Estadística

Linear mixed modeling is a useful approach for double mixed factorial designs with covariates. It is explained how these designs are appropriate for the study of human behavior as a function of characteristics of persons and situations and stimuli in the situations. The behavior of subjects nested in types of persons responding to stimuli nested in types of stimuli defines a mixed factorial design. The inclusion of additional covariates of the observational units can help to further explain the behavior under study. A linear mixed modeling approach for such designs allows a combined focus on fixed effects (general effects) and individual and stimulus differences in these effects. This combination has the potential to advance the integration of two different sub-disciplines of psychology: general psychology and differential psychology, so that they can borrow strength from each other. An application is presented with semantic categorization response time data from a factorial design with age groups by word types and with age of acquisition as an additional covariate of the words. The results throw light on the processes underlying the effect of age of acquisition and on individual differences and word differences.

Local equating is based on Lord's (1980) criterion of equity. It defines a family of true transformations that aiming at the ideal of equitable equating. van der Linden (this issue) offers a detailed discussion of common issues in observed score equating relative to this local approach. By assuming an underlying IRT model, one of the main features of local equating is that it adjusts the equated raw scores using conditional distributions of raw scores given an estimate of the ability of interest. In this paper we argue that when using a Rasch model for the estimation of the true transformation, this feature disappears, while the One Parameter Logistic Model (OPLM) and the 2PL do provide a local adjustment of the equated score.

criterion of equity of equating, van der Linden (this issue) revisits the so-called local equating method and offers alternative as well as new thoughts on several topics appearing in the field such as the types of transformations, symmetry, reliability, population invariance, used in equating. A remarkable aspect is to define equating as a standard statistical inference problem in which the true equating transformation is the parameter of interest which has to be estimated and assessed as any standard evaluation of an estimator of an unknown parameter in statistics. We believe that putting equating methods in a general statistical model framework would be an interesting an useful next step in the area. Van der Linden (this issue) conceptual paper on equating is certainly an important contribution to this task.

A power analysis of seven normality tests against the Ex-Gaussian distribution (EGd) is presented. The EGd is selected on the basis that it is a particularly well-suited distribution to accommodate positively skewed distributions such as those observed in reaction times data. A pre-assessment of the power of the selected tests across various types of distributions was done via a meta-analysis and a comparison with other power analyses reported in the literature was also performed. General recommendations are given as to which tests should be used to test normality in data suspected to resemble an EG distribution. Additionally, some topics for future research regarding the use of confidence intervals and the computation of accurate critical values are outlined.

Using the concept of reduction by sufficiency of a Bayesian model, the issue of Bayesian identifiability is discussed. Various statements given in the literature on Bayesian identifiability are revised. Particular attention is put on the possibility of updating unidentified parameters. This issue is discussed under a general framework and also carefully illustrated in a fully discrete Bayesian model.

Bayesian methods have become increasingly popular in social sciences due to its flexibility in accommodating numerous models from different fields. The domain of item response theory is a good example of fruitful research, incorporating in the lasts years new developments and models, which are being estimated using the Bayesian approach. This is partly because of the availability of free software such as WinBUGS and R, which has permitted researchers to explore new possibilities.

In social network studies, most often only a single relation (or link) between the actors is investigated. When more than one link has been recorded, the twoway sociomatrix becomes a three-way array with the set of links being the third way. In this paper, we present a model which simultaneously accounts for the three ways in the data. Random effects are used to model the between-actor variability, both on senders and receivers side. In addition, structural relations between the linking variables are investigated. The model is applied to a study of popularity and strength in a class of students. It is shown that popularity can be seen as a linear function of strength on the receivers' side, but not on the senders' side.

The paper offers a general review of the basic concepts of both statistical model and parameter identification, and revisit the conceptual relationships between parameter identification and both parameter interpretability and properties of parameter estimates. All these issues are then exemplified for the 1PL, 2PL and 1PL-G fixed-effects models. For the 3PL model, however, we provide a theorem proving that the item parameters are not  identified, do not have an empirical interpretation and that it is not possible to obtain consistent and unbiased estimates of them

Equating is a family of statistical models and methods that are used to adjust scores on two or more versions of a test, so that the scores from dierent tests may be used interchangeably. In this paper we present the R package SNSequate which implements both standard and nonstandard statistical models and methods for test equating. The package construction was motivated by need of having a modular, simple, yet comprehensive, and general software that carries out traditional and new equating methods. SNSequate currently implements the traditional mean, linear and equipercentile equating methods, as well as the mean-mean, mean-sigma, Haebara and Stocking-Lord IRT linking methods. It also supports newest methods such as local equating, kernel equating, and IRT parameter linking methods based on asymmetric item characteristic functions. Practical examples are given to illustrate the capabilities of the software. A list of other programs for equating is presented, highlighting the main dierences between them. Future directions for the package are also discussed.

Equating methods utilize functions to transform scores on two or more
versions of a test, so that they can be compared and used interchangeably. In common
practice, traditional methods of equating use parametric models where, apart
from the test scores themselves, no additional information is used for the estimation
of the equating transformation. We propose a flexible Bayesian nonparametric
model for test equating which allows the use of covariates in the estimation of the
score distribution functions that lead to the equating transformation. A major feature
of this approach is that the complete shape of the score distribution may change
as a function of the covariates. As a consequence, the form of the equating transformation
can change according to covariate values. We discuss applications of the
proposed model to real and simulated data. We conclude that our method has good
performance compared to alternative approaches

Este estudio espera contribuir a la discusión sobre calidad docente en Chile a través de dos enfoques: el primero se basa en estimaciones de valor agregado del profesor o "efecto Profesor" en el aprendizaje. En él se explora el rol de las características del profesor y del contexto en que éste trabaja (escuela y municipio) en la estimación del efecto Profesor. El segundo enfoque intenta predecir el desempeño de los alumnos utilizando las variables de contexto y características del profesor. Con respecto al primer enfoque, los resultados muestran inestabilidad en las estimaciones de valor agregado, ante el ajuste de distintos modelos que incorporan covariables progresivamente. Por otra parte, el análisis de desempeño de los alumnos arroja como resultado que, de las variables de contexto, la escuela es más importante que el municipio para la predicción. A nivel del profesor, sólo resulta estadísticamente significativa la participación en perfeccionamientos profesionales, desestimándose el valor de la experiencia y de la formación inicial.