ANALYSIS OF SHORT-ANSWERED AND MULTIPLE-CHOICE ITEMS SCORED VIA “1-0” AND WEIGHTED SCORING METHODS ACCORDING TO PSYCHOMETRIC CHARACTERISTICS OF TESTS

Goksu Gozen
Ankara University

Abstract

The common argument about the used for multiple-choice and short-answered items is that conventional method ignores partial information by giving identical weights to items and that it includes pseudo chance. Subject to insufficiency of conventional scoring method, several alternative scoring methods that specified weighting test items and/ or item options were suggested. The purpose of this research is to determine the effect of weighting short-answered and multiple choice items on individual test scores and test validities. The study was carried out in Ankara and Hacettepe Universities on 316 first year students attending the departments of Elementary Education and Counseling and Guidance in spring semester of the academic year 2001-2002. To be used in administration, two ability tests of multiple-choice and short-answered items dealing with verbal ability and reasoning were constructed. Consequent to the administration, test items were scored by the methods of (a) conventional scoring (1-0), (b) weighting by item difficulty (1-pj) and (c) weighting by item difficulty and item discrimination [(1-pj)rjx]. The relationship between the scores collected from various scoring methods and the difference between the validity coefficients were examined. Results showed that these three scoring methods did the same effect on the test scores and caused no significant difference in test validities.

Keywords

Item weighting, scoring methods, multiple-choice items.