CONFIDENCE LEVEL RATING IN MULTIPLE CHOICE TEST LEADING TO INCREASE IN RELIABILITY AND DECREASE IN STANDARD ERROR OF MEASUREMENT
Niyati Kapoor, V. Natarajan · 2012
In the present scenario, whenever any test taker is measured, evaluated and assessed, the number of right answers that a test taker provides, forms the basis of his assessment without understanding the fact that for responding a particular test item how much sure s/he has been. Also to be sure about the knowledge of the test taker, it is important to understand whether the test taker is confident about answering it or has the correct answer been the result of mere guesswork. This Classical Test Theory, applied Test and Item Analysis has a very low reliability and high standard error. A new and innovative technique can be introduced in the normal testing scenario which is called the Confidence Level Rating (CLR) which indicates the confidence s/he has in her/his response. Thus the purpose of this paper is to present how beneficial the confidence level rating is in the multiple choice test items and how this technique leads to increase in reliability and thereby reducing the standard error of measurement. Also this research paper is a by-product of an illuminative research project done by the first author and guided by the second author which offers various innovative practices in the field of learning and assessment which are conducive to the learning process of a learner in school education or otherwise,one of it being Confidence Level Rating. KEY WORDS Confidence level rating, Confidence-based learning, assessment of learning, assessment for learning, number right scores….. 1.