Abstract:We provide a framework for motivating and diagnosing the functional form in the structural part of nonlinear or linear structural equation models when the measurement model is a correctly specified linear confirmatory factor model. A mathematical population-based analysis provides asymptotic identification results for conditional expectations of a coordinate of an endogenous latent variable given exogenous and possibly other endogenous latent variables, and theoretically well-founded estimates of this conditional expectation are suggested. Simulation studies show that these estimators behave well compared to presently available alternatives. Practically, we recommend the estimator using Bartlett factor scores as input to classical non-parametric regression methods.

What problem does this paper attempt to address?

This paper attempts to solve the problem of how to non - parametrically estimate the form of relationships between endogenous latent variable coordinates in Structural Equation Models (SEMs). Specifically, when the measurement model is a correctly - specified linear confirmatory factor model, the author provides a framework to motivate and diagnose the functional form of the structural part in nonlinear or linear structural equation models. Through mathematical population analysis, the author provides asymptotic identification results of conditional expectations and proposes theoretically sound conditional expectation estimation methods. These estimators perform well in simulation studies and have advantages over currently available alternatives. The author recommends using Bartlett factor scores as inputs for classical non - parametric regression methods. In short, the main problems of the paper are: - How to non - parametrically estimate the form of relationships between endogenous latent variables in structural equation models? - Provide a theoretical basis for non - parametric estimation methods based on Bartlett factor scores to motivate and diagnose the functional form in structural equation models. The key contributions of the paper include: 1. Establish the conditions for identifying the conditional expectation function \( H \) using factor scores at the population level. 2. Prove some basic results of affine factor scores applicable to such analysis. 3. Present new asymptotic results, including the consistency of Bartlett scores in the mean - square sense, the normality of factor scores as measurement errors of factor estimates, and the conditions for the conditional expectation based on factor scores with reduced measurement errors to converge to the conditional expectation based on factors. 4. Suggest non - parametric methods based on Bartlett factor scores and evaluate the performance of these methods and those proposed by Kelava et al. (2017) through simulation studies. These contributions are helpful for more effective non - parametric estimation and model diagnosis in structural equation models.

Non-parametric Regression Among Factor Scores: Motivation and Diagnostics for Nonlinear Structural Equation Models

Optimal Estimation of Large-Dimensional Nonlinear Factor Models

Score predictor factor analysis: Reproducing observed covariances by means of factor score predictors

Factor models and variable selection in high-dimensional regression analysis

Path and Direction Discovery in Individual Dynamic Factor Models: A Regularized Hybrid Unified Structural Equation Modeling with Latent Variable

Multilevel Analysis of Mediation, Moderation, and Nonlinear Effects in Small Samples, Using Expected a Posteriori Estimates of Factor Scores

Factor modelling for high-dimensional functional time series

TESTING FOR STRUCTURAL CHANGES IN FACTOR MODELS VIA A NONPARAMETRIC REGRESSION

Factor-guided estimation of large covariance matrix function with conditional functional sparsity

Identification of the linear factor model

The Effect of Latent and Error Non-Normality on Measures of Fit in Structural Equation Modeling

Robust Nearly-Efficient Estimation of Large Panels with Factor Structures

A two-stage estimation procedure for non-linear structural equation models

Treating reflective indicators as causal-formative indicators in order to compute factor score estimates or unit-weighted scales

Score Predictor Factor Analysis as model for the identification of single-item indicators

A Simulation Study of Polychoric Instrumental Variable Estimation in Structural Equation Models

Consistent Partial Least Squares for Nonlinear Structural Equation Models

Covariance Function Estimation for High-Dimensional Functional Time Series with Dual Factor Structures

Informational Content of Factor Structures in Simultaneous Binary Response Models

A Schmid-Leiman based transformation resulting in perfect inter-correlations of three types of factor score predictors

The Kernel Trick for Nonlinear Factor Modeling