Effect of correction-for-guessing formula on psychometric characteristics of test

Abstract

The purpose of this research is to determine if there are any differences in test and item statistics when the correction for guessing formula is used and not used in a Turkish reading comprehension test consists of multiple-choice items. Proposed study is a correlational research and the study group consists 123 8th grade students in a secondary school. For the stated purpose, a Turkish language test is generated which consists of 15 items all of them chosen among the Turkish items of Central Educational Institutions Student Selection and Placement Exam (OKOSYS) and the Private Schools Examination (OOS). The first test application was conducted with the instruction of "correction for guessing formula will use for the test score". After a 30-day period, for impeding retention affect, same test applied with the "correction for guessing formula will not use for the test score" instruction. Student achievement, test and item statistics compared and interpreted according to the data obtained from both applications. Findings point out that there is a significant difference in test and item when statistics between the two applications of correction formula is used and not used. However the effect size was determined to be moderate. It has never been a significant difference on item statistics. Findings were discussed in context of related literature.