A general structural equation model with dichotomous,ordered categorical,and continuous latent variable indicators

A structural equation model is proposed with a generalized measurement part, allowing for dichotomous and ordered categorical variables (indicators) in addition to continuous ones. A computationally feasible three-stage estimator is proposed for any combination of observed variable types. This approach provides large-sample chi-square tests of fit and standard errors of estimates for situations not previously covered. Two multiple-indicator modeling examples are given. One is a simultaneous analysis of two groups with a structural equation model underlying skewed Likert variables. The second is a longitudinal model with a structural model for multivariate probit regressions.This research was supported by Grant No. 81-IJ-CX-0015 from the National Institute of Justice, by Grant No. DA 01070 from the U.S. Public Health Service, and by Grant No. SES-8312583 from the National Science Foundation. I thank Julie Honig for drawing the figures. Requests for reprints should be sent to Bengt Muthén, Graduate School of Education, University of California, Los Angeles, California 90024.     

Expectation confirmation: An examination of three competing models

We discuss three theoretical models from met expectations research in the fields of organizational behavior and consumer psychology. Based on the fundamental arguments in the models, we term these models: disconfirmation, ideal point, and experiences only. We present three-dimensional graphical and analytical representations of the models, with satisfaction being a function of expectations and experiences. We tested the models in the context of a new information system implementation in an organization, with expectations, experiences, and system satisfaction measured for both ease of use and usefulness, the focal constructs of the Technology Acceptance Model (TAM). We found that an experiences only model in which expectations had no measurable effect best explained the data for ease of use. The results for usefulness indicated a modified version of the experiences only model in which the positive effect of experiences becomes slightly stronger—i.e., more positive—as expectations increase.     

Factors predictive of a fetal alcohol spectrum disorder diagnosis: Parent and teacher ratings

Diagnostic assessment in Fetal Alcohol Spectrum Disorder (FASD) is informed by multidisciplinary assessment incorporating objective (i.e., test measures) and subjective means, such as parent and teacher behavior ratings. The purpose of this study was to extend our previous neuropsychological test findings by identifying parent and teacher ratings of academic achievement, attention, executive functioning, and adaptive functioning as predictors of an FASD diagnosis. The charts of 315 children and adolescents with prenatal alcohol exposure (PAE) who underwent assessment for FASD were retrospectively reviewed. Direct logistic regressions analyzed the contribution of different ratings on the likelihood of an FASD diagnosis. The results suggest that a number of rating measures do contribute toward accurately differentiating those with FASD from within a PAE population, including teacher ratings of learning problems, inattention, and adaptive skills. The classification accuracy for each regression was clinically significant (59.1–70.8%). Children with worse ratings on these variables are approximately 1.5 to 2 times more likely to receive an FASD diagnosis. Only teacher ratings (not parent) significantly contributed to whether a diagnosis was made, suggesting that teacher observational rating scales are a critical component of an FASD assessment. Together with our previous research examining neuropsychological evaluation and FASD diagnostic assessment, this study helps to further guide decisions to streamline care in multidisciplinary assessment and intervention planning.     

A second generation nonlinear factor analysis

Nonlinear common factor models with polynomial regression functions, including interaction terms, are fitted by simultaneously estimating the factor loadings and common factor scores, using maximum-likelihood-ratio and ordinary-least-squares methods. A Monte Carlo study gives support to a conjecture about the form of the distribution of the likelihood-ratio criterion.The research reported in this paper was partly supported by Natural Sciences and Engineering Research Grant No. A6346.     

Beyond individual justice facets: How (Mis)alignment between justice climates affects customer satisfaction through frontline customer extra-role service behavior

ABSTRACT

This study examines the combined effect of organizational justice facets on store-level customer extra-role service behavior, and subsequently on customer satisfaction. Hypotheses were tested on a sample of 1,951 employees in 121 business units from four countries, and on 55,731 customers of an international retailer. The results of polynomial regression and response surface analysis revealed that unit customer service performance and customer satisfaction are higher when justice facets are aligned at high levels, compared to when they are aligned at low levels. Moreover, we found evidence that the consequences of misalignment between justice facets are asymmetrical. Unit outcomes were higher when distributive justice (DJ) and procedural justice (PJ) climates were both higher than the interpersonal justice climate, compared to when the inverse was true. Conversely, unit outcomes increased when informational justice (INF-J) climate was higher than DJ and PJ climates, compared to when DJ and PJ climates were higher than INF-J climates. The observed effects of misalignment between justice facets were non-linear, as complex curvilinear relationships were moderator-dependent. Customer satisfaction was higher in stores with higher team customer service behavior, and team service behavior was found to be a significant conduit by which justice facets (mis)alignment influence customer satisfaction.     

