HKIER

Abstract

This article examines the results which have emerged from the first live administration of the Language Proficiency Assessment for Teachers of English (previously known as the English Language Benchmark Test). The article examines the method by which pass marks (or "cut scores") are calculated for each paper. In particular, the article examines the two methods by which pass marks may be arrived at for the two formal scale-based tests (the Writing Test that comprises five scales, and the Speaking Test that comprises six scales.) The first method requires test takers to achieve a pass (Level 3) on every scale (the "pass-every-scale" method); the second method bases a pass on test takers achieving an overall average of Level 3 (the "aggregate" method). In order to provide comparative data by which to view the results of the live administration, the article examines the results of the Writing Test and the Speaking Test from the Pilot Benchmark Assessment (English) (PBAE) administered in 1999. The PBAE was a test bed for the benchmark assessment framework now being implemented. By providing a context for the results of the live administration, the article describes how pass marks can vary by as much as 20% using the two different methods of calculation. The article concludes with the comment that the recommendation that test takers should pass every scale was proposed with the laudable intent of "raising standards." This well-intentioned move, however, is perhaps losing sight of the Education Commission's original intention with regard to the investigation of benchmarks, i.e., the establishing of a minimum standards test for English language teachers.

摘要

N.A.