David Magis

A general framework and an R package for the detection of dichotomous differential item functioning.

Differential item functioning (DIF) is an important issue of interest in psychometrics and educational measurement. Several methods have been proposed in recent decades for identifying items that function differently between two or more groups of examinees. Starting from a framework for classifying DIF detection methods and from a comparative overview of the most traditional methods, an R package for nine methods, called difR, is presented. The commands and options are briefly described, and the package is illustrated through the analysis of a data set on verbal aggression.

Effects of endotoxic shock on right ventricular systolic function and mechanical efficiency

OBJECTIVE To investigate the effects of endotoxin infusion on right ventricular (RV) systolic function and mechanical efficiency. METHODS Six anesthetized pigs (Endo group) received a 0.5 mg/kg endotoxin infusion over 30 min and were compared with six other anesthetized pigs (Control group) receiving placebo for 5 h. RV pressure-volume (PV) loops were obtained by the conductance catheter technique and pulmonary artery flow and pressure were measured with high-fidelity transducers. RESULTS RV adaptation to increased afterload during the early phase of endotoxin-induced pulmonary hypertension (T30) was obtained by both homeometric and hetereometric regulations: the slope of the end-systolic PV relationship of the right ventricle increased from 1.4+/-0.2 mmHg/ml to 2.9+/-0.4 mmHg/ml (P<0.05) and RV end-diastolic volume increased from 56+/-6 ml to 64+/-11 ml (P<0.05). Consequently, right ventricular-vascular coupling was maintained at a maximum efficiency. Ninety minutes later (T120), facing the same increased afterload, the right ventricle failed to maintain its contractility to such an elevated level and, as a consequence, right ventricular-vascular uncoupling occurred. PV loop area, which is known to be highly correlated with oxygen myocardial consumption, increased from 1154+/-127 mmHg/ml (T0) to 1798+/-122 mmHg/ml (T180) (P<0.05) while RV mechanical efficiency decreased from 63+/-2% (T0) to 45+/-5% (T270) (P<0.05). CONCLUSIONS In the very early phase of endotoxinic shock, right ventricular-vascular coupling is preserved by an increase in RV contractility. Later, myocardial oxygen consumption and energetic cost of RV contractility are increased, as evidenced by the decrease in RV efficiency, and right ventricular-vascular uncoupling occurs. Therefore, therapies aiming at restoring right ventricular-vascular coupling in endotoxic shock should attempt to increase RV contractility and to decrease RV afterload but also to preserve RV mechanical efficiency.

Alteration of Right Ventricular-Pulmonary Vascular Coupling in a Porcine Model of Progressive Pressure Overloading

In acute pulmonary embolism, right ventricular (RV) failure may result from exceeding myocardial contractile resources with respect to the state of vascular afterload. We investigated the adaptation of RV performance in a porcine model of progressive pulmonary embolism. Twelve anesthetized pigs were randomly divided into two groups: gradual pulmonary arterial pressure increases by three injections of autologous blood clot (n = 6) or sham-operated controls (n = 6). Right ventricular pressure-volume (PV) loops were recorded using a conductance catheter. Right ventricular contractility was estimated by the slope of the end-systolic PV relationship (Ees). After load was referred to as pulmonary arterial elastance (Ea) and assessed using a four-element Windkessel model. Right ventricular-arterial coupling (Ees/Ea) and efficiency of energy transfer (from PV area to external mechanical work [stroke work]) were assessed at baseline and every 30 min for 4 h. Eaincreased progressively after embolization, from 0.26 ± 0.04 to 2.2 ± 0.7 mmHg mL−1 (P < 0.05). Ees increased from 1.01 ±0.07 to 2.35 ± 0.27 mmHg mL−1 (P < 0.05) after the first two injections but failed to increase any further. As a result, Ees/Ea initially decreased to values associated with optimal SW, but the last injection was responsible for Ees/Ea values less than 1, decreased stroke volume, and RV dilation. Stroke work/PV area consistently decreased with each injection from 79% ± 3% to 39% ± 11% (P < 0.05). In response to gradual increases in afterload, RV contractility reserve was recruited to a point of optimal coupling but submaximal efficiency. Further afterload increases led to RV-vascular uncoupling and failure.

