Variable Structural Equation Modeling Research Articles

Normrank Correlations for Testing Associations and for Use in Latent Variable Models

Abstract Pearson’s correlation is widely used to test for an association between two variables and also forms the basis of several multivariate statistical procedures including many latent variable models. Spearman’s ρ \rho is a popular alternative. These procedures are compared with ranking the data and then applying the inverse normal transformation, or for short the normrank transformation. Using the normrank transformation was more powerful than Pearson’s and Spearman’s procedures when the distributions have less than normal kurtosis (platykurtic), when the distributions have greater than normal kurtosis (leptokurtic), and when the distribution is skewed. This is examined for testing if there is an association between two variables, identifying the number of factors in an exploratory factor analysis, identifying appropriate loadings in these analyses, and identifying relations among latent variables in structural equation models. R functions and their use are shown.

Exploring the Impact of the COVID-19 Pandemic on Automobile Sales: A Confirmatory Composite Analysis Using Google Trends Data to Develop a Structural Equation Model

Confirmatory factor analysis (CFA) analyzes latent variables in structural equation models representing unmeasurable attitudes or behaviors. In contrast, confirmatory composite analysis (CCA) uses emergent variables such as capacities, values, and indices. When significant secondary and Google trend data are used, CCA may be preferred over CFA. This study employs CCA and partial least squares to develop a structural equation model. It examines COVID-19’s impact on vehicle sales in Thailand. The study uses data from March 2020 to September 2021, including COVID-19 infection rates, newly registered cars, and Google Trends data on COVID-19 and vehicle sales. The results indicate that the overall model fit, and measurement model parameters are outstanding. This suggests a significant negative influence of the COVID-19 pandemic on car sales in Thailand. The study concluded that CCA and partial least squares were used to analyze the impact of COVID -19 on vehicle sales in Thailand. Analysis of secondary data and Google trend data revealed a significant negative impact of the pandemic on car sales, indicating a negative impact on the automotive industry in Thailand during the indicated period.

Determining Factors Affecting Perceived Customer Satisfaction on Public Utility Bus System in Occidental Mindoro, Philippines: A Case Study on Service Quality Assessment during Major Disruptions

The imposition of lockdown due to the COVID-19 pandemic has affected the majority of enterprises throughout the globe. The public transportation sector was greatly affected, especially in provincial areas in the Philippines. This study aimed to investigate the service quality of bus transits in Occidental Mindoro during the COVID-19 pandemic using Analytical Hierarchy Process (AHP) and SERVQUAL dimensions. A total of 200 individuals completed a 67-question online questionnaire. AHP was utilized to determine which bus providers need to be improved the most. Then, the SERVQUAL approach was used. The five dimensions were linked with the use of new technologies, including the COVID-19 protocol, service quality, and customer satisfaction as latent variables in structural equation modeling. The AHP results indicated that safety accounts for the most significant weight for bus passengers. Moreover, service quality, responsiveness, reliability, empathy, and COVID-19 protocol significantly enhance service and customer satisfaction. The findings of this research study can now serve as a service quality assessment and guidelines to provide a sustainable public bus transportation; it can also help determine the significant and insignificant factors in enhancing the service quality and improving the customer satisfaction of bus providers worldwide.

Products of Variables in Structural Equation Models

A general method is introduced in which variables that are products of other variables in the context of a structural equation model (SEM) can be decomposed into the sources of variance due to the multiplicands. The result is a new category of SEM which we call a Products of Variables Model (PoV). Some useful and practical features of PoV models include the estimation of interactions between latent variables, latent variable moderators, manifest moderators with missing values, and manifest or latent squared terms. Expected means and covariances are analytically derived for a simple product of two variables and it is shown that the method reproduces previously published results for this special case. It is shown algebraically that using centered multiplicands results in an unidentified model, but if the multiplicands have non-zero means, the result is identified. The method has been implemented in OpenMx and Ωnyx and is applied in five extensive simulations.

Data Analysis Procedures with Structural Equation Modelling (SEM): Narrative Literature Review

The relationship between variables in structural equation modeling (SEM) forms a structural model. This structural model can be explained through structural equations, such as in regression analysis. This structural equation describes the prediction of the latent (exogenous) independent variable on the latent (endogenous) dependent variable. Researchers who use analysis with structural equation models need to know whether the model built with empirical data has a unique value or not so that the model can be estimated. If the model does not have a unique value, then the model cannot be identified (unidentified). The cause of a model is categorized as unidentified because the information contained in empirical data is not sufficient to produce a unique solution in calculating model estimation parameters. This literature review aims to describe the process of data analysis using SEM.

