@article {16, title = {Maximum information stratification method for controlling item exposure in computerized adaptive testing}, journal = {Psicothema}, volume = {18}, number = {1}, year = {2006}, note = {Barrada, Juan RamonMazuela, PalomaOlea, JulioResearch Support, Non-U.S. Gov{\textquoteright}tSpainPsicothemaPsicothema. 2006 Feb;18(1):156-9.}, month = {Feb}, pages = {156-159}, edition = {2007/02/14}, abstract = {The proposal for increasing the security in Computerized Adaptive Tests that has received most attention in recent years is the a-stratified method (AS - Chang and Ying, 1999): at the beginning of the test only items with low discrimination parameters (a) can be administered, with the values of the a parameters increasing as the test goes on. With this method, distribution of the exposure rates of the items is less skewed, while efficiency is maintained in trait-level estimation. The pseudo-guessing parameter (c), present in the three-parameter logistic model, is considered irrelevant, and is not used in the AS method. The Maximum Information Stratified (MIS) model incorporates the c parameter in the stratification of the bank and in the item-selection rule, improving accuracy by comparison with the AS, for item banks with a and b parameters correlated and uncorrelated. For both kinds of banks, the blocking b methods (Chang, Qian and Ying, 2001) improve the security of the item bank.M{\'e}todo de estratificaci{\'o}n por m{\'a}xima informaci{\'o}n para el control de la exposici{\'o}n en tests adaptativos informatizados. La propuesta para aumentar la seguridad en los tests adaptativos informatizados que ha recibido m{\'a}s atenci{\'o}n en los {\'u}ltimos a{\~n}os ha sido el m{\'e}todo a-estratificado (AE - Chang y Ying, 1999): en los momentos iniciales del test s{\'o}lo pueden administrarse {\'\i}tems con bajos par{\'a}metros de discriminaci{\'o}n (a), increment{\'a}ndose los valores del par{\'a}metro a admisibles seg{\'u}n avanza el test. Con este m{\'e}todo la distribuci{\'o}n de las tasas de exposici{\'o}n de los {\'\i}tems es m{\'a}s equilibrada, manteniendo una adecuada precisi{\'o}n en la medida. El par{\'a}metro de pseudoadivinaci{\'o}n (c), presente en el modelo log{\'\i}stico de tres par{\'a}metros, se supone irrelevante y no se incorpora en el AE. El m{\'e}todo de Estratificaci{\'o}n por M{\'a}xima Informaci{\'o}n (EMI) incorpora el par{\'a}metro c a la estratificaci{\'o}n del banco y a la regla de selecci{\'o}n de {\'\i}tems, mejorando la precisi{\'o}n en comparaci{\'o}n con AE, tanto para bancos donde los par{\'a}metros a y b correlacionan como para bancos donde no. Para ambos tipos de bancos, los m{\'e}todos de bloqueo de b (Chang, Qian y Ying, 2001) mejoran la seguridad del banco.}, keywords = {*Artificial Intelligence, *Microcomputers, *Psychological Tests, *Software Design, Algorithms, Chi-Square Distribution, Humans, Likelihood Functions}, isbn = {0214-9915 (Print)}, author = {Barrada, J and Mazuela, P. and Olea, J.} }