<![CDATA[The purpose of this research is to provide information and education and training to teachers, related to multidimensional item measurement for AKM on students' Literacy and Numeracy abilities. An item that is multidimensional, if measured using unidimensional it will cause errors in measurement. This research is an experimental research using software. The data used is empirical data, namely the value of student report cards. Data collection was carried out in January 2023. Data was obtained from the Ternate City Education Office in the field that handles Kindergarten and Elementary Schools regarding the use of assessment of student learning outcomes. The population used is all students of SDN in Ternate City. Of all the 29 elementary schools which have implemented this independent curriculum, the sample used was 1000 students which was carried out by random sampling. Data collection was carried out by observation, tests, and interviews with students and teachers The minimum value is 0.033 and the maximum value is 0.12. It is also shown that the length of the lower whisker and the upper whisker are the same. It is also shown that the length of the lower whisker and the upper whisker are not the same. This means that using a sample size of 1000 and a test length of 40, the multidimensional model is better than the unidimensional model because the whisker length in the multidimensional model is smaller than the unidimensional model. That is, the shorter the whisker, the smaller the variance or the more precise the data. The estimation results of the multidimensional test parameters have a minimum score of 0.464 in dimension one and a maximum score of 1.952, a minimum score in dimension two of 0.518 and a maximum score of 1.957. For the overall measure of difficulty (DIFF) for the first and second dimensions, both unidimensional and multidimensional, the mean > SD is obtained. This shows that the distribution of variations in test parameters and ability parameters is smaller so that it can be said that the estimation of multidimensional test items at a sample size of 1000 and a test length of 40 has a smaller level of deviation]]>
Copyrights © 2023
