Nonlinear Structural Equation Models Research Articles

Causal network learning with non-invertible functional relationships

Discovery of causal relationships from observational data is an important problem in many areas. Several recent results have established the identifiability of causal directed acyclic graphs (DAGs) with non-Gaussian and/or nonlinear structural equation models (SEMs). Focusing on nonlinear SEMs defined by non-invertible functions, which exist in many data domains, a novel test is proposed for non-invertible bivariate causal models. Algorithms are further developed to incorporate this test in structure learning of DAGs that contain both linear and nonlinear causal relations. Extensive numerical comparisons show that the proposed algorithms outperform existing DAG learning methods in identifying causal graphical structures. The practical application of the methods is illustrated by learning causal networks for combinatorial binding of transcription factors from ChIP-Seq data.

The use of items and item parcels in nonlinear structural equation models

Nonlinear structural equation models within the frequentist framework were developed to work with continuous items. Applied researchers who usually work with Likert-type items choose between two strategies to estimate such models: treat items as continuous variables or create item parcels. Two Monte Carlo studies were conducted to evaluate the effects of each strategy on estimates and Type I errors for models with interaction and quadratic effects estimated using LMS. The first study evaluated the effect of asymmetry type and item quantity. The second assessed the use of item parcels and parcel configuration under equivalent conditions. Results reveal that treating items as continuous variables is not problematic when item categories are symmetrical or have opposite-direction asymmetries; however, meaningful parameter bias and increased Type I errors are produced in the case of same-direction asymmetry. Use of parcels does not overcome these problems. The results are discussed to provide recommendations for applied researchers.

A Marginal Maximum Likelihood Approach for Extended Quadratic Structural Equation Modeling with Ordinal Data

ABSTRACT The literature on non-linear structural equation modeling is plentiful. Despite this fact, few studies consider interactions between exogenous and endogenous latent variables. Further, it is well known that treating ordinal data as continuous produces bias, a problem which is enhanced when non-linear relationships between latent variables are incorporated. A marginal maximum likelihood-based approach is proposed in order to fit a non-linear structural equation model including interactions between exogenous and endogenous latent variables in the presence of ordinal data. In this approach, the exact gradient of the approximated observed log-likelihood is calculated in order to attain the approximated maximum likelihood estimator. A simulation study shows that the proposed method provides estimates with low bias and accurate coverage probabilities.

The moderating role of leisure participation in work–leisure conflict for the reduction of burnout in hotels and travel agencies

AbstractThis study investigates the influence of work–leisure conflict (WLC) on the performance of employees in Egyptian hotels and travel agencies through the mediating role of employee burnout related to emotional exhaustion, cynicism and professional efficacy. In addition, this study examines how leisure participation moderates the effects of WLC on both employee burnout and performance. Data used for this study were collected from 788 hotel and travel agencies operating in Egypt. Given the purpose of the study, a nonlinear partial least squares structural equation modelling (PLS‐SEM) was used to analyse the data. The findings support the proposed model and thus indicated that WLC has a negative influence on the performance of employees. Moreover, it has also been determined that leisure participation plays an important role in decreasing the negative influence of WLC on emotional exhaustion, cynicism and reduced professional efficacy of employees in hotels and travel agencies. These findings hold important implications for tourism and hospitality firms.

Specification issues in nonlinear SEM: The moderation that wasn't.

Analysis of interaction or moderation effects between latent variables is a common requirement in the social sciences. However, when predictors are correlated, interaction and quadratic effects become more alike, making them difficult to distinguish. As a result, when data are drawn from a quadratic population model and the analysis model specifies interactions only, misleading results may be obtained. This article addresses the consequences of different types of specification error in nonlinear structural equation models using a Monte Carlo study. Results show that fitting a model with interactions when quadratic effects are present in the population will almost certainly lead to erroneous detection of moderation effects, and that the same is true in the opposite scenario. Simultaneous estimation of interactions and quadratic effects yields correct results. Simultaneous estimation of interaction and quadratic effects prevents detection of spurious or misleading nonlinear effects. Results are discussed and recommendations are offered to applied researchers.