Parameter estimation of structural equation models with misclassification: The MC-SIMEX approach

The random errors in the measurement process, called measurement error or misclassification, are inevitable and cause bias and inconsistent parameter estimates. Misclassification Simulation Extrapolation (MC-SIMEX) is a simulation based measurement error estimation method to obtain reduced parameter bias under misclassification. The main purpose of this study is an adaptation of MC-SIMEX method on Structural Equation Modeling (SEM). The effects of misclassification on the parameter estimates of a binary explanatory variables in SEM and the performance of MC-SIMEX method investigated with both Monte Carlo and an empirical study. According to the main results, finding the best extrapolant function is just as important as estimating the misclassification matrix although MC-SIMEX corrected a part of the bias.

Policy attitudes toward adolescents transitioning gender.

Anti-transgender policies and state legislative initiatives that focus on school bathroom use and hormone use have emerged in recent years. These policies are generally written by and voted on by cisgender people, and as such, it is crucial to understand influences on nonaffirming attitudes toward policies that can impact trans youth. The present study aimed to extend research on transphobic attitudes in general to attitudes toward policies that impact youth undergoing transition. Latent variable covariances and structural equation modeling were used to test the relations between transphobia, genderism, homophobia, need for closure, sexual orientation, social dominance orientation, attitudes toward sexual minorities, beliefs about gender roles, aggression, religious fundamentalism, and contact with sexual and gender minority individuals, as they are related to attitudes toward hormone use and bathroom use for trans youth. Analyses of data from a sample of 248 cisgender adults indicated that genderism and transphobia was associated with attitudes toward gender-affirming hormone use and bathroom use for trans youth; need for closure was associated with gender-affirming attitudes toward bathroom use, but was not associated with hormone use. Sexual orientation was linked to attitudes toward gender-affirming policies, such that nonheterosexual participants had more affirming attitudes toward trans youths' bathroom use, but not hormone use. Implications for future research, advocacy efforts to promote rights for trans youth, and clinical work with trans youth and/or parents/guardians of trans youth are discussed. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

The Success Factors for Growing Health Tech Startups in Thailand

Startups have served as one of the Thai government’s mechanisms for promoting economic stability and growth. The health tech startup is a global trend toward new innovative industries. In Thailand, there are many health tech startups born each year. However, the healthcare business characteristic and business model differing from other businesses urge the need to understand the relevant factors of success. This knowledge can help you lower your chances of failing and enhance your chances of succeeding. The success elements that influence success, as well as the ones that are most crucial for health tech firms, were investigated in this study. This research used quantitative methodologies to create a systemic approach. The researchers used data from 240 samples to assess basic statistics and confirmatory second-order factor analysis based on the idea of latent variables in Structural Equation Modeling (SEM). It was found that health tech startups in Thailand can be successful because of 6 elements, which are government policy, startup support, human capital, market, finance, and culture. Culture is the most influential factor, followed by startup support, human capital, market, government policy, and finance. It's worth noting that tax relief and low-interest financing aren't high on the priority list. Internal elements such as entrepreneurial inspiration or knowledge, experience, and product prominence, on the other hand, take precedence.

Two-Stage Structural Equation Modeling: Controlling for Omitted and Confounding Variables in Structural Equation Models

Two-Stage Structural Equation Modeling: Controlling for Omitted and Confounding Variables in Structural Equation Models

A metric-based analysis on the effects of riparian and catchment landuse on macroinvertebrates

Woody riparian vegetation along rivers and streams provides multiple functions beneficial for aquatic macroinvertebrate communities. They retain fine sediments, nutrients and pesticides, improve channel hydromorphology, control water temperature and primary production through shading and provide leaves, twigs and large wood.In a recent conceptual model (Feld et al., 2018), woody riparian functions were considered either independent from large-scale landuse stressors (e.g. shading, input of organic matter), or dependent on landuse at larger spatial scales (e.g. fine sediment, nutrient and pesticide retention).We tested this concept using high-resolution data on woody riparian vegetation cover and empirical data from 1017 macroinvertebrate sampling sites in German lowland and mountain streams. Macroinvertebrate metrics indicative for individual functions were used as response variables in structural equation models (SEM), representing the hierarchical structure between the different considered stressors at different spatial scales: catchment, upstream riparian, local riparian and local landuse cover along with hydromorphology and water quality.The analysis only partly confirmed the conceptual model: Biotic integrity and water quality were strongly related to large-scale stressors as expected (absolute total effect 0.345–0.541), but against expectations, fine sediments retention, considered scale-dependent in the conceptual model, was poorly explained by large-scale stressors (absolute total effect 0.027–0.231). While most functions considered independent from large-scale landuse were partly explained by riparian landuse cover (absolute total effect 0.023–0.091) they also were nonetheless affected by catchment landuse cover (absolute total effect 0.017–0.390).While many empirical case studies at smaller spatial scales clearly document the positive effects of restoring woody riparian vegetation, our results suggest that most effects of riparian landuse cover are possibly superimposed by larger-scale stressors. This does not negate localized effects of woody riparian vegetation but helps contextualize limitations to successful restoration measures targeting the macroinvertebrate community.

