Swain: Stata module to correct the SEM chi-square overidentification test in small sample sizes or complex models. Statistical Software Components S457617, Boston College Department of Economics.

Swain corrects the chi-square overidentification test (i.e., likelihood ratio test of fit) for structural equation models whethr with or without latent variables. The chi-square statistic is asymptotically correct; however, it does not behave as expected in small samples and/or when the model is complex (cf. Herzog, Boomsma, & Reinecke, 2007). Thus, particularly in situations where the ratio of sample size (n) to the number of parameters estimated (p) is relatively small (i.e., the p to n ratio is large), the chi-square test will tend to overreject correctly specified models. To obtain a closer approximation to the distribution of the chi-square statistic, Swain (1975) developed a correction; this scaling factor, which converges to 1 asymptotically, is multiplied with the chi-square statistic. The correction better approximates the chi-square distribution resulting in more appropriate Type 1 reject error rates (see Herzog & Boomsma, 2009; Herzog, et al., 2007).

Country effects in ISSP-1993 environmental data: Comparison of SEM approaches

Structural equation models (SEM) are commonly used to analyze the relationship between variables some of which may be latent, such as individual ``attitude'' to and ``behavior'' concerning specific issues. A number of difficulties arise when we want to compare a large number of groups, each with large sample size, and the manifest variables are distinctly non-normally distributed. Using an specific data set, we evaluate the appropriateness of the following alternative SEM approaches: multiple group versus MIMIC models, continuous versus ordinal variables estimation methods, and normal theory versus non-normal estimation methods. The approaches are applied to the ISSP-1993 Environmental data set, with the purpose of exploring variation in the mean level of variables of ``attitude'' to and ``behavior''concerning environmental issues and their mutual relationship across countries. Issues of both theoretical and practical relevance arise in the course of this application.

A step beyond--the relevance of depressed mood and mastery in the interplay between the number of social roles and alcohol use.

OBJECTIVES: The present study examines whether depressed mood and external control mediate or moderate the relationship between the number of social roles and alcohol use. PARTICIPANTS: The analysis was based on a national representative sample of 25- to 45-year-old male and female drinkers in Switzerland. METHOD: The influence of depressed mood and external control on the relationship between the number of social roles (parenthood, partnership, employment) and alcohol use was examined in linear structural equation models (mediation) and in multiple regressions (moderation) stratified by gender. All analyses were adjusted for age and education level. RESULTS: Holding more roles was associated with lower alcohol use, lower external control and lower depressed mood. The study did not find evidence of depressed mood or external control mediating the social roles-alcohol relationship. A moderation effect was identified among women only, whereby a protective effect of having more roles could not be found among those who scored high on external control. In general, a stronger link was observed between roles and alcohol use, while depressed mood and external control acted independently on drinking. With the exception of women with high external control, the study found no link between a higher number of social roles and greater alcohol use. CONCLUSION: Our results indicate that drinking behaviours are more strongly linked to external control and depressed mood than they are to the number of social roles. The study also demonstrates that in any effective alcohol prevention policy, societal actions that enable individuals to combine more social roles play a central role.

Selecting statistical models to study the relationship between soybean yield and soil physical properties

Statistical models allow the representation of data sets and the estimation and/or prediction of the behavior of a given variable through its interaction with the other variables involved in a phenomenon. Among other different statistical models, are the autoregressive state-space models (ARSS) and the linear regression models (LR), which allow the quantification of the relationships among soil-plant-atmosphere system variables. To compare the quality of the ARSS and LR models for the modeling of the relationships between soybean yield and soil physical properties, Akaike's Information Criterion, which provides a coefficient for the selection of the best model, was used in this study. The data sets were sampled in a Rhodic Acrudox soil, along a spatial transect with 84 points spaced 3 m apart. At each sampling point, soybean samples were collected for yield quantification. At the same site, soil penetration resistance was also measured and soil samples were collected to measure soil bulk density in the 0-0.10 m and 0.10-0.20 m layers. Results showed autocorrelation and a cross correlation structure of soybean yield and soil penetration resistance data. Soil bulk density data, however, were only autocorrelated in the 0-0.10 m layer and not cross correlated with soybean yield. The results showed the higher efficiency of the autoregressive space-state models in relation to the equivalent simple and multiple linear regression models using Akaike's Information Criterion. The resulting values were comparatively lower than the values obtained by the regression models, for all combinations of explanatory variables.

