Fits bifactor exploratory structural equation models (B-ESEM),
together with standard exploratory structural equation modeling (ESEM)
and confirmatory factor analysis (CFA), for continuous and
ordinal data. Continuous models use 'lavaan' native efa()
blocks with robust maximum likelihood (MLR) estimation.
Ordinal ESEM defaults to the 'lavaan' weighted least squares
mean- and variance-adjusted (WLSMV) estimator; ordinal B-ESEM uses a
custom diagonally weighted least squares (DWLS) path with polychoric
correlations from 'psych', rotation-delta standard errors via 'numDeriv',
and a mean- and variance-adjusted chi-square. Target, geomin, and oblimin
rotations use 'GPArotation'; the bifactor ESEM approach follows Morin,
Arens and Marsh (2016)