References of "Magis, David"
     in
Bookmark and Share    
Full Text
Peer Reviewed
See detailIdentification of differential item functioning in multiple-group settings: A multivariate outlier detection approach
Magis, David ULg; De Boeck, Paul

in Multivariate Behavioral Research (2011), 46

This paper focuses on the identification of differential item functioning (DIF) when more than two groups of examinees are considered. We propose to consider items as elements of a multivariate space ... [more ▼]

This paper focuses on the identification of differential item functioning (DIF) when more than two groups of examinees are considered. We propose to consider items as elements of a multivariate space, where DIF items are outlying elements. Following this approach, the situation of multiple groups is a quite natural case. A robust statistics technique is proposed to identify DIF items as outliers in the multivariate space. For low dimensionalities, up to two three groups, also a simple graphical tool is derived. We illustrate our approach with a re-analysis of data from Kim, Cohen, and Park (1995) on using calculators for a mathematics test. [less ▲]

Detailed reference viewed: 30 (5 ULg)
Full Text
Peer Reviewed
See detailA generalized logistic regression procedure to detect differential item functioning among multiple groups
Magis, David ULg; Raîche, Gilles; Béland, Sébastien et al

in International Journal of Testing (2011), 11

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 ... [more ▼]

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, non uniform 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. [less ▲]

Detailed reference viewed: 20 (2 ULg)
Full Text
Peer Reviewed
See detailBeyond matching on the mean in developmental disabilities research
Facon, Bruno; Magis, David ULg; Belmont, John M

in Research in Developmental Disabilities (2011), 32

The matching of groups is a traditional way to control for confounding variables in developmental disability research. The equivalency of means across groups is routinely checked for these variables, but ... [more ▼]

The matching of groups is a traditional way to control for confounding variables in developmental disability research. The equivalency of means across groups is routinely checked for these variables, but not the homogeneity of their variances or the shapes of their distributions. In the present paper, it is argued that group matching can go seriously wrong unless it directly confronts the distributional concerns by the use of well-known statistical indices and very simple graphical displays of the distributions. The question of the equivalency of item response profiles is also addressed since two participants or two groups of participants can obtain the same overall score on the matching variable by passing different items. In this case, the matching cannot be considered satisfactory because of poor concordance between the molar (overall score) and molecular (item scores) levels of matching. Angoff's Delta plot method, a statistical approach for detecting differential item functioning across small groups is described. It is promising as a simple way to prove whole test/individual item correspondence and, in addition, a useful tool for making post-hoc statistical analyses at the item level on the dependant variables. [less ▲]

Detailed reference viewed: 9 (1 ULg)
Full Text
Peer Reviewed
See detailcatR: an R package for computerized adaptive testing
Magis, David ULg; Raîche, Gilles

in Applied Psychological Measurement (2011), 35

Computerized adaptive testing (CAT) is an active current research field in psychometrics and educational measurement. However, there is very little software available to handle such adaptive tasks. The R ... [more ▼]

Computerized adaptive testing (CAT) is an active current research field in psychometrics and educational measurement. However, there is very little software available to handle such adaptive tasks. The R package catR was developed to perform adaptive testing with as much flexibility as possible, in an attempt to provide a developmental and testing platform to the interested user. Several item-selection rules and ability estimators are implemented. The item bank can be provided by the user or randomly generated from parent distributions of item parameters. Three stopping rules are available. The output can be graphically displayed. [less ▲]

Detailed reference viewed: 25 (6 ULg)
See detailDes mécanismes pour assurer la validité de l’interprétation de la mesure en éducation. Tome 2: l'évaluation.
Raîche, Gilles; Paquette-Côté, Karine; Magis, David ULg

Book published by Presses de l'Université du Québec (2011)

Detailed reference viewed: 7 (2 ULg)
See detailDes mécanismes pour assurer la validité de l’interprétation de la mesure en éducation. Tome 1 : la mesure.
Raîche, Gilles; Paquette-Côté, Karine; Magis, David ULg

Book published by Presses de l'Université du Québec (2011)

Detailed reference viewed: 8 (3 ULg)
See detailComparaison empirique des méthodes classiques de détection du fonctionnement différentiel d’items en psychométrie
Magis, David ULg; De Boeck, Paul; Raîche, Gilles

in Raîche, Gilles; Paquette-Côté, Karine; Magis, David (Eds.) Des mécanismes pour assurer la validité de l’interprétation de la mesure en éducation. Tome 1 : la mesure. (2011)