Estimation of nonlinear structural equation models with dichotomous indicator variables: a Monte Carlo comparison of methods

Nonlinear structural equation models (SEMs), which include interactions among latent predictors, as well as quadratic or higher order terms, have been the focus of research over the last three decades, beginning with Kenny and Judd (1984). The great majority of that work has focused on the case where the indicator variables are continuous in nature. However, in practice many nonlinear SEMs will involve the use of responses to items on scales, which are categorical. The focus of the current simulation study was on comparing several methods for modelling nonlinear SEMs when indicator variables were dichotomous. Results of the study showed that a Bayesian approach, as well as a method based on 2-stage least squares, provided the most accurate parameter estimates, the highest power, and the best control over the Type I error rate for the interaction effect. Implications of these findings for practice are discussed.

Estimation of nonlinear structural equation models with dichotomous indicator variables: a Monte Carlo comparison of methods

Nonlinear structural equation models (SEMs), which include interactions among latent predictors, as well as quadratic or higher order terms, have been the focus of research over the last three decades, beginning with Kenny and Judd (1984). The great majority of that work has focused on the case where the indicator variables are continuous in nature. However, in practice many nonlinear SEMs will involve the use of responses to items on scales, which are categorical. The focus of the current simulation study was on comparing several methods for modelling nonlinear SEMs when indicator variables were dichotomous. Results of the study showed that a Bayesian approach, as well as a method based on 2-stage least squares, provided the most accurate parameter estimates, the highest power, and the best control over the Type I error rate for the interaction effect. Implications of these findings for practice are discussed.

THE EFFECT OF SELF-LEADERSHIP AND PHYSICAL VITALITY ON LIFE AND WORK SATISFACTION

What role does self-leadership play for individuals’ physical vitality and subjective well-being? The authors examined the data from 177 job incumbents and 165 students. Nonlinear structural equation modeling revealed a latent moderator effect of self-leadership on physical vitality and well-being. For individuals with low self-leadership, more physical vitality led to increased well-being, while for those with high self-leadership, no effect was detected. Understanding the role of high vs. low selfleadership in current and future physical vitality and well-being adds insight as to how health-related interventions should be structured to optimize well-being and accompanying positive working behavior

Uncovering the roles of retail brand experience and brand love in the apparel industry: Non-linear structural equation modelling approach

providing a pleasant shopping experience for consumers is vital in the apparel retail industry. A comprehensive multi-dimensional scale to measure “retail brand experience” (RBE) was recently introduced; however, this scale lacks application. Hence, the purpose of this study is to examine a conceptual model that shows the associations between RBE and customer behavioural outcomes (willingness to pay more, word of mouth communication, and continuous purchase intention) in the retail apparel industry using the multi-dimensional scale operationalisation of RBE. The present study also attempts to uncover the moderating role of brand love in the proposed model. Using the survey method, 416 usable responses were collected. The findings indicate that all the direct effects are statistically supported. The study also shows non-linear relationships between word of mouth communication and willingness to pay more, and word of mouth communication and continuous purchase intention. With respect to the moderating role of brand love, it enhances the influence of both RBE and word of mouth communication on willingness to pay more. Overall, the present study contributes to the retailing literature, particularly in the apparel-retailing context, branding literature, and customer behaviour literature. Practically, this study provides several recommendations to enhance customer willingness to pay more, their continuous purchase intentions, and word of mouth communication by providing better retail brand experience.

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

Applications of structural equation models (SEMs) are often restricted to linear associations between variables. Maximum likelihood (ML) estimation in non-linear models may be complex and require numerical integration. Furthermore, ML inference is sensitive to distributional assumptions. In this article, we introduce a simple two-stage estimation technique for estimation of non-linear associations between latent variables. Here both steps are based on fitting linear SEMs: first a linear model is fitted to data on the latent predictor and terms describing the non-linear effect are predicted by their conditional means. In the second step, the predictions are included in a linear model for the latent outcome variable. We show that this procedure is consistent and identifies its asymptotic distribution. We also illustrate how this framework easily allows the association between latent variables to be modeled using restricted cubic splines, and we develop a modified estimator which is robust to non-normality of the latent predictor. In a simulation study, we compare the proposed method to MLE and alternative two-stage estimation techniques.

A Nonlinear Dynamic Latent Class Structural Equation Model

In this article, we propose a nonlinear dynamic latent class structural equation modeling (NDLC-SEM). It can be used to examine intra-individual processes of observed or latent variables. These processes are decomposed into parts which include individual- and time-specific components. Unobserved heterogeneity of the intra-individual processes are modeled via a latent Markov process that can be predicted by individual- and time-specific variables as random effects. We discuss examples of sub-models which are special cases of the more general NDLC-SEM framework. Furthermore, we provide empirical examples and illustrate how to estimate this model in a Bayesian framework. Finally, we discuss essential properties of the proposed framework, give recommendations for applications, and highlight some general problems in the estimation of parameters in comprehensive frameworks for intensive longitudinal data.