A Didactic Presentation of Snijders’s lz* Index of Person Fit With Emphasis on Response Model Selection and Ability Estimation:

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.

A generalized logistic regression procedure to detect differential item functioning among multiple groups

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.

Beyond matching on the mean in developmental disabilities research

The matching of groups is a traditional way to control for confounding variables in developmental disabilities 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. Angoffs 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 dependent variables.

The sentence repetition task: A powerful diagnostic tool for French children with specific language impairment

This study assesses the diagnostic accuracy and construct validity of a sentence repetition task that is commonly used for the identification of French children with specific language impairment (SLI). Thirty-four school-aged children with a confirmed, diagnostically based diagnosis of SLI, and 34 control children matched on age and nonverbal abilities performed the sentence repetition task. Two general scoring measures took into account the verbatim repetition of the sentence and the number of words accurately repeated. Moreover, five other scoring measures were applied to their answers in order to separately take into account their respect of lexical items, functional items, syntax, verb morphology, and the general meaning of the sentence. Results show good to high levels of sensitivity and specificity at the three cut-off points for all scoring measures. A principal component analysis revealed two factors. Scoring measures for the respect of functional words, syntax and verb morphology provided the largest loadings to the first factor, while scoring measures for the respect of lexical words and general semantics provided the largest loadings to the second factor. Sentence repetition appears to be a valuable tool to identify SLI in French children, and the ability to repeat sentences correctly is supported by two factors: a morphosyntactic factor and a lexical factor.

A Note on the Item Information Function of the Four-Parameter Logistic Model

This article focuses on four-parameter logistic (4PL) model as an extension of the usual three-parameter logistic (3PL) model with an upper asymptote possibly different from 1. For a given item with fixed item parameters, Lord derived the value of the latent ability level that maximizes the item information function under the 3PL model. The purpose of this article is to extend this result to the 4PL model. A generic and algebraic method is developed for that purpose. The result is practically illustrated by an example and several potential applications of this result are outlined.

Angoff's delta method revisited: improving DIF detection under small samples.

Most methods for detecting differential item functioning (DIF) are suitable when the sample sizes are sufficiently large to validate the null statistical distributions. There is no guarantee, however, that they will still perform adequately when there are few respondents in the focal group or in both the reference and the focal group. Angoffs delta plot is a potentially useful alternative for small-sample DIF investigation, but it suffers from an improper DIF flagging criterion. The purpose of this paper is to improve this classification rule under mild statistical assumptions. This improvement yields a modified delta plot with an adjusted DIF flagging criterion for small samples. A simulation study was conducted to compare the modified delta plot with both the classical delta plot approach and the Mantel-Haenszel method. It is concluded that the modified delta plot is consistently less conservative and more powerful than the usual delta plot, and is also less conservative and more powerful than the Mantel-Haenszel method as long as at least one group of respondents is small.

Detection of Differential Item Functioning Using the Lasso Approach

This article proposes a novel approach to detect differential item functioning (DIF) among dichotomously scored items. Unlike standard DIF methods that perform an item-by-item analysis, we propose the “LR lasso DIF method”: logistic regression (LR) model is formulated for all item responses. The model contains item-specific intercepts, an effect of the sum score, and item-group interaction (i.e., DIF) effects, with a lasso penalty on all DIF parameters. Optimal penalty parameter selection is investigated through several known information criteria (Akaike information criterion, Bayesian information criterion, and cross validation) as well as through a newly developed alternative. A simulation study was conducted to compare the global performance of the suggested LR lasso DIF method to the LR and Mantel–Haenszel methods (in terms of false alarm and hit rates). It is concluded that for small samples, the LR lasso DIF approach globally outperforms the LR method, and also the Mantel–Haenszel method, especially in the presence of item impact, while it yields similar results with larger samples.