Item bias detection using loglinear irt

A method is proposed for the detection of item bias with respect to observed or unobserved subgroups. The method uses quasi-loglinear models for the incomplete subgroup × test score × Item 1 × ... × itemk contingency table. If subgroup membership is unknown the models are Haberman's incomplete-latent-class models.The (conditional) Rasch model is formulated as a quasi-loglinear model. The parameters in this loglinear model, that correspond to the main effects of the item responses, are the conditional estimates of the parameters in the Rasch model. Item bias can then be tested by comparing the quasi-loglinear-Rasch model with models that contain parameters for the interaction of item responses and the subgroups.The author thanks Wim J. van der Linden and Gideon J. Mellenbergh for comments and suggestions and Frank Kok for empirical data.     

Gender bias and construct validity in vocational interest measurement: Differential item functioning in the Strong Interest Inventory

Item response theory was used to address gender bias in interest measurement. Differential item functioning (DIF) technique, SIBTEST and DIMTEST for dimensionality, were applied to the items of the six General Occupational Theme (GOT) and 25 Basic Interest (BI) scales in the Strong Interest Inventory. A sample of 1860 women and 1105 men was used. The scales were not unidimensional and contain both primary and minor dimensions. Gender-related DIF was detected in two-thirds of the items. Item type (i.e., occupations, activities, school subjects, types of people) did not differ in DIF. A sex-type dimension was found to influence the responses of men and women differently. When the biased items were removed from the GOT scales, gender differences favoring men were reduced in the R and I scales but gender differences favoring women remained in the A and S scales. Implications for the development, validation and use of interest measures are discussed.     

Enhancing measurement validity in diverse populations: Modern approaches to evaluating differential item functioning

When developing and evaluating psychometric measures, a key concern is to ensure that they accurately capture individual differences on the intended construct across the entire population of interest. Inaccurate assessments of individual differences can occur when responses to some items reflect not only the intended construct but also construct-irrelevant characteristics, like a person's race or sex. Unaccounted for, this item bias can lead to apparent differences on the scores that do not reflect true differences, invalidating comparisons between people with different backgrounds. Accordingly, empirically identifying which items manifest bias through the evaluation of differential item functioning (DIF) has been a longstanding focus of much psychometric research. The majority of this work has focused on evaluating DIF across two (or a few) groups. Modern conceptualizations of identity, however, emphasize its multi-determined and intersectional nature, with some aspects better represented as dimensional than categorical. Fortunately, many model-based approaches to modelling DIF now exist that allow for simultaneous evaluation of multiple background variables, including both continuous and categorical variables, and potential interactions among background variables. This paper provides a comparative, integrative review of these new approaches to modelling DIF and clarifies both the opportunities and challenges associated with their application in psychometric research.     

Differential item functioning of the Rosenberg Self‐Esteem Scale in the US and China: Measurement bias matters

Using an item‐response theory‐based approach (i.e. likelihood ratio test with an iterative procedure), we examined the equivalence of the Rosenberg Self‐Esteem Scale (RSES) in a sample of US and Chinese college students. Results from the differential item functioning (DIF) analysis showed that the RSES was not fully equivalent at the item level, as well as at the scale level. The two cultural groups did not use the scale comparably, with the US students showing more extreme responses than the Chinese students. Moreover, we evaluated the practical impact of DIF and found that cultural differences in average self‐esteem scores disappeared after the DIF was taken into account. In the present study, we discuss the implications of our findings for cross‐cultural research and provide suggestions for future studies using the RSES in China.     

