Equating Two Alternate Forms of Basic Statistics Test Using the Single Group Random Design: An Application of the Traditional Equating Methods

Abstract

Most testing situations in higher education institutions utilise different tests which measure similar or the same psychological construct to safeguard the security of the various test and to improve the validity of scores obtained from the examinations. This requires strategies like developing parallel tests and equating the scores. This paper demonstrates the application of traditional equating methods to equate two alternate forms of Basic Statistics test using the single group random design. Two alternate test forms (Form X and Form Y) were developed by test experts with measurement and statistics background. The development of the test forms was closely guided by the test specification table and item specification blueprint. The tests were administered to 146 students who were sampled through the convenience sampling technique. Half of the sample (n=73) took Form X and the other half were administered Form Y. The findings showed that relatively, the equipercentile equating appeared to produce scores that were similar and also within the range of the Form X scores. Consequently, the equipercentile equating was found to be statistically accurate for equating.