Building predictive models of soil particle-size distribution

Is it possible to build predictive models (PMs) of soil particle-size distribution (psd) in a region with complex geology and a young and unstable land-surface? The main objective of this study was to answer this question. A set of 339 soil samples from a small slope catchment in Southern Brazil was used to build PMs of psd in the surface soil layer. Multiple linear regression models were constructed using terrain attributes (elevation, slope, catchment area, convergence index, and topographic wetness index). The PMs explained more than half of the data variance. This performance is similar to (or even better than) that of the conventional soil mapping approach. For some size fractions, the PM performance can reach 70 %. Largest uncertainties were observed in geologically more complex areas. Therefore, significant improvements in the predictions can only be achieved if accurate geological data is made available. Meanwhile, PMs built on terrain attributes are efficient in predicting the particle-size distribution (psd) of soils in regions of complex geology.

Predicting overall service quality: a structural equation modelling approach

In this article, the results of a modified SERVQUAL questionnaire (Parasuraman et al., 1991) are reported. The modifications consisted in substituting questionnaire items particularly suited to a specific service (banking) and context (county of Girona, Spain) for the original rather general and abstract items. These modifications led to more interpretable factors which accounted for a higher percentage of item variance. The data were submitted to various structural equation models which made it possible to conclude that the questionnaire contains items with a high measurement quality with respect to five identified dimensions of service quality which differ from those specified by Parasuraman et al. And are specific to the banking service. The two dimensions relating to the behaviour of employees have the greatest predictive power on overall quality and satisfaction ratings, which enables managers to use a low-cost reduced version of the questionnaire to monitor quality on a regular basis. It was also found that satisfaction and overall quality were perfectly correlated thus showing that customers do not perceive these concepts as being distinct

Minimization of the experimental workload for the prediction of pollutants propagation in rivers. Mathematical modelling and knowledge re-use

The present thesis in focused on the minimization of experimental efforts for the prediction of pollutant propagation in rivers by mathematical modelling and knowledge re-use. Mathematical modelling is based on the well known advection-dispersion equation, while the knowledge re-use approach employs the methods of case based reasoning, graphical analysis and text mining. The thesis contribution to the pollutant transport research field consists of: (1) analytical and numerical models for pollutant transport prediction; (2) two novel techniques which enable the use of variable parameters along rivers in analytical models; (3) models for the estimation of pollutant transport characteristic parameters (velocity, dispersion coefficient and nutrient transformation rates) as functions of water flow, channel characteristics and/or seasonality; (4) the graphical analysis method to be used for the identification of pollution sources along rivers; (5) a case based reasoning tool for the identification of crucial information related to the pollutant transport modelling; (6) and the application of a software tool for the reuse of information during pollutants transport modelling research. These support tools are applicable in the water quality research field and in practice as well, as they can be involved in multiple activities. The models are capable of predicting pollutant propagation along rivers in case of both ordinary pollution and accidents. They can also be applied for other similar rivers in modelling of pollutant transport in rivers with low availability of experimental data concerning concentration. This is because models for parameter estimation developed in the present thesis enable the calculation of transport characteristic parameters as functions of river hydraulic parameters and/or seasonality. The similarity between rivers is assessed using case based reasoning tools, and additional necessary information can be identified by using the software for the information reuse. Such systems represent support for users and open up possibilities for new modelling methods, monitoring facilities and for better river water quality management tools. They are useful also for the estimation of environmental impact of possible technological changes and can be applied in the pre-design stage or/and in the practical use of processes as well.

Thymocyte migration: an affair of multiple cellular interactions?

Cell migration is a crucial event in the general process of thymocyte differentiation. The cellular interactions involved in the control of this migration are beginning to be defined. At least chemokines and extracellular matrix proteins appear to be part of the game. Cells of the thymic microenvironment produce these two groups of molecules, whereas developing thymocytes express the corresponding receptors. Moreover, although chemokines and extracellular matrix can drive thymocyte migration per se, a combined role for these molecules appears to contribute to the resulting migration patterns of thymocytes in their various stages of differentiation. The dynamics of chemokine and extracellular matrix production and degradation is not yet well understood. However, matrix metalloproteinases are likely to play a role in the breakdown of intrathymic extracellular matrix contents. Thus, the physiological migration of thymocytes should be envisioned as a resulting vector of multiple, simultaneous and/or sequential stimuli involving chemokines, adhesive and de-adhesive extracellular matrix proteins, as well as matrix metalloproteinases. Accordingly, it is conceivable that any pathological change in any of these loops may result in the alteration of normal thymocyte migration. This seems to be the case in murine infection by the protozoan parasite Trypanosoma cruzi, the causative agent of Chagas' disease. A better knowledge of the physiological mechanisms governing thymocyte migration will provide new clues for designing therapeutic strategies targeting developing T cells.

