Item Response Model Using In Longitudinal Muti-level Data

In the field of potential capacity development in the research. Researchers havebeen a lack of effective mathematical methods and corresponding model to handlehaving a longitudinal, multi-level feature data. For this situation, We have establisheda two-tier hierarchical model: The first layer is a layer of time, described the potentialcapacity of the individual’s development over time; The second layer is a layer factor, studied the factors that subjects already have potential ability before enrollment in thepaper. According to case that subjects answer in the exam (questions are drawn fromthe exam), combined with a hierarchical model and item response models (simulationonly normal ogive model), we estimate the subject’s ability potential trends over time, Meanwhile, speculate his(her) development in future under subjects already havedifferent potential ability before enrollment. In the process of model’s parameterestimation. We use Gibbs sampling,that belong to Markov Chain Monte Carlo(MCMC). And using the parameters of the posterior mean iteration trajectories andautocorrelation function of plans to test receipt and inspection of the its estimated.Finally, simulation results illustrate that the model established in this paper can be agood deal of data having a longitudinal, multi-level features in the field of potentialcapacity development.