No abstract

This paper focuses on two likelihood-based indices of person fit, the index lz and the Snijders’s modified index lz *. The first one is commonly used in practical assessment of person fit, although its asymptotic standard normal distribution is not valid when true abilities are replaced by sample ability estimates. The lz * index is a generalization of lz , which corrects for this sampling variability. Surprisingly, it is not yet popular in the psychometric and educational assessment community. Moreover, there is some ambiguity about which type of item response model and ability estimation method can be used to compute the lz * index. The purpose of this article is to present the index lz * in a simple and didactic approach. Starting from the relationship between lz and lz *, we develop the framework according to the type of logistic item response theory (IRT) model and the likelihood-based estimators of ability. The practical calculation of lz * is illustrated by analyzing a real data set about language skill assessment.

We present an extension of the logistic regression procedure to identify dichotomous differential item functioning (DIF) in the presence of more than two groups of respondents. Starting from the usual framework of a single focal group, we propose a general approach to estimate the item response functions in each group and to test for the presence of uniform DIF, nonuniform DIF, or both. This generalized procedure is compared to other existing DIF methods for multiple groups with a real data set on language skill assessment. Emphasis is put on the flexibility, completeness, and computational easiness of the generalized method.

Des chercheurs discutaient récemment de l’importance de rester à jour au sujet des plus récentes avancées en méthodes quantitatives. À ce titre, de nombreux auteurs ont exposé leur souhait de voir les chercheurs abandonner le populaire coefficient alpha de Cronbach. C’est dans une optique de diffusion et de vulgarisation que ce court article a comme objectif de présenter l’alternative qui semble la plus prometteuse pour mesurer la fidélité d’un test, le coefficient omega de McDonald, qui est basée sur l’analyse factorielle à un facteur commun.

The scientific treatment of missing data has been the subject of research for nearly a century. Strangely, interest in missing data is quite new in the fields of educational science and psychology (Peugh & Enders, 2004;Schafer & Graham, 2002). It is now important to better understand how various common methods for dealing with missing data can affect widely-used psychometric coefficients. The purpose of this study is to compare the impact of ten common fill-in methods on Cronbach's α (Cronbach, 1951). We use simulation studies to investigate the behavior of α in various situations. Our results show that multiple imputation is the most effective method. Furthermore, simple imputation methods like Winer imputation, item mean, and total mean are interesting alternatives for specific situations. These methods can be easily used by non-statisticians such as teachers and school psychologists.