Implementation of the "Streamlined" Canadian Language Benchmarks Placement Test (CLBPT)
FUNDED BY CITIZENSHIP AND IMMIGRATION CANADA

Project Update
June 22, 2002
The CLBPT was launched at TESL Canada Conference in Regina, Saskatchewan in May 2002. For more information about the CLBPT implementation please see our website. Also for CLBPT Training and Support project please see 2002-2003 Projects above.

CLBPT will be launched at TESL Canada
May 2002 Conference for use in ESL and LINC

Project Goal and Outcomes
The goal of this project is to carry out widespread implementation of the new "streamlined" Canadian Language Benchmark Placement Test (CLBPT) for use in LINC and non-LINC ESL programs. The CLBPT includes:

• CLBPT Test Kit consisting of: Guide Book, Reading and Writing Tests, a Listening/Speaking Assessment Log, an accompanying photo-spread and audio tape, cover sheets for the CLBPT and addendum notes
  
  
• CLBPT Training package to include: CLBPT Training Manual and Video Cassette (10 taped interviews)
  

Rationale/Need
Use of the streamlined CLBPT will enable LINC Assessment Centres to reduce backlog in high-volume intake periods in LINC Assessment Centres across Canada, and will also provide a reliable tool for ESL instructors to use to test and place students in remote areas and other locations where CLBA-trained assessors are not available.

Background:
In 1999-2000 the CCLB initiated a project to develop a task-based placement instrument to determine the ESL proficiency of adults in each of the four language skill areas: listening, speaking, reading and writing. The "Streamlined Test" (CLBPT) was to provide an alternative to the traditional and more thorough Canadian Language Benchmarks Assessment (CLBA) instrument in order to assist ESL instructors in the testing and placement of students in remote areas and other locations where CLBA-trained assessors are not available; and to help CLBA-trained assessors reduce backlog in high-volume intake periods in centres across Canada. The contract for the development of the CLBPT was awarded to the Association of Adult and Continuing Education School Board Administrators of Ontario (CESBA).

FUNDING PROVIDED BY THE ONTARIO GOVERNMENT.

In July 2001 a pilot study was conducted in order to measure the comparability of the Streamlined Placement Test (CLBPT) to the CLBA in scoring reliability and administrative efficiency. Where the traditional CLBA assessment test takes an average of 2.5 hours to administer to Stage II students, initial testing showed that the CLBPT typically takes 1.25 hours (30 minutes for each of Reading and Writing and 15 minutes for Listening and Speaking combined). The results of the study, based on a cross-Canada pilot over the summer of 2001, are described in the Executive Summary of the Final Report below.

FUNDING PROVIDED BY CITIZENSHIP AND IMMIGRATION CANADA, AS WELL AS THE ONTARIO AND ALBERTA GOVERNMENTS AND IN-KIND CONTRIBUTIONS FROM THE MANITOBA GOVERNMENT.

Examination of the Validity of the Canadian Language Benchmarks Placement Test
Executive Summary, August 14, 2001
By: Philip Nagy The Ontario Institute for Studies in Education of the University of Toronto

An examination was undertaken of the levels of agreement between the Canadian Language Benchmark Placement Test (CLBPT) and the Canadian Language Benchmarks Assessment (CLBA), and the time saved by using the CLBPT. The sample available, just over 500, was not as large as the 1600 that had been hoped for. The sample was adequate to establish the overall level of agreement between the two instruments, but not the level of agreement at each benchmark. We can be certain, with 95% confidence, that the CLBPT agrees perfectly with CLBA between 33% and 41% (37% ± 4%) of the time, and agrees within one benchmark between 79% and 85% (82% ±3%) of the time. Considering that the CLBA is itself considered accurate within ± one benchmark, this is quite a good level of agreement. Except that the CLBPT gave slightly higher average results in Listening/Speaking and in Reading at the middle benchmarks, the disagreements were symmetrically distributed. The CLBPT can be used for the same low stakes decisions for which the CLBA was designed. Time saved with the shorter test increased with benchmark level. Total testing time saved was from about 15 minutes at Benchmark 1 to 75 minutes at Benchmark 8. However, there were numbers of cases where the CLBPT test took as long or longer than the CLBA test: for Listening/Speaking, 30% of respondents, for Reading, 20%, and for Writing, 20%. Attempts at diagnosing the source of disagreement were hampered by the small samples. Not only were there lower levels of test writing than expected, but the high level of agreement further lowered the available number of cases of disagreement. Nevertheless, it was possible to identify some sites where Listening/Speaking agreement was lower than others. Administrators at the pilot assessment sites provided substantial feedback on the validation process. Many of the difficulties identified focused on the lack of time available for the validation. The tapes used in the Listening/Speaking tests, and the written forms for all tests, need to be upgraded to production quality. Test instructions and training procedures require clarification and fine-tuning. Some test items require revisions. A working group has been struck to deal with these revisions as a prelude to the implementation process. Further validation data will be collected as required.