29 July 2017
Development of Assessment Instruments within the Evidence-Centered Design (ECD) Approach
The course covers the following topics:
Bayesian Statistics: The lecture will give a brief overview of how to develop and use Bayesian models for extending probabilistic reasoning about student’s proficiencies beyond reasoning conducted in the traditional Frequentist statistical framework. The lecture will describe how Bayesian statistics can be a guiding principle for reasoning about complex psychometric models which are not tractable under the frequentist approach. In addition, the lecture will briefly touch some modern computing algorithms like Markov Chain Monte Carlo (MCMC) which can solve problems cast in the Bayesian statistical framework.
Graphical Models: The lecture will be concerned with answering a question of how the underlying principles of many assessment designs – the psychometric model should reflect the cognitive model in a manner that fits the purpose of the assessment – could be addressed by graphical networks. This lecture provides some basic foundations of graph theory and graphical models and links them to Bayesian networks (i.e., types of Bayesian networks – graphical models in which all of the variables are discrete – are very popular in artificial intelligence communities and gaining some popularity among psychometricians).
Putting It All Together: The lecture concerns with putting ECD, Bayesian statistics and graphical networks together for the purpose of describing evidence of student’s KSA, task, assembly, and presentation models necessary for the assessment. The lecture will briefly discuss how to build Bayesian probability models that are embedded in a graphical structure that reflect our knowledge about the student’s KSA, defined by ECD, and propagate evidence through the graphical model in order to make a claim about a student.
Mark Zelman (PhD, World Bank Group Consultant)
People working with social measurements;
Master and Ph.D. students.
EUR 670: Includes all materials, meals, accommodation and transportation from Moscow and back.
For several participants who will be selected through competition the participation is free (including meals and accommodation). NRU HSE provides 10 grants covering 100% of participation fee. Additionally, 5 grants, covering 50% of fee.