Bayesian longitudinal item response modeling with multivariate asymmetric serial dependencies

It is usually impossible to impose experimental conditions in large-scale longitudinal (observational) studies in education. This increases the risk of bias due to for instance unobserved heterogeneity, missing background variables, and dropouts. A flexible statistical model is required for the nature of the observational assessment data and to account for the unexplained heterogeneity. A general class of longitudinal item response theory (IRT) models is proposed, where growth in performance can be monitored using a skewed multivariate normal distribution for the latent variables. Change in performance and unexplained heterogeneity is addressed through structured covariance patterns and skewed multivariate latent variable distributions. The Cholesky decomposition of the covariance matrix is considered to model the dependence structure. A novel multivariate skew-normal distribution is defined by the antedependence model with centered skew-normal distributed errors. A hybrid MCMC approach is developed for parameter estimation, model-fit assessment, and model comparison. Results of simulation studies show good parameter recovery. A longitudinal assessment study by the Brazilian federal government is considered to show the performance of the general LIRT model.

[1]  Silvia Cagnone,et al.  A Composite Likelihood Inference in Latent Variable Models for Ordinal Longitudinal Responses , 2012, Psychometrika.

[2]  Frank B. Baker,et al.  Item Response Theory : Parameter Estimation Techniques, Second Edition , 2004 .

[3]  R. Kohn,et al.  On Gibbs sampling for state space models , 1994 .

[4]  Dani Gamerman,et al.  Markov Chain Monte Carlo: Stochastic Simulation for Bayesian Inference, Second Edition , 2006 .

[5]  A. Béguin,et al.  MCMC estimation and some model-fit analysis of multidimensional IRT models , 2001 .

[6]  Heleno Bolfarine,et al.  Bayesian inference for a skew-normal IRT model under the centred parameterization , 2011, Comput. Stat. Data Anal..

[7]  Sandip Sinharay.  Bayesian item fit analysis for unidimensional item response theory models. , 2006, The British journal of mathematical and statistical psychology.

[8]  Andrew Gelman,et al.  General methods for monitoring convergence of iterative simulations , 1998 .

[9]  D. Zimmerman,et al.  Modeling Nonstationary Longitudinal Data , 2000, Biometrics.

[10]  Jean-Paul Fox,et al.  A Bayesian generalized multiple group IRT model with model-fit assessment tools , 2012, Comput. Stat. Data Anal..

[11]  A. Azzalini.  A class of distributions which includes the normal ones , 1985 .

[12]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[13]  Hal S. Stern,et al.  Posterior Predictive Assessment of Item Response Theory Models , 2006 .

[14]  Juan L. Padilla,et al.  Multidimensional multiple group IRT models with skew normal latent trait distributions , 2018, J. Multivar. Anal..

[15]  Adelchi Azzalini,et al.  The Skew-Normal and Related Families , 2018 .

[16]  Mark D. Reckase,et al.  Item Response Theory: Parameter Estimation Techniques , 1998 .

[17]  Susan E. Embretson,et al.  A multidimensional latent trait model for measuring learning and change , 1991 .

[18]  Jean-Paul Fox,et al.  Bayesian longitudinal item response modeling with restricted covariance pattern structures , 2014, Statistics and Computing.

[19]  H. Bolfarine,et al.  Parameter recovery for a skew-normal IRT model under a Bayesian approach: hierarchical framework, prior and kernel sensitivity and sample size , 2012 .

[20]  M. Pourahmadi.  Joint mean-covariance models with applications to longitudinal data: Unconstrained parameterisation , 1999 .

[21]  D. Andrade,et al.  Item response theory for longitudinal data: Item and population ability parameters estimation , 2006 .

[22]  R. Jennrich,et al.  Unbalanced repeated-measures models with structured covariance matrices. , 1986, Biometrics.

[23]  J. Fox,et al.  Longitudinal multiple-group IRT modelling: covariance pattern selection using MCMC and RJMCMC , 2015 .

[24]  C. Bayes.  Bayesian inference for the skewness parameter of the scalar skew-normal distribution , 2007 .

[25]  D. Zimmerman,et al.  Antedependence Models for Longitudinal Data , 2009 .

[26]  Erling B. Andersen,et al.  Estimating latent correlations between repeated testings , 1985 .

[27]  S. Frühwirth-Schnatter.  Data Augmentation and Dynamic Linear Models , 1994 .

[28]  N. Henze.  A Probabilistic Representation of the 'Skew-normal' Distribution , 1986 .

[29]  D. Andrade,et al.  Item response theory for longitudinal data: population parameter estimation , 2005 .

[30]  D. Dunson.  Dynamic Latent Trait Models for Multidimensional Longitudinal Data , 2003 .

[31]  A multiple group item response theory model with centered skew-normal latent trait distributions under a Bayesian framework , 2013 .

[32]  S. E. Ahmed,et al.  Markov Chain Monte Carlo: Stochastic Simulation for Bayesian Inference , 2008, Technometrics.

[33]  Marc G. Genton,et al.  Skew-elliptical distributions and their applications : a journey beyond normality , 2004 .

[34]  S. Rabe-Hesketh,et al.  An autoregressive growth model for longitudinal item analysis , 2016, Psychometrika.

[35]  A. Pewsey.  Problems of inference for Azzalini's skewnormal distribution , 2000 .