A protocol for modelling generalised biological responses using latent variables in structural equation models

In this paper, we consider the problem of how to quantitatively characterise the degree to which a study object exhibits a generalised response. By generalised response, we mean a multivariate response where numerous individual properties change in concerted fashion due to some internal integration. In latent variable structural equation modelling (LVSEM), we would typically approach this situation using a latent variable to represent a general property of interest (e.g. performance) and multiple observed indicator variables that reflect the specific features associated with that general property. While ecologists have used LVSEM in a number of cases, there is substantial potential for its wider application. One obstacle is that LV models can be complex and easily over-specified, degrading their value as a means of generalisation. It can also be challenging to diagnose causes of misspecification and understand which model modifications are sensible. In this paper, we present a protocol, consisting of a series of questions, designed to guide the researchers through the evaluation process. These questions address: (1) theoretical development, (2) data requirements, (3) whether responses to perturbation are general, (4) unique reactions by individual measures and (5) how far generality can be extended. For this illustration, we reference a recent study considering the potential consequences of maintaining biodiversity as part of agricultural management on the overall quality of grapes used for wine-making. We extend our presentation to include the complexities that occur when there are multiple species with unique reactions.

The Impact of Scaling Methods on the Properties and Interpretation of Parameter Estimates in Structural Equation Models with Latent Variables

ABSTRACT Latent variables in structural equation models do not have an observable scale. Hence, researchers resort to scaling methods, such as fixed marker, effects coding, or fixed factor, to assign scales to the latent variables. The use of such procedures results in numerically different estimates, in spite of a single underlying population model. In this paper, we provide a framework which not only allows for a translation between estimates obtained under different scaling methods, but also helps to explore the relation between the underlying population parameters and their estimates, thus providing a basis for the interpretation of estimated parameters. Additionally, the framework proves useful for demonstrating that the choice of scaling method affects the power of the Wald test for testing parameters’ significance.

Model-Based Manifest and Latent Composite Scores in Structural Equation Models

Composite scores are commonly used in the social sciences as dependent and independent variables in statistical models. Typically, composite scores are computed prior to statistical analyses. In this paper, we demonstrate the construction of model-based composite scores that may serve as outcomes or predictors in structural equation models (SEMs). Model-based composite scores of manifest variables are useful in the presence of ignorable missing data, as full-information maximum likelihood estimation can be used for parameter estimation. Model-based composite scores of latent variables account for measurement error in the aggregated variables. We introduce the pseudo-indicator model (PIM) for the construction of three composite scores: (a) the sum score, (b) the weighted sum score, and (c) the average score of manifest and latent variables in SEM. The utility of manifest model-based composite scores in the case of missing values is shown by a simulation study. The use of multiple manifest and latent model-based composite scores in SEM is illustrated with data from motivation research.

Using phantom variables in structural equation modeling to assess model sensitivity to external misspecification.

External misspecification, the omission of key variables from a structural model, can fundamentally alter the inferences one makes without such variables present. This article presents 2 strategies for dealing with omitted variables, the first a fixed parameter approach incorporating the omitted variable into the model as a phantom variable where all associated parameter values are fixed, and the other a random parameter approach specifying prior distributions for all of the phantom variable's associated parameter values under a Bayesian framework. The logic and implementation of these methods are discussed and demonstrated on an applied example from the educational psychology literature. The argument is made that such external misspecification sensitivity analyses should become a routine part of measured and latent variable modeling where the inclusion of all salient variables might be in question. (PsycINFO Database Record

A least squares approach to latent variables extraction in formative–reflective models

A new least-squares based procedure for the extraction of latent variables in structural equation models with formative–reflective schemes is developed and illustrated. The procedure is a valuable alternative to PLS-PM and SEM since it is fully consistent with the causal structure of formative–reflective schemes and it extracts the factor scores without substantial identification or indeterminacy problems. Moreover, the new methodology involves the optimization of an explicit and simple to interpret objective function, provides a natural way to check the correct specification of the model and is computationally light. The superiority of the new algorithm over its competitors is proved through examples involving both simulated and real data.