A Multifaceted Approach regarding the Association of the DsbA-L Gene with the Risk of Obesity-related Diseases Based on Clinical Pharmacogenetics

Adiponectin, the most abundant adipose tissue-derived adipocytokine, improves insulin sensitivity and has anti-inflammatory properties. Disulfide-bond A oxidoreductase-like protein (DsbA-L) is a key molecule in the multimerization of adiponectin (i.e., activation of adiponectin). In mice, liver-specific knockout of the Dsba-L gene impaired the mitochondrial function in hepatocytes and exacerbated the high-fat-diet-induced fatty liver. In addition, the DsbA-L mRNA level is negatively correlated with body mass index (BMI) in humans. We recently investigated the clinical impact of the DsbA-L gene on lifestyle-related diseases in Japanese subjects. We confirmed the influence of the common DsbA-L rs1917760 polymorphism on the multimerization of adiponectin, as well as the association of the polymorphism with the risk of obesity and non-alcoholic fatty liver disease, using prediction models based on a non-linear mixed effect model and/or structural equation models among elderly participants in a health screening program. We also observed a decreasing effect of DsbA-L polymorphism on the DsbA-L mRNA level in peripheral blood mononuclear cells, and an increasing effect of the polymorphism on the prevalence of excessive weight among schizophrenia patients at a high risk for obesity. These findings suggest that DsbA-L may be a key molecule associated with the development and progression of obesity and its related diseases. Therefore, genotyping the DsbA-L polymorphism and identifying patients at a high risk of developing obesity may help prevent obesity and its complications by facilitating targeted prevention and treatment programs for high-risk individuals.

Statistical inference for nonignorable missing-data problems: a selective review

ABSTRACTNonignorable missing data are frequently encountered in various settings, such as economics, sociology and biomedicine. We review statistical inference for nonignorable missing-data problems, including estimation, influence analysis and model selection. For estimation of mean functionals, we review semiparametric method and empirical likelihood (EL) approach. For estimation of parameters in exponential family nonlinear structural equation models, we introduce expectation-maximisation algorithm, Bayesian approach, and Bayesian EL method. For influence analysis, we investigate the case-deletion method and local influence analysis method from the frequentist and Bayesian viewpoints. For model selection, we present the modified Akaike information criterion and penalised method.

Study on the Optimization Modeling of Construction Time Considering the Dependence of Durations

In order to solve the problem of project time optimization considering the dependence of durations, and on the basis of the existing literature the Pearson product-moment correlation coefficient and the nonlinear structural equation model are adopted to measure serial duration dependence and parallel dependence separately in the construction projects, and which are used to update the duration data. The mathematical model of the optimization of construction time is set up while the cost as a constraint, at the same time the optimization model is modified by considering the duration dependence, and finally the time optimization is achieved by using genetic algorithm. By taking into account the optimization modeling of construction time considering the dependence of durations, an empirical study is carried out to get the ideal results. This method, which has good guidance and practicability, can be used to solve the problem of time delay in the actual project.

Study on the Method of Dependence Measure of Durations in Construction Schedule Network

Previous research of construction schedule is based on the classic model of network planning technique as the background, assume the duration of each process were independent of each other in the schedule network planning. In the process of the practical project, each duration of process is not independent of each other, but acting on each other. In this paper, through research, based on a series of assumptions, we used Product-Moment Correlation Coefficient to measure serial dependence between duration variables. We also propose to regard parallel duration dependency, the common influence factors and organization management as potential variables, establishing the nonlinear structural equation model to measure parallel dependency among duration variables. Finally, this paper applies dependence measure and duration forecast considered dependence to a practical project progress network, results show that the method has good applicability and scientificity.

