Compensatory and non‐compensatory multidimensional randomized item response models

Fox, Jean-Paul; Entink, Rinke Klein; Avetisyan, Marianna

doi:10.1111/bmsp.12012

Cited by 10 publications

(14 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, in their models, the means of LVs and variances and covariances among LVs do not depend on any individual information. In general, a multidimensional item response theory model makes it possible to investigate correlations between LVs, to test simultaneously effects of an explanatory variable on several LVs and to test for differential effects of explanatory variables on various LVs [22]. To address this issue, Fox et al [22] developed a multidimensional item response model with randomized response techniques for cross-sectional data.…”

Section: Introductionmentioning

confidence: 99%

“…In general, a multidimensional item response theory model makes it possible to investigate correlations between LVs, to test simultaneously effects of an explanatory variable on several LVs and to test for differential effects of explanatory variables on various LVs [22]. To address this issue, Fox et al [22] developed a multidimensional item response model with randomized response techniques for cross-sectional data. Klein Entink et al [23] and Fox and Marianti [24] jointly modeled the accuracy and speed of test takers using two MLIRT models (one for accuracy and one for speed).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multidimensional latent trait linear mixed model: an application in clinical studies with multivariate longitudinal outcomes

Wang

Luo

2017

Statist. Med.

View full text Add to dashboard Cite

Multilevel item response theory (MLIRT) models have been widely used to analyze the multivariate longitudinal data of mixed types (e.g., categorical and continuous) in clinical studies. The MLIRT models often have unidimensional assumption, i.e., the multiple outcomes are clinical manifestations of a univariate latent variable. However, the unidimensional assumption may be unrealistic because some diseases may be heterogeneous and characterized by multiple impaired domains with variable clinical symptoms and disease progressions. We relax this assumption and propose a multidimensional latent trait linear mixed model (MLTLMM) to allow multiple latent variables and within-item multidimensionality (one outcome can be a manifestation of more than one latent variable). We conduct extensive simulation studies to assess the unidimensional MLIRT model and the proposed MLTLMM model. The simulation studies suggest that the MLTLMM model outperforms unidimensional model when the multivariate longitudinal outcomes are manifested by multiple latent variables. The proposed model is applied to two motivating studies of amyotrophic lateral sclerosis (ALS): a clinical trial of ceftriaxone and the Pooled Resource Open-Access ALS Clinical Trials (PRO-ACT) database.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Multidimensional latent trait linear mixed model: an application in clinical studies with multivariate longitudinal outcomes

Wang

Luo

2017

Statist. Med.

View full text Add to dashboard Cite

show abstract

“…We restricted ourselves to these methods, because these are broadly applicable and yield correlation and regression coefficients that closely resemble those from standard statistical models. However, future versions of package RRreg could include multivariate RR methods for more specific situations such as computing sum scores of RR variables (Cruyff 2008), constructing scales comprised of RR items (Himmelfarb 2008), using RR in cross-sectional (Frank, Van den Hout, and Van der Heijden 2009) or cross-national studies (De Jong, Pieters, and Stremersch 2012), combining RR and propensity scoring (Gingerich 2010), or generalizing single and multidimensional item response theory to RR (Böckenholt and Van der Heijden 2007;Fox 2005;Fox, Entink, and Avetisyan 2014).…”

Section: Discussionmentioning

confidence: 99%

RRreg: An R Package for Correlation and Regression Analyses of Randomized Response Data

Heck¹,

Moshagen²

2018

J. Stat. Soft.

View full text Add to dashboard Cite

The randomized response (RR) technique was developed to improve the validity of measures assessing attitudes, behaviors, and attributes threatened by social desirability bias. The RR removes any direct link between individual responses and the sensitive attribute to maximize the anonymity of respondents and, in turn, to elicit more honest responding. Since multivariate analyses are no longer feasible using standard methods, we present the R package RRreg that allows for multivariate analyses of RR data in a user-friendly way. We show how to compute bivariate correlations, how to predict an RR variable in an adapted logistic regression framework (with or without random effects), and how to use RR predictors in a modified linear regression. In addition, the package allows for power analysis and robustness simulations. To facilitate the application of these methods, we illustrate the benefits of multivariate methods for RR variables using an empirical example.

show abstract

“…The α i was assumed to be independent of b i and β i for two reasons. First, previous studies (e.g., Bolt and Lall, 2003; Fox et al, 2014) indicate that the correlations between the time-discrimination α i and the other item parameters ( b i and β i ) provide negligible information about the item quality or person latent traits, especially the relationship between speed and accuracy among test takers. Thus, by following the convention of jointly estimating an RT model and an IRT model, the covariances related to time-discrimination α i were ignored.…”

Section: Methodsmentioning

confidence: 99%

How Do Test Takers Interact With Simulation-Based Tasks? A Response-Time Perspective

et al. 2019

View full text Add to dashboard Cite

Many traditional educational assessments use multiple-choice items and constructed-response items to measure fundamental skills. Virtual performance assessments, such as game- or simulation-based assessments, are designed recently in the field of educational measurement to measure more integrated skills through the test takers’ interactive behaviors within an assessment in a virtual environment. This paper presents a systematic timing study based on data collected from a simulation-based task designed recently at Educational Testing Service. The study is intended to understand the response times in complex simulation-based tasks so as to shed light on possible ways of leveraging response time information in designing, assembling, and scoring of simulation-based tasks. To achieve this objective, a series of five analyses were conducted to first understand the statistical properties of the timing data, and then investigate the relationship between the timing patterns and the test takers’ performance on the items/task, demographics, motivation level, personality, and test-taking behaviors through use of different statistical approaches. We found that the five analyses complemented each other and revealed different useful timing aspects of this test-taker sample’s behavioral features in the simulation-based task. The findings were also compared with notable existing results in the literature related to timing data.

show abstract

Compensatory and non‐compensatory multidimensional randomized item response models

Cited by 10 publications

References 34 publications

Multidimensional latent trait linear mixed model: an application in clinical studies with multivariate longitudinal outcomes

Multidimensional latent trait linear mixed model: an application in clinical studies with multivariate longitudinal outcomes

RRreg: An R Package for Correlation and Regression Analyses of Randomized Response Data

How Do Test Takers Interact With Simulation-Based Tasks? A Response-Time Perspective

Contact Info

Product

Resources

About