Structural equation modeling of partial and total annoyances due to urban road traffic and aircraft noises

Data of a French in situ socio-acoustic survey were used to model partial annoyance due to urban road traffic noise, partial annoyance due to aircraft noise and total annoyance due to these combined noises. Structural equation modeling carried out on the in situ data showed that long-term noise annoyance depends on noise exposure but also on noise disturbance, dwelling satisfaction, visibility of a main road from the dwelling and noise sensitivity. Both noise exposure and noise sensitivity were introduced as independent variables in structural equation modeling of partial and total noise annoyances. Their contributions to the models highlight the necessity to consider these two variables in annoyance model prediction. Finally, in total noise annoyance models, whereas partial annoyance due to aircraft noise contributes to total noise annoyance as much as partial road traffic annoyance, aircraft noise exposure contributes to total noise annoyance much more than road traffic noise. Several reasons may explai...

Testing Baron and Kenny’s Preliminary Conditions for Mediating or Moderating Variables in Structural Equation Modeling

A common misnomer in statistical analysis is the identification of intervening variables as either moderating or mediating as a result of placement of another latent construct or variables between exogenous and endogenous variables in a Structural Equation Model (SEM). However, placement does not determine variable status. Determination of variable status requires the application of Baron and Kenny’s preliminary conditions and analysis for determining whether a variable is moderating, mediating, or in fact, intervening. This procedure aids in statistically representing the actual strength and direction of variables in order to improve reporting predictor results on outcomes. An SEM was developed using the results of the analysis of the Biomedical Engineering Interdepartmental (BEI) Survey conducted in 2011. This document illustrates the application of Baron and Kenny’s method to the BEI survey SEM model.

Alcohol Use Disorders: Translational Utility of DSM-IV Liabilities to the DSM-5 System.

Young adults have some of the highest rates of problem drinking and alcohol use disorders (AUDs) relative to any other age. However, recent evidence suggests that the DSM-IV hierarchical classification system of AUDs does not validly represent symptoms in the population; instead, it evinces a unitary, dimensional classification scheme. The DSM-5 has been altered to fit this changing, evidence-based conceptualization. Nevertheless, little is understood about the degree to which known risk factors for DSM-IV AUD diagnoses will transfer to the new DSM-5 guidelines in this group of high-risk drinkers. The current study built a coherent model of liabilities for DSM-IV AUDs in young adults and tested for transferability to DSM-5. N = 496 college students (51.10% male) were assessed on a variety of factors related to AUD risk, including demographics, substance use (past 90-days), and drinking motives. Liability models were created using all variables in Structural Equation Modeling to test direct and indirect effects on DSM diagnostic status. The best model under the DSM-IV was chosen based on fit and parsimony. This model was then applied to the DSM-5 system to test for transferability. The best the fitting model for DSM-IV included direct influences of drug use, quantity-frequency of alcohol consumption, and social and coping drinking motives. Improved model fit was found when the DSM-5 system was the outcome. Knowledge of risk factors for AUDs appear to transfer well to the new diagnostic system.

Gilbert A. Churchill Jr.'s Editorship of Journal of Marketing Research, 1979–1982

Gil Churchill served at Journal of Marketing Research (JMR) during an important time of maturation for the field. Journal of Marketing (JM) and JMR were firmly established, and both encouraged articles across the realm of marketing (e.g., segmentation, pricing, channels, advertising). Whereas JM was positioned a bit more for immediate relevance to practitioners, JMR was oriented more toward methods of conducting research across those marketing realms. The balancing act was one that Gil handled brilliantly, as evidenced by the Tables of Contents across his issues. For example, consider the article by Aaker and Bagozzi (1979), “Unobservable Variables in Structural Equation Models with an Application in Industrial Selling,” which nicely melds a methodological contribution with one in a substantive area of inquiry. Naturally, most articles emphasize one facet more than the other, but Gil achieved balance within each issue. For example, in the November 1978 issue (Vol. 15), there are two articles on sales force issues and two on measurement; in the August 1979 issue (Vol. 16), there are four articles on advertising and four articles on statistical issues.

Do sensitive parents foster kind children, or vice versa? Bidirectional influences between children’s prosocial behavior and parental sensitivity.

Bidirectional theories of social development have been around for over 40 years (Bell, 1968), yet they have been applied primarily to the study of antisocial development. In the present study, the reciprocal relationship between parenting behavior and children's socially competent behaviors were examined. Using the National Institute of Child Health and Development Study of Early Child Care data set (NICHD Early Child Care Research Network, 2005), bidirectional relationships between parental sensitivity and children's prosocial behavior were modeled using latent variables in structural equation modeling for mothers and fathers, separately. Children and their parents engaged in structured interactions when children were 54-month-olds, 3rd graders, and 5th graders, and these interactions were coded for parental sensitivity. At 3rd, 5th, and 6th grades, teachers and parents reported on children's prosocial behavior. Parental education and child gender were entered as covariates in the models. The results provide support for a bidirectional relationship between children's prosocial behavior and maternal sensitivity (but not paternal sensitivity) in middle childhood. The importance of using a bidirectional approach to examine the development of social competence is emphasized.