<![CDATA[This study aims to describe the estimation point of measurement error in mathematics tests using the classic theory and to find out the effect of the estimation model of measurement errors. This study population is all participant answer sheets of UN SMA mathematic test conducted in 2009/2010 academic year in the city of Yogyakarta in the form of a multiple choice test. The sample was established by stratified proportional systematic random sampling. The data were in the form of documentation and analyzed with the Iteman program. The model used for the analysis is Thorndike, , Kristoff, and Feldt models The research results can be summarized as follows. For the error estimation model with the classic theory, the average error estimation of the mathematics test is in the range of 2.65742 £ SEM £ 3.544352. In the classic theory, Thorndike model yields the biggest measurement error estimation, which means Thorndike model is the most effective model to estimate measurement error in a mathematic test.]]>
Copyrights © 2012
