skip to main content skip to footer

Effects of Amount of Time Allowed on the Test of Written English TWE

Author(s):
Hale, Gordon A.
Publication Year:
1992
Report Number:
RR-92-27
Source:
ETS Research Report
Document Type:
Report
Page Count:
42
Subject/Key Words:
English (Second Language), English Tests, Performance Factors, Test Length, Test of Written English (TWE), Timed Tests

Abstract

This study examined students' essay performance on topics from the TWE® (Test of Written English™) under two time limits--30 minutes, as on the current TWE, and 45 minutes. In the main groups of the study, each student wrote an essay on one topic under one time limit and on another topic under the other time limit (with orders counterbalanced). The correlation between scores for 30- versus 45-minute conditions was relatively high and approached the parallel-form reliability of the task, as indicated by data from students who wrote essays on separate topics under the same time limits. Thus, the provision of additional time apparently had little effect on the standing of the students in relation to each other. Both parallel-form reliability and interrater reliability were approximately the same for the 30- and 45-minute conditions. Mean scores on the 6- point TWE scale were found to be higher by about 1/4 to 1/3 point under the 45-minute condition than the 30-minute condition, indicating that provision of additional time produced a modest but reliable increase in scores. The magnitude of the effect was roughly comparable for students of low versus high proficiency, and for students in intensive English programs versus students in academic coursework. Responses to a questionnaire indicated that the students regarded 45 minutes as more sufficient for accomplishing the task than 30 minutes. The results are discussed in relation to the literature on time effects and to practical implications for the TWE.

Read More