A model-based standardization approach that separates true bias/DIF from group ability differences and detects test bias/DTF as well as item bias/DIF

A model-based modification (SIBTEST) of the standardization index based upon a multidimensional IRT bias modeling approach is presented that detects and estimates DIF or item bias simultaneously for several items. A distinction between DIF and bias is proposed. SIBTEST detects bias/DIF without the usual Type 1 error inflation due to group target ability differences. In simulations, SIBTEST performs comparably to Mantel-Haenszel for the one item case. SIBTEST investigates bias/DIF for several items at the test score level (multiple item DIF called differential test functioning: DTF), thereby allowing the study of test bias/DIF, in particular bias/DIF amplification or cancellation and the cognitive bases for bias/DIF.This research was partially supported by Office of Naval Research Cognitive and Neural Sciences Grant N0014-90-J-1940, 4421-548 and National Science Foundation Mathematics Grant NSF-DMS-91-01436. The research reported here is collaborative in every respect and the order of authorship is alphabetical. The assistance of Hsin-hung Li and Louis Roussos in conducting the simulation studies was of great help. Discussions with Terry Ackerman, Paul Holland, and Louis Roussos were very helpful.     

On the power of the test for cluster bias

Cluster bias refers to measurement bias with respect to the clustering variable in multilevel data. The absence of cluster bias implies absence of bias with respect to any cluster‐level (level 2) variable. The variables that possibly cause the bias do not have to be measured to test for cluster bias. Therefore, the test for cluster bias serves as a global test of measurement bias with respect to any level 2 variable. However, the validity of the global test depends on the Type I and Type II error rates of the test. We compare the performance of the test for cluster bias with the restricted factor analysis (RFA) test, which can be used if the variable that leads to measurement bias is measured. It appeared that the RFA test has considerably more power than the test for cluster bias. However, the false positive rates of the test for cluster bias were generally around the expected values, while the RFA test showed unacceptably high false positive rates in some conditions. We conclude that if no significant cluster bias is found, still significant bias with respect to a level 2 violator can be detected with an RFA model. Although the test for cluster bias is less powerful, an advantage of the test is that the cause of the bias does not need to be measured, or even known.     

Measurement bias and error correction in a two-stage estimation for multilevel IRT models

Among current state-of-the-art estimation methods for multilevel IRT models, the two-stage divide-and-conquer strategy has practical advantages, such as clearer definition of factors, convenience for secondary data analysis, convenience for model calibration and fit evaluation, and avoidance of improper solutions. However, various studies have shown that, under the two-stage framework, ignoring measurement error in the dependent variable in stage II leads to incorrect statistical inferences. To this end, we proposed a novel method to correct both measurement bias and measurement error of latent trait estimates from stage I in the stage II estimation. In this paper, the HO-IRT model is considered as the measurement model, and a linear mixed effects model on overall (i.e., higher-order) abilities is considered as the structural model. The performance of the proposed correction method is illustrated and compared via a simulation study and a real data example using the National Educational Longitudinal Survey data (NELS 88). Results indicate that structural parameters can be recovered better after correcting measurement biases and errors.     

Assessing the Robustness of Mixture Models to Measurement Noninvariance

Abstract

Recent work reframes direct effects of covariates on items in mixture models as differential item functioning (DIF) and shows that, when present in the data but omitted from the fitted latent class model, DIF can lead to overextraction of classes. However, less is known about the effects of DIF on model performance—including parameter bias, classification accuracy, and distortion of class-specific response profiles—once the correct number of classes is chosen. First, we replicate and extend prior findings relating DIF to class enumeration using a comprehensive simulation study. In a second simulation study using the same parameters, we show that, while the performance of LCA is robust to the misspecification of DIF effects, it is degraded when DIF is omitted entirely. Moreover, the robustness of LCA to omitted DIF differs widely based on the degree of class separation. Finally, simulation results are contextualized by an empirical example.     

Consequences of Ignoring Guessing Effects on Measurement Invariance Analysis

Pseudo-guessing parameters are present in item response theory applications for many educational assessments. When sample size is not sufficiently large, the guessing parameters may be ignored from the analysis. This study examines the impact of ignoring pseudo-guessing parameters on measurement invariance analysis, specifically, on item difficulty, item discrimination, and mean and variance of ability distribution. Results show that when non-zero guessing parameters are ignored from the measurement invariance analysis, item discrimination estimates tend to decrease particularly for more difficult items, and item difficulty estimates decrease unless the items are highly discriminating and difficult. As the guessing parameter increases, the size of the decrease in item discrimination and difficulty tends to increase, and the estimated mean and variance of ability distribution tend to be inaccurate. When two groups have heterogeneous ability distributions, ignoring the guessing parameter affects the reference group and the focal group differently. Implications of result findings are discussed.     

