Detection of differential item functioning in Rasch models by boosting techniques

Detecting Differential Item Functioning in Multidimensional Graded Response Models With Recursive Partitioning

Differential item functioning (DIF) is a common challenge when examining latent traits in large scale surveys. In recent work, methods from the field of machine learning such as model-based recursive partitioning have been proposed to identify subgroups with DIF when little theoretical guidance and many potential subgroups are available. On this basis, we propose and compare recursive partitioning techniques for detecting DIF with a focus on measurement models with multiple latent variables and ordinal response data. We implement tree-based approaches for identifying subgroups that contribute to DIF in multidimensional latent variable modeling and propose a robust, yet scalable extension, inspired by random forests. The proposed techniques are applied and compared with simulations. We show that the proposed methods are able to efficiently detect DIF and allow to extract decision rules that lead to subgroups with well fitting models.

Yale Food Addiction Scale 2.0 (YFAS 2.0) and modified YFAS 2.0 (mYFAS 2.0): Rasch analysis and differential item functioning

BACKGROUND

Food addiction (FA) is a prevalent concern that may manifest as poorly controlled food consumption and promote overweight/obesity. Thus, having a well-established instrument for assessment may facilitate better prevention and treatment. The current study investigated the psychometric properties of two common measures of FA (i.e., the Yale Food Addiction Scale [YFAS] 2.0 and its modified version, mYFAS 2.0) using a robust statistical analysis (Rasch model).

METHODS

In this cross-sectional study, the scales were sent to 974 students studying in higher education (60% females) in Taiwan through online media including email and social networks. Rasch modeling was used to assess dimensionality, difficulty level, and item misfit and hierarchy. Differential item functioning (DIF) was performed to examine consistency of the items across gender and weight status.

RESULTS

Rasch analysis indicated 3 items of the 35 items belonging to the YFAS 2.0 (8.6%) and none belonging to the mYFAS 2.0 were misfit. Unidimensionality and construct validity of both scales were supported by appropriate goodness-of-fit for diagnostic criteria. The person separation was 3.14 (reliability = 0.91) for the YFAS 2.0 and 2.17 (reliability = 0.82) for mYFAS 2.0, indicating the scales could distinguish participants into more than 3 strata. Only one substantial DIF was found for diagnostic criteria of "Failure to fulfill major role obligation" in the YFAS 2.0 across gender.

CONCLUSION

According to Rasch modeling, both the YFAS 2.0 and mYFAS 2.0 have acceptable construct validity in Chinese-speaking youth. Scoring methods using either diagnostic criteria or symptom counts for both the YFAS 2.0 and mYFAS 2.0 are supported by the present Rasch findings.

Semi-automated Rasch analysis using in-plus-out-of-questionnaire log likelihood

Rasch analysis is a popular statistical tool for developing and validating instruments that aim to measure human performance, attitudes and perceptions. Despite the availability of various software packages, constructing a good instrument based on Rasch analysis is still considered to be a complex, labour-intensive task, requiring human expertise and rather subjective judgements along the way. In this paper we propose a semi-automated method for Rasch analysis based on first principles that reduces the need for human input. To this end, we introduce a novel criterion, called in-plus-out-of-questionnaire log likelihood (IPOQ-LL). On artificial data sets, we confirm that optimization of IPOQ-LL leads to the desired behaviour in the case of multi-dimensional and inhomogeneous surveys. On three publicly available real-world data sets, our method leads to instruments that are, for all practical purposes, indistinguishable from those obtained by Rasch analysis experts through a manual procedure.

Item-Focused Trees for the Detection of Differential Item Functioning in Partial Credit Models

Various methods to detect differential item functioning (DIF) in item response models are available. However, most of these methods assume that the responses are binary, and so for ordered response categories available methods are scarce. In the present article, DIF in the widely used partial credit model is investigated. An item-focused tree is proposed that allows the detection of DIF items, which might affect the performance of the partial credit model. The method uses tree methodology, yielding a tree for each item that is detected as DIF item. The visualization as trees makes the results easily accessible, as the obtained trees show which variables induce DIF and in which way. In the present paper, the new method is compared with alternative approaches and simulations demonstrate the performance of the method.

An Update on Statistical Boosting in Biomedicine

Statistical boosting algorithms have triggered a lot of research during the last decade. They combine a powerful machine learning approach with classical statistical modelling, offering various practical advantages like automated variable selection and implicit regularization of effect estimates. They are extremely flexible, as the underlying base-learners (regression functions defining the type of effect for the explanatory variables) can be combined with any kind of loss function (target function to be optimized, defining the type of regression setting). In this review article, we highlight the most recent methodological developments on statistical boosting regarding variable selection, functional regression, and advanced time-to-event modelling. Additionally, we provide a short overview on relevant applications of statistical boosting in biomedicine.

Item-focussed Trees for the Identification of Items in Differential Item Functioning

A novel method for the identification of differential item functioning (DIF) by means of recursive partitioning techniques is proposed. We assume an extension of the Rasch model that allows for DIF being induced by an arbitrary number of covariates for each item. Recursive partitioning on the item level results in one tree for each item and leads to simultaneous selection of items and variables that induce DIF. For each item, it is possible to detect groups of subjects with different item difficulties, defined by combinations of characteristics that are not pre-specified. The way a DIF item is determined by covariates is visualized in a small tree and therefore easily accessible. An algorithm is proposed that is based on permutation tests. Various simulation studies, including the comparison with traditional approaches to identify items with DIF, show the applicability and the competitive performance of the method. Two applications illustrate the usefulness and the advantages of the new method.