Detailed reference viewed: 20 (12 ULg)
See detailPour assurer la validité de l'interprétation de la mesure en éducation: aspects pratiques
Magis, David ULg; Raîche, Gilles; Paquette-Côté, Karine

in Raîche, Gilles; Paquette-Côté, Karine; Magis, David (Eds.) Des mécanismes pour assurer la validité de l'interprétation de la mesure en éducation. Tome 2: l'évaluation. (2011)

Detailed reference viewed: 11 (6 ULg)
See detailÉtude du fonctionnement différentiel d'item dans le cadre du test de classement en anglais-langue seconde (TCALS-II)
Béland, Sébastien; Magis, David ULg; Raîche, Gilles

in Blais, Jean-Guy; Gilles, Jean-Luc (Eds.) Evaluation des apprentissages et technologies de l'information et de la communication : Le futur est à notre porte (2011)

Detailed reference viewed: 12 (0 ULg)
See detailConditions d'efficacité de la détection des patrons de réponses inappropriés lors de l'administration d'épreuves adaptatives
Raîche, Gilles; Magis, David ULg; Béland, Sébastien et al

in Blais, Jean-Guy; Gilles, Jean-Luc (Eds.) Evaluation des apprentissages et technologies de l'information et de la communication : Le futur est à notre porte (2011)

Detailed reference viewed: 4 (0 ULg)
Full Text
Peer Reviewed
See detailDo Raven’s Colored Progressive Matrices function in the same way in typical and clinical populations? Insights from the intellectual disability field
Facon, Bruno; Magis, David ULg; Nuchadee, Marie-Laure et al

in Intelligence (2011), 39

Standardized tests are used widely in comparative studies of clinical populations, either as dependent or control variables. Yet, one cannot always be sure that the test items measure the same constructs ... [more ▼]

Standardized tests are used widely in comparative studies of clinical populations, either as dependent or control variables. Yet, one cannot always be sure that the test items measure the same constructs in the groups under study. In the present work, 460 participants with intellectual disability of undifferentiated etiology and 488 typical children were tested using Raven’s Colored Progressive Matrices (RCPM). Data were analyzed using binomial logistic regression modeling designed to detect differential item functioning (DIF). Results showed that 12 items out of 36 function differentially between the two groups, but only 2 items exhibit at least moderate DIF. Thus, a very large majority of the items have identical discriminative power and difficulty levels across the two groups. It is concluded that RCPM can be used with confidence in studies comparing participants with and without intellectual disability. In addition, it is suggested that methods for investigating internal bias of tests used in cross-cultural, cross-linguistic or cross gender comparisons should also be regularly employed in studies of clinical populations, particularly in the field of developmental disability, to show the absence of systematic measurement error (i.e. DIF) affecting item responses. [less ▲]

Detailed reference viewed: 25 (2 ULg)
Full Text
Peer Reviewed
See detailA test-length correction to the estimation of extreme proficiency levels
Magis, David ULg; Béland, Sébastien; Raîche, Gilles

in Applied Psychological Measurement (2011), 35

In this paper, the estimation of extremely large or extremely small proficiency levels, given the item parameters of a logistic item response model, is investigated. On the one hand, the estimation of ... [more ▼]

In this paper, the estimation of extremely large or extremely small proficiency levels, given the item parameters of a logistic item response model, is investigated. On the one hand, the estimation of proficiency levels by maximum likelihood (ML), despite being asymptotically unbiased, may yield infinite estimates. On the other hand, with an appropriate prior distribution, the Bayesian approach of maximum a posteriori (MAP) yields finite estimates, but it suffers from severe estimation bias at the extremes of the proficiency scale. In a first step, we propose a simple correction to the MAP estimator in order to reduce this estimation bias. The correction factor is determined through a simulation study and depends only on the length of the test. In a second step, some additional simulations highlight that the corrected estimator behaves like the ML estimator and outperforms the standard MAP method for extremely small or extremely large abilities. Although based on the Rasch model, the method could be adapted to other logistic item response models. [less ▲]

Detailed reference viewed: 15 (4 ULg)
See detailLa correction d'indices de détection de patrons de réponses inappropriés paramétriques: une étude de type Monte Carlo
Béland, Sébastien; Raîche, Gilles; Magis, David ULg

Conference (2010, November)

L’une des conditions à respecter afin d’interpréter adéquatement les scores obtenus à un test est de s'assurer que les sujets ont répondu selon leur habileté réelle. Or, il se peut que certains individus ... [more ▼]