Exact Tests for Contemporaneous Correlation of Disturbances in Seemingly Unrelated Regressions.

This paper proposes finite-sample procedures for testing the SURE specification in multi-equation regression models, i.e. whether the disturbances in different equations are contemporaneously uncorrelated or not. We apply the technique of Monte Carlo (MC) tests [Dwass (1957), Barnard (1963)] to obtain exact tests based on standard LR and LM zero correlation tests. We also suggest a MC quasi-LR (QLR) test based on feasible generalized least squares (FGLS). We show that the latter statistics are pivotal under the null, which provides the justification for applying MC tests. Furthermore, we extend the exact independence test proposed by Harvey and Phillips (1982) to the multi-equation framework. Specifically, we introduce several induced tests based on a set of simultaneous Harvey/Phillips-type tests and suggest a simulation-based solution to the associated combination problem. The properties of the proposed tests are studied in a Monte Carlo experiment which shows that standard asymptotic tests exhibit important size distortions, while MC tests achieve complete size control and display good power. Moreover, MC-QLR tests performed best in terms of power, a result of interest from the point of view of simulation-based tests. The power of the MC induced tests improves appreciably in comparison to standard Bonferroni tests and, in certain cases, outperforms the likelihood-based MC tests. The tests are applied to data used by Fischer (1993) to analyze the macroeconomic determinants of growth.

Semi-Parametric Weak Instrument Regressions with an Application to the Risk-Return Trade-off

Recent work shows that a low correlation between the instruments and the included variables leads to serious inference problems. We extend the local-to-zero analysis of models with weak instruments to models with estimated instruments and regressors and with higher-order dependence between instruments and disturbances. This makes this framework applicable to linear models with expectation variables that are estimated non-parametrically. Two examples of such models are the risk-return trade-off in finance and the impact of inflation uncertainty on real economic activity. Results show that inference based on Lagrange Multiplier (LM) tests is more robust to weak instruments than Wald-based inference. Using LM confidence intervals leads us to conclude that no statistically significant risk premium is present in returns on the S&P 500 index, excess holding yields between 6-month and 3-month Treasury bills, or in yen-dollar spot returns.

Tests multiples simulés et tests de normalité basés sur plusieurs moments dans les modèles de régression

Cet article illustre l’applicabilité des méthodes de rééchantillonnage dans le cadre des tests multiples (simultanés), pour divers problèmes économétriques. Les hypothèses simultanées sont une conséquence habituelle de la théorie économique, de sorte que le contrôle de la probabilité de rejet de combinaisons de tests est un problème que l’on rencontre fréquemment dans divers contextes économétriques et statistiques. À ce sujet, on sait que le fait d’ignorer le caractère conjoint des hypothèses multiples peut faire en sorte que le niveau de la procédure globale dépasse considérablement le niveau désiré. Alors que la plupart des méthodes d’inférence multiple sont conservatrices en présence de statistiques non-indépendantes, les tests que nous proposons visent à contrôler exactement le niveau de signification. Pour ce faire, nous considérons des critères de test combinés proposés initialement pour des statistiques indépendantes. En appliquant la méthode des tests de Monte Carlo, nous montrons comment ces méthodes de combinaison de tests peuvent s’appliquer à de tels cas, sans recours à des approximations asymptotiques. Après avoir passé en revue les résultats antérieurs sur ce sujet, nous montrons comment une telle méthodologie peut être utilisée pour construire des tests de normalité basés sur plusieurs moments pour les erreurs de modèles de régression linéaires. Pour ce problème, nous proposons une généralisation valide à distance finie du test asymptotique proposé par Kiefer et Salmon (1983) ainsi que des tests combinés suivant les méthodes de Tippett et de Pearson-Fisher. Nous observons empiriquement que les procédures de test corrigées par la méthode des tests de Monte Carlo ne souffrent pas du problème de biais (ou sous-rejet) souvent rapporté dans cette littérature – notamment contre les lois platikurtiques – et permettent des gains sensibles de puissance par rapport aux méthodes combinées usuelles.

Étude longitudinale des liens prédictifs entre les traits de personnalité et les symptômes intériorisés à l'adolescence.