Testing Differential Item Functioning in Small Samples

Abstract

Differential item functioning (DIF) is a pernicious statistical issue that can mask true group differences on a target latent construct. A considerable amount of research has focused on evaluating methods for testing DIF, such as using likelihood ratio tests in item response theory (IRT). Most of this research has focused on the asymptotic properties of DIF testing, in part because many latent variable methods require large samples to obtain stable parameter estimates. Much less research has evaluated these methods in small sample sizes despite the fact that many social and behavioral scientists frequently encounter small samples in practice. In this article, we examine the extent to which model complexity—the number of model parameters estimated simultaneously—affects the recovery of DIF in small samples. We compare three models that vary in complexity: logistic regression with sum scores, the 1-parameter logistic IRT model, and the 2-parameter logistic IRT model. We expected that logistic regression with sum scores and the 1-parameter logistic IRT model would more accurately estimate DIF because these models yielded more stable estimates despite being misspecified. Indeed, a simulation study and empirical example of adolescent substance use show that, even when data are generated from / assumed to be a 2-parameter logistic IRT, using parsimonious models in small samples leads to more powerful tests of DIF while adequately controlling for Type I error. We also provide evidence for minimum sample sizes needed to detect DIF, and we evaluate whether applying corrections for multiple testing is advisable. Finally, we provide recommendations for applied researchers who conduct DIF analyses in small samples.     

认知诊断框架下多级评分题目的DIF检测方法及其应用

本文对多级计分认知诊断测验的DIF概念进行了界定,并通过模拟实验以及实证研究对四种常见的多级计分DIF检验方法的适用性进行理论以及实践性的探索。研究结果表明：四种方法均能对多级计分认知诊断中的DIF进行有效的检验,且各方法的表现受模型的影响不大;相较于以总分为匹配变量,以KS为匹配变量时更利于DIF的检测;以KS为匹配变量的LDFA方法以及以KS为匹配变量的曼特尔检验方法在检测DIF题目时有着最高的检验力。     

分类数据测量等价性检验方法及其比较：项目阈值(难度)参数的组间差异性检验

测量工具满足等价性是进行多组比较的前提, 测量等价性的检验方法主要有基于CFA的多组比较法和基于IRT的DIF检验两类方法。文章比较了单维测验情境下基于CCFA的DIFFTEST检验方法和基于IRT模型的IRT-LR检验方法, 以及多维测验情境下DIFFTEST和基于MIRT的卡方检验方法的差异。通过模拟研究的方法, 比较了几种方法的检验力和第一类错误, 并考虑了样本总量、样本量的组间均衡性、测验长度、阈值差异大小以及维度间相关程度的影响。研究结果表明：(1)在单维测验下, IRT-LR是比DIFFTEST更为严格的检验方法; 多维测验下, 在测验较长、测验维度之间相关较高时, MIRT-MG比DIFFTEST更容易检验出项目阈值的差异, 而在测验长度较短、维度之间相关较小时, DIFFTEST的检验力反而略高于MIRT-MG方法。(2)随着阈值差值增加, DIFFTEST、IRT-LR和MIRT-MG三种方法的检验力均在增加, 当阈值差异达到中等或较大时, 三种方法都可以有效检验出测验阈值的不等价性。(3)随着样本总量增加, DIFFTEST、IRT-LR和MIRT-MG方法的检验力均在增加; 在总样本量不变, 两组样本均衡情况下三种方法的检验力均高于不均衡的情况。(4)违背等价性题目个数不变时, 测验越长DIFFTEST的检验力会下降, 而IRT-LR和MIRT-MG检验力则上升。(5) DIFFTEST方法的一类错误率平均值接近名义值0.05; 而IRT-LR和MIRT-MG方法的一类错误率平均值远低于0.05。     