L’une des conditions à respecter afin d’interpréter adéquatement les scores obtenus à un test est de s'assurer que les sujets ont répondu selon leur habileté réelle. Or, il se peut que certains individus tentent de se surclasser en trichant ou de se sous-classer intentionnellement. Par chance, de nombreuses méthodes ont été développées pour détecter les individus qui adoptent de telles stratégies. Dans le cadre de cette présentation, nous nous intéresserons plus particulièrement aux indices de détection de patrons de réponses inappropriés. L’utilisation des indices de détection paramétriques comporte de nombreuses limites (Meijer et Sijtsma, 2001). Par exemple, nous savons que plusieurs indices ne peuvent pas être interprétés à partir des quantiles d’une loi de distribution connue. De plus, il est connu que l’utilisation d’une estimation de l’habileté peut grandement affecter la distribution connue de certains indices. Afin de dépasser les problèmes précédents, Snijders (2001) a développé une approche permettant de corriger les indices de détection de patrons de réponses inappropriés lorsqu'une estimation de l’habileté est considérée. Les indices ainsi corrigés suivent asymptotiquement une loi normale standard, ce qui permet la définition de seuils de détection de patrons inappropriés. Nous présenterons comment cette correction permet d'améliorer l'identification des patrons de réponses problématiques, notamment grâce à une diminution de l'écart entre l’erreur de type I empirique et le niveau de signification nominal. Une approche par simulation de données de type Monte Carlo sera présentée, et différents estimateurs de l'habileté (maximum de vraisemblance, maximum de vraisemblance a posteriori, maximum de vraisemblance pondérée) seront utilisés. [less ▲]

Detailed reference viewed: 17 (1 ULg)
See detailDes modèles psychométriques pour corriger les résultats dans des enquêtes malgré les tentatives de fraude
Raîche, Gilles; Béland, Sébastien; Magis, David ULg

Conference (2010, November)

Dans le contextes des enquêtes, il arrive fréquemment que les répondants ne fournissent pas nécessairement des réponses appropriées. C'est le cas notamment en ce qui a trait aux cas de fraudes dans les ... [more ▼]

Dans le contextes des enquêtes, il arrive fréquemment que les répondants ne fournissent pas nécessairement des réponses appropriées. C'est le cas notamment en ce qui a trait aux cas de fraudes dans les épreuves d'évaluation des habiletés disciplinaires. À cette fin, des modèles psychométriques ont été proposés pour corriger le résultat d'un étudiant malgré ses tentatives de fraudes. Ces modèles reposent toutefois uniquement sur des réponses de type dichotomique ou seulement une bonne ou une mauvaise réponse est considérée. Ces comportements de fraude sont également présents à l'intérieur des enquêtes d'opinion qui s'intéressent aux perceptions des répondants. La plupart du temps les réponses à ces enquêtes présentent plus d'un choix de réponses et sont alors dits polytomiques. Récemment, de nouvelles modélisations appropriées pour ce type de réponses ont été proposées. Elles seront présentées et accompagnées d'exemples d'application. [less ▲]

Detailed reference viewed: 13 (2 ULg)
Full Text
See detailIdentification of differential item functioning by logistic regression: application to intellectual disability data
Magis, David ULg; Facon, Bruno

Conference (2010, October 14)

The purpose of this talk is to briefly present the logistic regression procedure to identify differential itm functioning (DIF) in psychometrics, and to illustrate the method on a practical data set about ... [more ▼]

The purpose of this talk is to briefly present the logistic regression procedure to identify differential itm functioning (DIF) in psychometrics, and to illustrate the method on a practical data set about intellectual disability. DIF is an important research fi eld and many statistical methods have been developed in this framework. Logistic regression has the asset of not being based on item response models, and it can detect both uniform and nonuniform DIF. In addition, it can be applied to the comparison of more than two groups. This method is illustrated by comparing two groups of French pupils, one group of normal ability students and one group with intellectually disabled students. Raven's Colored Progressive Matrices test is assigned and DIF is investigated between the two groups. It is concluded that none of the items can be flagged as functioning differently. [less ▲]

Detailed reference viewed: 25 (4 ULg)
Full Text
See detailOld and new issues in differential item functioning (DIF): How to overcome current research conservatism?
Magis, David ULg

Scientific conference (2010, October 05)

In the context of differential item functioning (DIF), some issues have been pointed out more than 20 years ago, whereas other ones have appeared much more recently (often due to an incredible increase of ... [more ▼]

In the context of differential item functioning (DIF), some issues have been pointed out more than 20 years ago, whereas other ones have appeared much more recently (often due to an incredible increase of available information). In addition, there is some uprising evidence that the topic of DIF is becoming less attractive for journal editors. Explaining the reasons for this lack of interest in DIF, and why it should be actively countered, are the main goals of this talk. Three specific issues will be discussed: (a) the supremacy of the Mantel-Haenszel method, (b) the problem of item purification, and (c) the identification of DIF among multiple groups. Some recent and ongoing research results will be discussed to highlight why DIF should not be left alone in the years to come. [less ▲]

