Bayesian Analysis of a Quantile Multilevel Item Response Theory Model

Multilevel item response theory (MLIRT) models are used widely in educational and psychological research. This type of modeling has two or more levels, including an item response theory model as the measurement part and a linear-regression model as the structural part, the aim being to investigate the relation between explanatory variables and latent variables. However, the linear-regression structural model focuses on the relation between explanatory variables and latent variables, which is only from the perspective of the average tendency. When we need to explore the relationship between variables at various locations along the response distribution, quantile regression is more appropriate. To this end, a quantile-regression-type structural model named as the quantile MLIRT (Q-MLIRT) model is introduced under the MLIRT framework. The parameters of the proposed model are estimated using the Gibbs sampling algorithm, and comparison with the original (i.e., linear-regression-type) MLIRT model is conducted via a simulation study. The results show that the parameters of the Q-MLIRT model could be recovered well under different quantiles. Finally, a subset of data from PISA 2018 is analyzed to illustrate the application of the proposed model.

THE IMPACT OF LIFESTYLE BEHAVIORS ON HEALTHY AGING TRAJECTORIES: THE ATHLOS PROJECT

The number of people above 60 years old will double by 2050. There is a considerate variability in the health status of older people. The identification of the different trajectories that people follow as they grow older constitutes one of the aims of the ATHLOS project. In the current study, we created a metric of health in the four available waves (2001, 2003, 2012, 2015) of the Mexican Health and Aging Study (MHAS) by employing Bayesian multilevel Item Response Theory. Growth mixture modelling indicated that older Mexicans (n=14,143) age by following four distinct pathways (i.e. high-stable, moderate-stable, low-stable, decliners). Adherence to healthy lifestyle behaviours (i.e. physical activity, non-smoking, limited alcohol consumption) was associated with better health trajectories. Preliminary analyses in the ATHLOS harmonised dataset also suggest that older people age by following four distinct pathways. The impact of lifestyle behaviours within the harmonised dataset will be investigated and also presented.

CREATION OF A COMMON METRIC OF HEALTH STATUS IN THE HARMONIZED DATASET OF THE ATHLOS PROJECT

Although life longevity has increased across the world, evidence suggests some heterogeneity of the ageing process across individuals. To investigate different ageing patterns, the ATHLOS project harmonised data from 411,000 individuals across 17 existing cohort studies. The harmonised dataset provides comparable information on functioning measures, cognition, mental health, sociodemographic and lifestyle behaviours. To measure the process of healthy ageing across time and cohorts, we employed a Bayesian Multilevel Item Response Theory(IRT) and created a common metric of health status by using items of functioning. The IRT measurement model includes parameters describing the difficulty and discriminatory power of each item. We adopted the Bayesian Multilevel framework as it allows item parameters to vary among studies and the simultaneous estimation of all parameters under a Markov Chain Monte Carlo(MCMC) method. Finally, we assessed the predictive validity of the metric against mortality by performing a Receiver Operating Characteristic(ROC) curve analysis.

Assessing the Violent Offending and Violent Victimization Overlap Among a Sample of Chinese Youth and Young Adults

Although an increasing number of studies have examined offending and victimization in China, little is known about the extent of the victim–offender overlap and what factors may explain differential tendencies toward offending versus victimization in the Chinese context. To fill this gap and to broaden the empirical base of the current literature on the victim–offender overlap, we examine the association of a number of prominent theoretical constructs with overall violent encounters and role differentiation. Using data collected from over 2,000 students from six schools in Changzhi, a city in Northern China, and multilevel item response theory modeling, we find a nontrivial overlap between violent offending and victimization and differential tendencies toward offending versus victimization. We also find that most of our theoretical constructs are associated with overall violent encounters, but only moral beliefs, peer delinquency, drinking, and gender are related to role differentiation. Findings for theory and research are discussed.

That Door You Just Kicked in Was Locked for Your Protection, Not Mine

Objectives: This research investigates tendencies for individuals to preference adopting weaponry to protect their home over unarmed defensive measures such as installing a lock or alarm. We extend the subculture of violence perspective to account for specific choices and test this approach against hypotheses related to situational reactions to threat. Methods: To explore differential preferences in crime prevention choices for protecting the home, we use data from 1,961 Seattle adults, collected during 2002 to 2003. We employ Osgood and Schreck’s multilevel item response theory–based method as our statistical approach. Results: The results indicate that those who endorse the values of the subculture of violence are more likely to have defensive weaponry among their countermeasures against crime, although the results also show that those who believe the police act justly are more likely to procure weapons. Situational reactions to threat also influenced choices, though not always in the predicted direction. Conclusions: Beliefs may be an important determinant for how people protect themselves against crime. Further, criminological theory can successfully explain crime prevention choices in the general population, indicating considerable untapped future directions for research.

Introduction to Multilevel Item Response Theory Analysis

Item response theory (IRT) models are the main psychometric approach for the development, evaluation, and refinement of multi-item instruments and scaling of latent traits, whereas multilevel models are the primary statistical method when considering the dependence between person responses when primary units (e.g., students) are nested within clusters (e.g., classes). This article introduces multilevel IRT (MLIRT) modeling, and provides the basic information to conduct, interpret, and report results based on an analysis using MLIRT modeling. The procedures are demonstrated using a sample data set based on the National Institute for the Evaluation of School System survey completed in Italy by fifth-grade students nested in classrooms to assess math achievement. The data and command files (Stata, M plus, flexMIRT) needed to reproduce all analyses and plots in this article are available as supplemental online materials at http://jea.sagepub.com/supplemental .

Measuring Teachers' Instruction with Multilevel Item Response Theory

The purpose of this study was to describe an approach for measuring teachers' uses of instruction as it relates to students' achievement through classroom observations. Despite significant work on the substantive content of observation systems chronicling teachers' instruction, literature has largely relied on simple counts of instructional features or the average of quality indicators to describe teachers' instruction. However, such coarse summaries generally do not reflect current theories of instruction, prior empirical evidence, and the framework of most observation systems. The approach presented in this paper builds on evidence that teachers' instruction varies across lessons and that instructional features or quality indicators do not necessarily contribute equally to our understanding of effective instruction. To align theory, data and methods, this study applied multilevel item response theory to the study of early literacy instruction as it relates to students' achievement. This model provided a more complex, but more precise and theoretically grounded, view of instruction by linking components of instruction theory to model parameters. Empirical results suggested that multilevel item response models encouraged precision in the specification of theory, data collection, and models that is absent in simpler models.