Especificación del algoritmo para un test adaptativo informatizado de analogías verbales / Algorithm specification for a computerized adaptive test of verbal analogies

Se presentan los pasos seguidos en el diseño de un Test Adaptativo Informatizado de Analogías Verbales. El algoritmo adaptativo se determinó teniendo en cuenta las características de un Banco de Ítems de Analogías Verbales construido a partir de la Teoría de Respuesta al Ítem, los objetivos de evaluación, la población por evaluar y las características del software. El Banco está compuesto por 64 ítems unidimensionales calibrados según el Modelo Logístico de Tres Parámetros, informativos en todo el rango del rasgo y libres de funcionamiento diferencial entre géneros. Se programó el algoritmo adaptativo con el FastTEST Professional Testing System utilizando las siguientes especificaciones: a) inicio aleatorio entre niveles levemente inferiores a la media del rasgo, b) estimación del rasgo por Máxima Verosimilitud Condicional, c) selección progresiva de los ítems con el Método de Máxima Información de Fisher y d) criterio mixto de finalización.

This paper presents the steps followed in the design of a Computerized Adaptive Test to measure Verbal Analogies. The adaptive algorithm was determined upon the characteristics of a Verbal Analogies’ Item Bank constructed on the basis of Item Response Theory, the assessment objectives, the studied population and software features. The Bank consists of 64 unidimensional items calibrated with the Three Parameter Logistic Model, which are informative throughout the entire latent trait’s range and free of differential item functioning between genders. The adaptive algorithm was programmed via FastTEST Professional Testing System using the following specifications:a) random start between levels slightly below average trait, b) Conditional Maximum Likelihood Trait Estimation, c) progressive items selection by Maximum Fisher Item Information, and d) combined stopping rule.

Exploring methods of the difficulty estimate of medical microbiology examination questions / 中华医学教育探索杂志

In the item bank construction of medical microbiology examination questions,the accurate difficulty coefficient is an important parameter to guarantee the quality of the bank.Thus,the accurate difficulty estimate plays an important role in the item bank construction of examination questions.Methods of estimating the difficulty of medical microbiology test questions were explored,based on the knowledge points,the type of test questions,the structure of test questions and the degree of students’ familiarity with the questions and so on.Finally,the feasibility of methods was confirmed through the simulated tests.

A Development of Computer-Based Examination(CBE) System for Medical Students / 한국의학교육

A Computer-Based Examination(CBE) System is developed to take the effective examination for medical student. The server system is operated with Windows NT(Korean ver. 4.0) and the clients system with Windows 95(later than ver. OSR 2), and the Microsoft SQL server(ver. 7.0) is used for database server, and the Inprise Delphi(ver. 4.02) for development tool. This system consists of five subsystems(item bank, item selection, implementation, item analysis). The CBE system is designed to execute the multimedia data(image, sound, movie), and for professors to build question items, to extract the items for examinations on this system, and for students to conduct the examination on the client computer systems. It will reduce time to mark examination papers and to analyze the items, and can be applied for self-studying(computer assisted learning, CAL) with linking to internet or knowledge-base system.