Test translation in a South African context using the peabody picture vocabulary test-revised.

Abstract

This study explored the feasibility of using test translation to adapt a standard English vocabulary test for Zulu speakers. It was motivated by the difficulties associated with assessing speakers of the African languages, in the absence of Zulu speech-language therapists and Zulu assessment tools. The adaptation process in the present study began with a direct translation of the PPVT-R into Zulu by twenty Zulu university students. Based on the degree of consensus on translations and judgements of cultural appropriacy (using the committee approach), items were selected for the purposes of pilot testing of the translations on 107 Zulu pupils from six to eleven years, in the study area. The results of the first pilot study revealed that it was not feasible to standardise the first translated version of the test, as only seventeen percent of the items were found suitable for inclusion in a normative study. Further test development using translations from twenty educators, their judgements of cultural appropriacy of stimuli as well as application of the back translation test to determine semantic equivalence of translations resulted in the development of a revised translated version of the PPVT-R, consisting of mUltiple translations for some items. This was administered to 544, six to eleven year old Zulu children. The results of the second pilot study revealed that only 31.2% of the translations administered were appropriate for the purposes of developing norms, across all the age groups tested, with 80% of these translations applicable for six and seven year old Zulu children only. Therefore, the translated version of the PPVT-R, despite modifications, showed significantly reduced applicability for Zulu speakers. Using the information derived from the present study a test comprising thirty six Zulu words has been compiled for the purposes of screening the receptive vocabulary skills of six and seven year old Zulu children in the study area. The findings of the study confirmed the difficulties in using test translation as a procedural option in adapting a test for Zulu children; implications, further modifications and investigations are suggested in this regard.