Introduction: While back translation has been widely used in medical research surveying linguistically diverse populations, research literature often fails to document this complex translation process. Our study examines inadequacies in the use of back translation, suggests improvements, as well as suggesting where other translation strategies may be more appropriate.Areas covered: This paper cites numerous metastudies showing how back translation is often uncritically adopted in validation of research instruments, pointing to potential methodological failings, before examining the back-translation processes in an Australian study of non-English speaking cancer patients. Our study of back translation applied to patient self-report questionnaires demonstrates that appropriate renditions of items are critically dependent upon both translator and researcher awareness of item purpose, overall project specifications and identification of linguistic ambiguities in source test items. The poor implementation and documentation of back-translation processes in many studies indicate alternatives to back translation may be appropriate.Expert opinion: Where translations are used in research, translation processes need to be made explicit in research protocols and reports, and translation experts need to be part of the research team, with translation guidance and advice integrated into all stages of research design.