Assessment of Differential Rater Functioning in Latent Classes with New Mixture Facets Models

Multifaceted data are very common in the human sciences. For example, test takers' responses to essay items are marked by raters. If multifaceted data are analyzed with standard facets models, it is assumed there is no interaction between facets. In reality, an interaction between facets can occur, referred to as differential facet functioning. A special case of differential facet functioning is the interaction between ratees and raters, referred to as differential rater functioning (DRF). In existing DRF studies, the group membership of ratees is known, such as gender or ethnicity. However, DRF may occur when the group membership is unknown (latent) and thus has to be estimated from data. To solve this problem, in this study, we developed a new mixture facets model to assess DRF when the group membership is latent and we provided two empirical examples to demonstrate its applications. A series of simulations were also conducted to evaluate the performance of the new model in the DRF assessment in the Bayesian framework. Results supported the use of the mixture facets model because all parameters were recovered fairly well, and the more data there were, the better the parameter recovery.     

The input bias: The misuse of input information in judgments of outcomes

In this paper we identify an input bias, the systematic misuse of input information in judgments of outcome quality. In many settings irrelevant input measures, such as the amount of time an employee spends in the office, influence outcome assessments, such as performance reviews. Across four studies we find that input values subtly, but significantly distort judgments of outcome quality. Irrelevant input information predictably influences outcome assessments even when people recognize that input measures should not matter and believe that input information did not matter. We examine the mechanics of the input bias, and suggest that because input measures are often easy to manipulate or misrepresent, the input bias is likely to have broad implications for managerial judgment and decision making.     

Evaluation on types of invariance in studying extreme response bias with an IRTree approach

In recent years, item response tree (IRTree) approaches have received increasing attention in the response style literature for their ability to partial out response style latent variables as well as associated item parameters. When an IRTree approach is adopted to measure extreme response styles, directional and content invariance could be assumed at the latent variable and item parameter levels. In this study, we propose to evaluate the empirical validity of these invariance assumptions by employing a general IRTree model with relaxed invariance assumptions. This would allow us to examine extreme response biases, beyond extreme response styles. With three empirical applications of the proposed evaluation, we find that relaxing some of the invariance assumptions improves the model fit, which suggests that not all assumed invariances are empirically supported. Specifically, at the latent variable level, we find reasonable evidence for directional invariance but mixed evidence for content invariance, although we also find that estimated correlations between content-specific extreme response latent variables are high, hinting at the potential presence of a general extreme response tendency. At the item parameter level, we find no directional or content invariance for thresholds and no content invariance for slopes. We discuss how the variant item parameter estimates obtained from a general IRTree model can offer useful insight to help us understand response bias related to extreme responding measured within the IRTree framework.     

情绪预测偏差的成因及干预

情绪预测偏差是一种对未来事件发生时情绪反应的预测和实际体验之间的分离现象。梳理该领域最近10年(2009~2019年)研究文献可知, 热点研究主题涉及偏差的现象、成因和干预, 相应呈现为三点主要发现: 情绪预测偏差十分普遍, 情绪预测偏差成因多源, 情绪预测偏差可以干预。未来研究应着力关注情绪预测偏差的发生模式和心理机制, 着力揭示具体偏差的神经心理机制, 并从进化和文化视角综合考察偏差的发生机制。     

Differential Item Functioning of the Emotional and Behavioral Screener for Caucasian and African American Elementary School Students

The present study examined the psychometric properties of a universal screening instrument called the Emotional and Behavioral Screener (EBS), which is designed to identify students exhibiting emotional and behavioral problems. The primary purposes of this study were to assess the measurement invariance of EBS items between Caucasian and African-American students and to assess the impact of differential item functioning (DIF) on EBS scores. The sample consisted of 946 elementary students from throughout the U.S. The findings suggested that EBS items exhibited small to negligible levels of DIF, and that DIF did not significantly impact EBS scores. The results supported the EBS as universal screening instrument that is fair in measuring the emotional and behavioral risk of elementary students. Research limitations and implications for school professionals are discussed.