Examinations, Marks, Grades and Scales: A Working Paper.

Comparisons of students' educational performances are usually based on test and examination results. However, for such comparisons to be valid, it is suggested that evaluations must be made on some common basis since many educational and employment decisions are based on these evaluations. Standardized tests, often used when comparisons are to be made, offer some guarantee of a common evaluative basis, even though they may contain measurement errors. Among the many transforming and scaling work techniques that have been suggested, two measurements are considered useful in terms of groups of scores, that is, central tendency (mean and median) and dispersion (standard deviation and interquartile range) measures. It is suggested that while transformations are required in order to validate results from different examinations, one should consider what purpose the transformation is to serve. Also, discussed are transformations that convert marks to T scores and the differences among percentiles, ranks, and standard scores. It is concluded that certain basic educational problems are revealed when scaling or transformation techniques are used. (JS)