Cette étude longitudinale visait à vérifier si les traits de personnalité (selon le modèle en cinq facteurs, « Big Five ») au début de l’adolescence (12-13 ans) permettent de prédire les symptômes intériorisés deux ans plus tard (14-15 ans), en contrôlant pour le niveau initial de symptômes intériorisés ainsi que l’influence de plusieurs facteurs de risque connus. Les données employées proviennent d’une étude longitudinale prospective. L’échantillon compte 1036 adolescents provenant de huit écoles secondaires québécoises. Les adolescents ont répondu à un questionnaire autorévélé. Des modèles d’équations structurales ont d’abord démontré la pertinence de conceptualiser les symptômes intériorisés comme une variable latente. D’autres modèles ont démontré que certains traits de personnalité prédisent effectivement les symptômes intériorisés ultérieurs. Cependant, contrairement aux études effectuées auprès d’adultes, le rôle de la Stabilité émotionnelle et de l’Extraversion n’est pas significatif après que l’influence de facteurs de risque connus et du sexe ait été contrôlée. Ce sont plutôt le Contrôle et l’Amabilité qui sont significativement reliés aux symptômes intériorisés ultérieurs dans la présente étude. Les résultats soulignent également le rôle important des facteurs de risque liés aux relations avec les pairs. Finalement, des modèles d’équations structurales multi-groupes ont mis en évidence des différences sexuelles significatives dans les relations prédictives. Cette étude confirme que les traits de personnalité des adolescents peuvent jouer un rôle dans le développement des symptômes intériorisés, ce qui leur confère une pertinence théorique et clinique.

Étude longitudinale des liens prédictifs entre les traits de personnalité et les comportements antisociaux à l’adolescence

Cette étude longitudinale visait à évaluer si les traits de personnalité des adolescents permettent de prédire leurs comportements antisociaux ultérieurs, après avoir contrôlé pour l’effet du niveau initial du comportement antisocial ainsi que celui de plusieurs facteurs de risque connus de ces comportements. L’échantillon utilisé compte 1036 adolescents provenant de huit écoles secondaires québécoises. Les adolescents ont été évalués à deux reprises, soit en secondaire 1 (12-13 ans) et en secondaire 3 (14-15 ans). Ils ont répondu à un questionnaire autorévélé. Des modèles d’équations structurales ont d’abord confirmé que la covariation entre différents comportements antisociaux des adolescents peut être expliquée par une variable latente. Les résultats ont confirmé que les traits de personnalité des adolescents à 12 et 13 ans prédisent leurs comportements antisociaux à 14 et 15 ans. En accord avec les études antérieures, l’Extraversion, le Contrôle et la Stabilité émotionnelle prédisent les comportements antisociaux futurs. Toutefois, l’effet de l’Amabilité disparait une fois que le niveau initial est contrôlé. Finalement, des modèles d’équations structurales multi-groupes ont permis de démontrer que certaines relations prédictives sont différentes selon le sexe. Les résultats de cette étude soulignent l’importance des traits de personnalité pour les théories du comportement antisocial ainsi que pour la pratique clinique.

Predicting random level and seasonality of hotel prices. A structural equation growth curve approach

This article examines the effect on price of different characteristics of holiday hotels in the sun-and-beach segment, under the hedonic function perspective. Monthly prices of the majority of hotels in the Spanish continental Mediterranean coast are gathered from May to October 1999 from the tour operator catalogues. Hedonic functions are specified as random-effect models and parametrized as structural equation models with two latent variables, a random peak season price and a random width of seasonal fluctuations. Characteristics of the hotel and the region where they are located are used as predictors of both latent variables. Besides hotel category, region, distance to the beach, availability of parking place and room equipment have an effect on peak price and also on seasonality. 3- star hotels have the highest seasonality and hotels located in the southern regions the lowest, which could be explained by a warmer climate in autumn

Correcting Measurement Error Bias in Interaction Models with Small Samples

Several methods have been suggested to estimate non-linear models with interaction terms in the presence of measurement error. Structural equation models eliminate measurement error bias, but require large samples. Ordinary least squares regression on summated scales, regression on factor scores and partial least squares are appropriate for small samples but do not correct measurement error bias. Two stage least squares regression does correct measurement error bias but the results strongly depend on the instrumental variable choice. This article discusses the old disattenuated regression method as an alternative for correcting measurement error in small samples. The method is extended to the case of interaction terms and is illustrated on a model that examines the interaction effect of innovation and style of use of budgets on business performance. Alternative reliability estimates that can be used to disattenuate the estimates are discussed. A comparison is made with the alternative methods. Methods that do not correct for measurement error bias perform very similarly and considerably worse than disattenuated regression