Search

Describes a Monte Carlo investigation of four methods for treating incomplete data: (1) resemblance based hot-deck imputation (RBHDI); (2) iterated stochastic regression imputation; (3) structured model expectation maximization; and (4) saturated model expectation maximization. Results favored the expectation maximization methods. (SLD)

Proposes generalized least squares methods in structural equation models to estimate potential mean structure parameters and to evaluate whether the given model can be augmented successfully with a mean structure. A simulation shows that a method that takes variability due to the estimate of covariance structure parameters into account performs best. (SLD)

Investigated the conditions under which the matrix of factor loadings from the factor analysis model with equal unique variances will give a good approximation to the matrix of factor loadings from the regular factor analysis model. Extends the results to the image factor analysis model. Discusses implications for practice. (SLD)

Adapts robust schemes to mean and covariance structures, providing an iteratively reweighted least squares approach to robust structural equation modeling. Each case is weighted according to its distance, based on first and second order moments. Test statistics and standard error estimators are given. (SLD)

Studied whether the standard z-statistic that evaluates whether a factor loading is statistically necessary is correctly applied in such situations and more generally when the variables being analyzed are arbitrarily rescaled. An example illustrates that neither the factor loading estimates nor the standard error estimates possess scale equivariance or invariance. (SLD)

Shows that phantom factors are unnecessary to achieve the objectives of a hierarchical regression and gives a direct approach to computing hierarchical or fixed-order regressions that is equivalent to that proposed by P. de Jong (1999).(SLD)

The adequacy of "rule of thumb" conventional cutoff criteria and several alternatives for fit indices in covariance structure analysis was evaluated through simulation. Analyses suggest that, for all recommended fit indexes except one, a cutoff criterion greater than (or sometimes smaller than) the conventional rule of thumb is required for model selection or evaluation. (SLD)

Studied the small sample behavior of several test statistics based on the maximum-likelihood estimator but designed to perform better with nonnormal data. Monte Carlo results indicate the satisfactory performance of the "F" statistic recently proposed by K. Yuan and P. Bentler (1997). (SLD)

Two bootstrap tests are proposed to test the independence hypothesis in a two-way cross table. Monte Carlo studies are used to compare the traditional asymptotic test with these bootstrap methods, and the bootstrap methods are found superior in two ways: control of Type I error and statistical power. (SLD)

Proposes a two-stage estimation method for the analysis of covariance structure models with ordinal ipsative data (OID). A goodness-of-fit statistic is given for testing the hypothesized covariance structure matrix, and simulation results show that the method works well with a large sample. (SLD)

A test for linear trend among a set of eigenvalues of a correlation matrix is described. It is a generalization of G. Anderson's (1965) test for the equality of eigenvalues and extends the present authors' previous work on linear trends in eigenvalues of a covariance matrix. The linear trend hypothesis is discussed. (SLD)

Compares two statistical approaches to modeling growth across time: the multilevel model and latent curve analysis. A longitudinal data set from a school-based substance-use-prevention model for adolescents (2,779 students) illustrates differences and similarities between the approaches and shows very compatible results for both. (SLD)

One of the goals of the present work was to study the relationship between child-to-parent violence (CPV) and other types of intra-family violence such as inter-parental violence and parent-to-child violence, in order to verify which of these two types of domestic violence is a more relevant risk factor for CPV and to analyze the presence of gender differences in the bi-directionality of violence. Another purpose was to identify the psychological profile of perpetrators. The sample comprised 485 adolescents from the province of Gipuzkoa (Spain), of both sexes, taken from nine schools and aged 12 to 18. Parent-to-child violence and inter-parental violence were significant risk factors for CPV. Evidence was found in support of a social learning taking into account gender: boys were more likely to be physically aggressive toward the mother if she was also physically victimized by the father. Differences were found in the profiles of adolescents who behave violently toward their parents (inappropriate upbringing by mother, social maladjustment, and drug abuse) depending on gender. Keywords Child-to-parent violence Adolescents * Family violence * Maladjustment * Gender differences, Violence against parents is understood to be any act by a child to gain power and control over parents by causing them physical, psychological, emotional or financial harm (Cottrell 2001). [...]

n real data analysis with structural equation modeling, data are unlikely to be exactly normally distributed. If we ignore the non-normality reality, the parameter estimates, standard error estimates, and model fit statistics from normal theory based methods such as maximum likelihood (ML) and normal theory based generalized least squares estimation (GLS) are unreliable. On the other hand, the asymptotically distribution free (ADF) estimator does not rely on any distribution assumption but cannot demonstrate its efficiency advantage with small and modest sample sizes. The methods which adopt misspecified loss functions including ridge GLS (RGLS) can provide better estimates and inferences than the normal theory based methods and the ADF estimator in some cases. We propose a distributionally weighted least squares (DLS) estimator, and expect that it can perform better than the existing generalized least squares, because it combines normal theory based and ADF based generalized least squares estimation. Computer simulation results suggest that model-implied covariance based DLS (DLSM) provided relatively accurate and efficient estimates in terms of RMSE. In addition, the empirical standard errors, the relative biases of standard error estimates, and the Type I error rates of the Jiang-Yuan rank adjusted model fit test statistic (TJY) in DLSM were competitive with the classical methods including ML, GLS, and RGLS. The performance of DLSM depends on its tuning parameter a. We illustrate how to implement DLSM and select the optimal a by a bootstrap procedure in a real data example. [ABSTRACT FROM AUTHOR]

Chi-square tests based on maximum likelihood (ML) estimation of covariance structures often incorrectly over-reject the null hypothesis: Σ = Σ θ when the sample size is small. Reweighted least squares (RLS) avoids this problem. In some models, the vector of parameter must contain means, variances, and covariances, yet whether RLS also works in mean and covariance structures remains unexamined. This research extends RLS to mean and covariance structures, evaluating a generalized least squares function with ML parameter estimates. A Monte Carlo simulation study was carried out to examine the statistical performance of ML vs RLS with multivariate normal data. Based on empirical rejection frequencies and empirical averages of test statistics, this study shows that RLS performs much better than ML in mean and covariance structure models when sample sizes are small, whereas it does not perform better than ML to reject misspecified models. [ABSTRACT FROM AUTHOR]