The influence of external factors on children's travel mode: A comparison of school trips and non-school trips

This study examined travel patterns of pupils from four secondary schools in Austria and Germany. Their mobility behavior was examined using a one-week travel diary in a typical school week. This paper examines objective determinants (in particular settlement pattern and trip characteristics) of mode choice. We used a Bayesian approach for nonlinear Structural Equation Modeling (SEM) of binary response variables to assess the effects of external factors on the choice of travel modes. The focus lies on the competitiveness between car and transit. The results indicate that children's modal choices are influenced by trip length and the service quality of motorized modes. A key finding is that school trips and non-school trips are very different. School trips are quite affine to transit even in rural areas, given a sufficient service quality, which can easily be provided by a school bus system. Long school trips increase the frequency of transit use. Non-school trips, however, are much more affine to car ridership, if trip length exceeds the range for walking and cycling.

Investigating the moderating role of Export Promotion Programmes using evidence from North-Africa

PurposeBased on an institutional approach to explaining firms’ internationalisation, this paper aims to empirically investigate the role of Export Promotion Programmes (EPPs) in moderating the influence of export barriers perceptions on small and medium enterprises’ (SMEs) propensity to export.Design/methodology/approachThe study uses evidence from Algeria, the largest North-African country. The data were collected using an online questionnaire, targeting SMEs operating in the manufacturing sector. The study considers the influence of procedural, informational, environmental and functional barriers on export propensity, to uncover the moderating role of trade missions, trade shows and export seminars and workshops on such relationships. To examine these links, five main hypotheses are proposed and tested through a non-linear partial least squares structural equation modelling on a sample of 128 Algerian SMEs.FindingsThe results show that while internal barriers decrease firms’ export propensity, EPPs including trade fairs and shows may independently pose either a positive or negative influence on such relationships.Research limitations/implicationsThe study confirms the applicability of the institutional perspective to explaining firms’ internationalisation. More importantly, the present study highlights the role of EPPs in moderating the influence of export barriers perceptions on SMEs’ international market entry, a role neglected by the extant empirical literature.Practical implicationsThe current findings hold important implications to export promotion organisations operating in African countries. Notably, the results reveal that some programmes could have a negative influence if they are not delivered appropriately.Originality/valueThis study offers a rare focus on the moderating role of EPPs in the relationship between export barriers and export propensity, within the setting of a North-African country.

Predicting intention to adopt solar technology in Canada: The role of knowledge, public engagement, and visibility

Solar power (i.e., solar photovoltaic) accounts for about 0.3% of total electricity production in Canada. To enhance this contribution to energy supply from solar power, financial incentives and technological breakthroughs alone may not guarantee change. Drawing on a national survey of 2065 Canadian residents, we identify the determinants of technology adoption intention with the exemplary case of rooftop solar. Using a combination of latent and observed variables within a non-linear structural equation model, our analysis quantifies how a set of individual and community level factors affect adoption intention. Analysis reveals that the visibility of solar technology has a particularly strong effect on intention, lending support to social learning and social network theories of diffusion of innovation. Our findings also show that the perceived knowledge of energy systems and being publicly engaged in energy issues significantly increases adoption intention. These conclusions encourage policy options that enhance public engagement and the visibility of solar technology within neighborhoods and communities.

ASYMMETRIES IN THE RESPONSES OF REGIONAL JOB FLOWS TO OIL PRICE SHOCKS

This paper studies the effect of oil price innovations on manufacturing job flows across U.S. states. First, I estimate a nonlinear structural equation model and compute impulse response functions by Monte Carlo integration. I find asymmetries in the responses of job flows to positive and negative oil price innovations. Yet, these asymmetries do not pass a test of symmetry on the impulse responses, especially after accounting for data mining. Third, I use a test for the absence of job reallocation to evaluate whether an unexpected increase in the real price of oil price triggers an important change in job reallocation. I find that oil price shocks have limited regional allocative effects. (JEL E24, E32, Q43)

Estimation of truncated spline in nonlinear structural equation modeling using weighted least squares method

Structural equation modeling (SEM) consists of two parts, the measurement model and the structural model. In this paper, we develop a SEM model that the measurement model is linear and the structural model is nonlinear. Structural model explain about relationship between latent variables. Truncated spline is used for modeling the nonlinear relationship between the latent variables. Trought weighted least squares (WLS) method, It is found that estimation of parameter truncated spline in nonlinear SEM for two exogen and an endogen latent variables is resulted by optimizingminγ{‖Σ−12(ωη−Γξκγ)‖2}and it gave an estimatorγ^=(Γ′ξkΣ−1Γξk)−1Γ′ξkΣ−1ωηThe estimator influenced by the location and some knot points κ1 and κ2.