Short-cut estimators of criterion-referenced test consistency
- 1 June 1990
- journal article
- research article
- Published by SAGE Publications in Language Testing
- Vol. 7 (1) , 77-97
- https://doi.org/10.1177/026553229000700106
Abstract
The purpose of this paper is to present relatively easy-to-calculate estimates for the consistency of criterion-referenced tests (CRTs). Four techniques are demonstrated using data from a CRT administered in the intermediate ESL reading classes in the English Language Institute at the University of Hawaii at Manoa. First, the threshold loss agreement approach is represented by the agreement coefficient (Subkoviak 1980) and kappa coefficient (Cohen 1960). Until recently, these coefficients could only be applied in test-retest CRT situa tions. Alternative and simpler methods (suggested by Subkoviak 1988) are given for estimating these coefficients from the results of a single CRT administration. Second, the squared-error loss agreement approach is repre sented by the phi(lambda) dependability index, for which simplified calcula tions, derived from Brennan (1980), are given. Finally, the domain score dependability approach is represented by a short-cut method for estimating a general purpose phi coefficient. This short-cut estimate is derived from Brennan's 1983 and 1984 discussions of the phi coefficient.Keywords
This publication has 15 references indexed in Scilit:
- A criterion-referenced measurement approach to ESL achievement testingLanguage Testing, 1984
- A CATEGORICAL INSTRUMENT FOR SCORING SECOND LANGUAGE WRITING SKILLSLanguage Learning, 1984
- AN INTRODUCTION TO GENERALIZABILITY THEORY IN SECOND LANGUAGE RESEARCH1Language Learning, 1982
- Improving the Psychometric, Criterion-Referenced, and Practical Qualities of Integrative Language TestsTESOL Quarterly, 1982
- Signal/Noise Ratios for Domain-Referenced TestsPsychometrika, 1977
- AN INTERPRETATION OF LIVINGSTON'S RELIABILITY COEFFICIENT FOR CRITERION‐REFERENCED TESTSJournal of Educational Measurement, 1972
- A “UNIVERSE‐DEFINED” SYSTEM OF ARITHMETIC ACHIEVEMENT TESTS1Journal of Educational Measurement, 1968
- The Signal/Noise Ratio in the Comparison of Reliability CoefficientsEducational and Psychological Measurement, 1964
- Instructional technology and the measurement of learing outcomes: Some questions.American Psychologist, 1963
- A Coefficient of Agreement for Nominal ScalesEducational and Psychological Measurement, 1960