Abstract :
Item response theory (IRT) is used commonly for design of tests, test assembly, test scaling and calibration, construction of test item banks, investigations of test item bias, and other common procedures in the test development process. Logistic function is the most popular unidimensional IRT model. Applications of IRT models require obtaining the set of values for logistic function parameters that best fit an empirical data set. However, success in obtaining such set of values does not guarantee that the constructs they represent actually exist, for the adequacy of a model is not sustained by the possibility of estimating parameters. In this study, an Equal ARea Logistic estimation algorithm (EARL) is proposed. The EARL algorithm can estimate the parameters, which are discrimination, difficulty and pseudo-guessing level, of the logistic model. Numerical results of two-parameter and three-parameter estimation are presented to show the stability, accuracy and time-saving of the algorithm.
Keywords :
design for testability; logistics; parameter estimation; design of tests; equal area logistic estimation algorithm; item response theory; logistic function; test assembly; test item banks; test item bias; test scaling; three-parameter estimation; three-parameter logistic model; unidimensional IRT model; Assembly; Calibration; Conference management; Iterative algorithms; Logistics; Maximum likelihood estimation; Parameter estimation; Technology management; Testing; Transportation; equal area construction; item response theory; logistic model;
