Scholarly record

EFFECT OF USING L1 AND L2 PROMPTS ON WRITING PERFORMANCE IN HIGH STAKES TESTS

Maria Sikolova, I. Mrozkova, L. Kolackova, K. Kucerova, J. Dooley

Abstract

This paper is a contribution to the field of language assessment, namely testing writing. It focuses on the design of test tasks and discusses the advantages and disadvantages of delivering the instructions and prompts in the target (L2) versus native (L1) languages. It analyses the impact of target and native language prompts on the test takerВ’s performance. The authors analyse the writing ability test results of a high-stakes proficiency examination using two sets of prompts, one given in L1, and one in L2. Both sets of writings were assessed according to the standard scoring procedure, and the results were compared as to the language level achieved. Additionally, the test takersВ’ perception of the possible difficulties in comprehending the prompts was assessed by questionnaires filled in after completing both sets of tests. The consolidated results showed that the majority of the test takers achieved the same language proficiency level on both tasks regardless of whether the prompts were delivered in L1 or L2. When the results differed, in almost all cases the test takers achieved better results in writing using native language prompts, compared to only one candidate who performed better using the L2 prompts. The perception of the difficulty of the instructions and prompts in L1 and L2 by the test takers has shown that fewer than 50% of them believed that the writing test with the instructions and prompts given in the target language was more difficult than those who received directions in L1. The primary assumption that providing instructions and prompts in the native language would positively influence results on the L2 writing ability test was thus not confirmed. Moreover, the majority of the test takers did not consider the difference to be a significant factor in their performance.

Publication Impact Profile

No metrics available.

PLUMX see details

DOI resolver: doi.org/10.5593/sgemsocial2018/3.5/S13.050

Publication details

Back to publication list