Comparison between Rush Model Parameters to Completed and Lost Data by Different Methods of Processing Missing Data

Abstract

The current study aims to compare between the assessments of the Rush model’s parameters to the missing and completed data in various ways of processing the missing data. To achieve the aim of the present study, the researcher followed the following steps: preparing Philip Carter test for the spatial capacity which consists of (20) items on a group of (250) sixth scientific stage students in the directorates of Baghdad Education at Al–Rusafa (1^st, 2^nd and 3^rd) for the academic year (2018-2019). Then, the researcher relied on a single-parameter model to analyze the data. The researcher used Bilog-mg3 model to check the hypotheses, data and match them with the model. In addition, the researcher relied on chi-squared value for each item at (0.05).  After that, the researcher found out the parameters of the missing data after relying on a loss percentage (10%) and used three ways to treat them (mean, regression, likelihood). The results showed that the comparison between the parameters completed and missing data by using three ways of processing the missing data is in favor of the parameters of the completed data, and the likelihood way is the suitable way to treat the completed data. 

     The conclusions, recommendations and suggestions have been drawn based on the findings.