Detailed reference viewed: 18 (0 ULg)
See detailMultidimensional fluctuation, pseudo-guessing and carelessness IRT person parameters with polytomous item response models
Raîche, Gilles; Magis, David ULg; Béland, Sébastien

Conference (2010, July)

Frequently, candidates at aptitude multiple choice test miss attention, motivation or preparation and show underachievement or overachievement. Similarly, at surveys respondents show frequently misfitting ... [more ▼]

Frequently, candidates at aptitude multiple choice test miss attention, motivation or preparation and show underachievement or overachievement. Similarly, at surveys respondents show frequently misfitting patterns of responses. Their result does not correspond any more to their true aptitude or attitude, an inappropriate response pattern being obtained. New multidimensional models specific to polytomous responses circumvent these situations and diminish considerably the associated person bias. Multidimensional polytomous item response models adding new person parameters to the trait of the candidate are proposed. In the spirit of previous Raiche’s dichotomous responses IRT propositions, like the discrimination and don’t know item parameters, these models offer fluctuation and don’t know person parameters. Estimation methods, results from simulation showing the efficacy of these models and recommendations for the design of testing situations will be presented. [less ▲]

Detailed reference viewed: 59 (5 ULg)
Full Text
See detailOutlier identification of differential item functioning in multiple groups
Magis, David ULg; De Boeck, Paul

Conference (2010, July)

Differential item functioning (DIF) has received increasing focus in the past decades. Recently, Magis and De Boeck (2009) proposed to identify differentially functioning items as outliers in a one ... [more ▼]

Differential item functioning (DIF) has received increasing focus in the past decades. Recently, Magis and De Boeck (2009) proposed to identify differentially functioning items as outliers in a one-dimensional space of DIF measures, using robust statistical tools for outlier identification. The purpose of this talk is to present an extension of this approach for the case of more than one focal group. In this multiple group framework, items can be characterized by multiple vectors of DIF measures, one for each focal group, so that a multivariate DIF space is obtained. DIF items can then be identified as outliers in this multivariate space, based on robust multivariate estimators of location and dispersion. The MCD (Minimum Covariance Determinant) estimator is shown to be adequate for this purpose. A major asset of the method that it can rely on existing DIF indices to define the DIF vectors, and that it does not need a purification step. Alternatively, it can be used to determine on an anchor set. The method will be illustrated by an example about calculator effects on mathematics test items. [less ▲]

Detailed reference viewed: 27 (2 ULg)
Full Text
See detailDetection of differential item functioning with the difR package
Magis, David ULg; Béland, Sébastien; Tuerlinckx, Francis et al

Poster (2010, July)

The purpose of this talk is to briefly present the package difR of the software R. This package permits to identify differential item functioning (DIF) with several parametric or nonparametric methods ... [more ▼]

The purpose of this talk is to briefly present the package difR of the software R. This package permits to identify differential item functioning (DIF) with several parametric or nonparametric methods. Nonparametric methods include: Mantel-Haenszel, standardization, logistic regression and Breslow-Day. Available parametric methods are the Lord’s chi-square test, the Raju’s area method, and the likelihood-ratio test. Some methods for DIF identification in multiple groups (generalized Mantel-Haenszel, generalized Lord’s test) are also available. The presentation focuses on the package and its potential, rather than on the included methods. Its performance will be illustrated with an example data set. [less ▲]

Detailed reference viewed: 71 (4 ULg)
See detailUsing Snijders’ approach to standardize some parametric person fit indexes with estimated ability levels: a simulation study
Béland, Sébastien; Magis, David ULg; Raîche, Gilles

Conference (2010, July)

Several authors (Molenaar & Hoijtink, 1990; Meijer & Sijtsma, 2001) have shown that several person-fit statistics present some important limitations. One issue is that the distribution of some parametric ... [more ▼]

Several authors (Molenaar & Hoijtink, 1990; Meijer & Sijtsma, 2001) have shown that several person-fit statistics present some important limitations. One issue is that the distribution of some parametric person-fit statistics is unknown. Another important issue is that the distribution of person-fit indexes is most often derived under the true ability level. In this situation, replacing the true ability by some estimate can seriously affect the distribution of these indexes. Snijders (2001) proposed a method to correct the mean and the variance of many parametric person-fit statistics to be approximately standard normally distributed, and derived the corrected version of the lz index. The purpose of this paper is to apply this correction to other well-known parametric indexes, and to compare them with their classical versions. The simulation results indicate that the standardized indexes have empirical type I errors close to the nominal significance level, and that the corrected indexes outperform their classical versions in this regard. [less ▲]

Detailed reference viewed: 35 (0 ULg)