., bifactor models).
f1_name, f2_name, f3_name, f4_name, f5_name, f6_name: Optional names (character) for each specified latent factor.
fg_name: Optional name (character) for "general factor" (see argument fg_cols, defaults to "FactorG").
ordered: Whether to treat measured variables as ordinal (vs. continuous; see lavaan::sem()). FALSE by default.
missing: How to treat missing data (listwise deletion by default, see lavaan::lavOptions() for alternatives)
dynamic: Whether to estimate dynamic fit indices (McNeish & Wolf, 2023). To use, download the latest development version from Github using pak::pkg_install("dynamic") or devtools::install_github("melissagwolf/dynamic"). FALSE by default.
meas_invar: Character indicating column name to split-by when conducting measurement invariance across a categorical variable. NULL by default, assuming no measurement invariance analysis.
modify: Additional character or character vector to be attached to the SEM specification prior to model fitting (e.g., specifying correlated residual variances)
lfs_method: Character string indicating method for estimating latent factor scores (for details and default information, see lavaan::lavPredict())
lfs_transform: Whether to transform extracted factor scores to match model-implied mean and variance-covariance (for details, see lavaan::lavPredict()). TRUE by default, following best practice when latent factor scores are used in subsequent regression analysis.
chrome_bypass: Whether to bypass chrome-screenshot method for displaying results from running elfs(). FALSE by default.