Simplified method based on an intelligent model to obtain the extinction angle of the current for a single-phase half wave controlled rectifier with resistive and inductive load

With the aim of calculating the extinction angle of the current of a single-phase half wave controlled rectifier with resistive and inductive load, present work shows a method to obtain a regression model based on intelligent methods. This type of circuit is a typical non-linear case of study that requires a hard work to solve it by hand. To create the intelligent model, a dataset has been obtained with a computational method for the working range of the circuit. Then, with the dataset, to achieve the final solution, several methods of regression were tested from traditional to intelligent types. The model was verified empirically with electronic circuit software simulation, analytical methods and with a practical implementation. The advantage of the proposed method is its low computational cost. Then, the final solution is very appropriate for applications where high computational requirements are not possible, like low-performance microcontrollers or web applications.     

Investigating discrepancies in program quality related to youth volleyball athletes’ needs support

Program quality has been outlined as an important predictor of positive outcomes in youth development programs; however, little evidence exists to support this, particularly within sport. Objective: The purpose of this study was to examine the associations between researcher- and coach-assessed program quality scores as they relate to youth volleyball athletes’ basic needs support. Design: Observational data and self-report data were gathered from coaches and youth. Researchers completed 84 observations using a measure of program quality across 14 teams. Coaches completed the same measure at the end of the season and 138 athletes (M_age = 14.50) from the 14 teams completed a self-report questionnaire pertaining to needs support. Data were analyzed using polynomial regressions with response surface methodology. Results: Athletes’ needs support was significantly (p < 0.001) associated with all domains of researcher- and coach-assessed program quality (i.e., safe environment, supportive environment, interaction, engagement), and between 20 and 35% of the variance in athletes’ needs support was explained by these variables. The degree of discrepancy between researcher- and coach-assessed program quality increased when progressing through three domains of program quality (safe environment, supportive environment, engagement). Response surface methods indicated that as the degree of agreement increased between researchers’ and coaches’ ratings of the safety of the environment, supportiveness of the environment, and opportunities for engagement, so did athletes’ basic needs support scores in a linear fashion. Conclusions: Practical implications surrounding coach education and the importance of knowledge translation between academics and practitioners are noted, and future research directions are discussed.     

Asymptotic bias of normal-distribution-based maximum likelihood estimates of moderation effects with data missing at random

Moderation analysis is useful for addressing interesting research questions in social sciences and behavioural research. In practice, moderated multiple regression (MMR) models have been most widely used. However, missing data pose a challenge, mainly because the interaction term is a product of two or more variables and thus is a non-linear function of the involved variables. Normal-distribution-based maximum likelihood (NML) has been proposed and applied for estimating MMR models with incomplete data. When data are missing completely at random, moderation effect estimates are consistent. However, simulation results have found that when data in the predictor are missing at random (MAR), NML can yield inaccurate estimates of moderation effects when the moderation effects are non-null. Simulation studies are subject to the limitation of confounding systematic bias with sampling errors. Thus, the purpose of this paper is to analytically derive asymptotic bias of NML estimates of moderation effects with MAR data. Results show that when the moderation effect is zero, there is no asymptotic bias in moderation effect estimates with either normal or non-normal data. When the moderation effect is non-zero, however, asymptotic bias may exist and is determined by factors such as the moderation effect size, missing-data proportion, and type of missingness dependence. Our analytical results suggest that researchers should apply NML to MMR models with caution when missing data exist. Suggestions are given regarding moderation analysis with missing data.     

Factors predictive of a fetal alcohol spectrum disorder: Neuropsychological assessment

A variety of neurodevelopmental impairments related to fetal alcohol spectrum disorder (FASD) diagnoses have been consistently documented. However, it is not clear whether such variables are predictive of a diagnosis. The purpose of the present study is to use logistic regressions to identify predictors of FASD in neuropsychological assessment. Charts of 180 children and adolescents with prenatal alcohol exposure (PAE) who underwent psychological and diagnostic assessment for FASD were retrospectively reviewed. A total of 107 received an FASD diagnosis (the PAE-FASD group) and 73 did not (the PAE group). Following preliminary analyses, direct logistic regressions were performed to assess the contribution of different neuropsychological testing measures on the likelihood of a child or adolescent receiving an FASD diagnosis. The results indicate that the classification accuracy of the PAE-FASD and PAE groups is clinically significant across models of intelligence, academic achievement, memory, and executive functioning. Classification rates across the various models range from 67.1% to 75.5%, with models incorporating 10 intelligence subtests or 3 academic subtests emerging as superior to those using broad indices of intelligence and/or individual subtests of memory or executive functioning. A “test battery” model incorporating verbal intelligence, verbal/auditory working memory (digit span), basic reading and spelling skills, math calculations, delayed story recall, and spatial planning and problem-solving yielded a classification rate of 74.7%. These results suggest that neuropsychological testing is a critical component of FASD assessment and help guide decisions to maximize the efficiency and efficacy of the diagnostic process and treatment recommendations.