Selecting an Appropriate Method for Linking the Scores of Academic Achievement with those of the National University Entrance Examination

The purpose of this study was to find an appropriate method for linking the academic achievement scores to the National University Entrance Examination scores in Iran in comparison to the current practice (normalized method) based on Conditional Standard Error for Measurement. For this purpose, the scores of randomly selected of 5,000 high school graduates of the Mathematics and Physics in 2014 who had participated in National University Entrance Examination of 2015 were analyzed. Based on the criteria introduced by Dorans (2004) and the viewpoints of Pommerich et al. (2004), the concordance was chosen for linking the scores. Accordingly, linear, mean, and equipercentile methods were used to link scores on academic achievement and National University Entrance Examination tests. Results indicated that the equipercentile method were less sample dependent, so test scores linked using this method were more valid for making admissions decisions. Finally, the results showed that the conditional standard error of measurement of the linked scores using the equipercentile method is much lower than the method which is currently being used by the National Organization of Educational Testing in Iran.